Quasi-global, land-only, high-resolution and spatially averaged climate variables from downscaled CMIP6 models for climate impact research

Quasi-global, land-only, high-resolution and spatially averaged climate variables from downscaled CMIP6 models for climate impact research | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF data-descriptor Quasi-global, land-only, high-resolution and spatially averaged climate variables from downscaled CMIP6 models for climate impact research Sally Jahn, Katy A M Gaythorpe, Ilaria Dorigatti, Peter Winskill, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9508034/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Global climate models (GCMs) are essential tools for understanding the climate system and projecting its evolution under different scenarios. However, differences in model construction introduce uncertainties, and GCMs have coarse resolution and inherent biases, limiting their effectiveness for informing local or regional adaptation and mitigation planning. We apply the statistical Double Bias-Corrected Constructed Analogues (DBCCA) method to generate bias-corrected and downscaled climate projections at daily and 0.1° spatial resolution for 1985–2100, with a specific focus on supporting tropical health-related impact research. Our quasi-global, high-resolution projections are based on six GCMs from Phase 6 of the Coupled Model Intercomparison Project (CMIP6) under two emission scenarios (SSP2-4.5 and SSP5-8.5), covering 12 land-only domains between 60°N and 60°S. Moreover, many impact researchers rely on accessible data aggregated to administrative units, often weighted by population, rather than gridded data, as these align more directly with policy- and decision-making. Based on our projections, we also provide user-ready (population-weighted) spatially aggregated climate variables at administrative unit levels (0–2) for 104 countries prioritized for tropical disease research. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Background & Summary Global Climate Models (GCMs), such as those provided by the Coupled Model Intercomparison Project (CMIP) 1 , 2 , are highly complex tools used to simulate the Earth system and are widely applied in climate impact research. GCMs differ in their construction, which impacts the representation of physical processes, spatial resolution, climate sensitivity, and their intended scope for further applications. GCMs generally capture large-scale climate patterns, but, due to their relatively coarse spatial resolutions (about 1–3°), rely heavily on parameterizations, simplified representations of natural, often sub-grid-scale processes that are too complex or infeasible to simulate directly. Parameterizations, simplified physical and thermodynamic schemes, coarse spatial resolution, and incomplete representations of key climate system processes in GCMs are major sources of uncertainty in climate (change) modelling. These limitations lead to systematic model errors (biases) and to the poor, or even absent, representation of fine-scale features that are critical for accurate regional and local climate impact assessments 3 , 4 . Therefore, before GCM outputs can be effectively applied at local or regional scales, it is essential to correct model biases 5 – 7 , and to generate climate information at finer spatial resolutions than those provided by GCMs via appropriate bias-correction and downscaling (BC&D) methods 8 – 12 . Observational datasets serve as reference climatology, but variations in data sources, quality control, generation methods, and spatiotemporal resolution introduce an additional source of uncertainty in climate impact research 13 – 17 . Any deficiencies in the observational reference are often transferred to future climate projections. Hence, the choice of GCMs, reference observational datasets, and respective BC&D methods can lead to substantial differences in the fine-scale spatiotemporal patterns of the bias-corrected and downscaled climate projections. It is therefore critical to carefully consider these choices when selecting or generating robust high-resolution climate projections to support impact assessments in key sectors such as agriculture, water resources, and health. Downscaling approaches are typically categorized into two main types: dynamical and statistical. Dynamical downscaling uses outputs from GCMs as boundary conditions for driving limited-area, high-resolution Regional Climate Models (RCMs) to generate high-resolution climate projections. This approach is exemplified by the Coordinated Regional Climate Downscaling Experiment (CORDEX) 18 that was established and implemented to generate high-resolution dynamical downscaled outputs for the fifth phase of CMIP (CMIP5). However, RCMs are sensitive to the boundary conditions provided by their driving GCMs and remain subject to substantial errors, resulting in a strong dependence on the choice of GCMs, and regional biases that must be corrected for accurate climate impact assessments 6 , 19 , 20 . Moreover, dynamical downscaling is computationally expensive, requiring significant data storage and processing capacity, which limits the number of available models, simulation runs, and generated products 21 – 23 and are often limited in spatial extent 24 , 25 . For example, while CP4-Africa, a pan-African convection-permitting regional climate simulation, is one of the only initiatives that explicitly simulates convection without a respective parameterization and provides output at a higher spatial resolution than other regional climate downscaling exercises (4.5 km), it is currently based on a single driving model and emissions scenario (using an idealized future climate under Representative Concentration Pathway (RCP) 8.5), limited to the African continent, and offers only reduced temporal coverage based on two 10-year time periods 26 , 27 . In contrast, statistical downscaling uses hydrometeorological observations to adjust model biases and refine spatial resolution. Statistical BC&D methods range from simple approaches - such as delta change methods that correct only the mean of the variable of interest - to more detailed techniques that adjust biases across the full distribution of values, such as Quantile Delta Mapping (QDM) 6 , 28 . Quantile Mapping (QM) methods have been widely adopted in climate impact studies, especially at the global scale, due to their relatively low computational cost compared to alternative techniques 29 . A simple approach to generating downscaled climate information involves first simply interpolating coarse-resolution model outputs to the finer resolution of gridded observational datasets, followed by the application of a bias-correction method such as QDM independently to each grid cell. However, this cell-by-cell correction can disrupt the spatial covariance structure of climate variables, potentially leading to unrealistic spatial patterns 30 , 31 . To address these limitations, more sophisticated statistical downscaling methods have been developed, e.g., weather typing approaches like analogue-based methods 4 , 29 , 32 . Over recent decades, a variety of statistically downscaled climate projection datasets have been generated for a wide range of applications, utilizing a broad spectrum of BC&D approaches that span varying levels of methodological complexity, each with its own strengths, limitations, and scope of application. Recent climate impact assessments, e.g., those focused on vector-borne diseases (VBDs), have still often relied on projections generated using relatively simple BC&D methods 33 – 36 , such as linear scaling and delta change techniques, e.g., by utilizing the WorldClim dataset 37 , 38 . However, WorldClim’s methods for generating projections are effective primarily at conveying broad-scale climate change signals, and its outputs are limited to monthly climatological means averaged over 20-year periods, making it insufficient for modern impact studies which are increasingly focused on assessing changes in seasonality and extreme events. The Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) has also developed downscaled and bias-corrected climate projections from CMIP6 models specifically targeting at impact assessment across a variety of sectors and was already applied in recent VBD outbreak research 39 , 40 , but it has a relatively coarse spatial resolution (0.5°) 41 . The ISMIP bias-correction algorithm was also applied to dynamically downscaled output; for example, one version was used to produce global bias-corrected daily datasets derived from CORDEX at half-degree resolution 20 . Another widely used dataset is NASA’s Earth Exchange (NEX) Global Daily Downscaled Projections (GDDP) dataset 42 , 43 , employing a more advanced BC&D method known as Bias-Correction and Spatial Disaggregation (BCSD). However, despite its widespread adoption 44 , 45 , BCSD has notable limitations, as it corrects only the mean and variance, leading to inadequate preservation of trends especially in the distribution tails. Additionally, the dataset relies on the Global Meteorological Forcing Dataset 46 as its reference, a reanalysis product that is no longer maintained and has declined in widespread use 47 . A recent high-resolution, daily global dataset of downscaled CMIP6 models 48 , was developed using the statistical Bias-Correction Constructed Analogues with Quantile Mapping Reordering (BCCAQ) method, which demonstrates superior skill in representing the behaviour and statistical characteristics of hydrologic extremes, and is hence particularly well suited for hydrological assessments. Furthermore, Gergel, et al. 49 published a dataset for climate impact research based on a novel spatial BC&D approach to generate global downscaled climate projections designed to better preserve trends in the distribution tails. The method is highly sensitive to the choice of reference dataset, and ERA5 reanalysis was used despite its less reliable performance in tropical regions, particularly for precipitation estimates, for which it is primarily recommended for extratropical monitoring 50 . Hence, the practitioner’s dilemma is no longer primarily the lack of observational data sources or bias-corrected and downscaled climate projections, but rather how to select and evaluate an appropriate dataset for a specific application. In practice, products are often chosen based on availability, familiarity with the data provider, and, most importantly, convenience of format, as most data on weather, climate and climate change are typically provided in specialized formats (e.g., NetCDF) and often require substantial processing before they can be used for downstream impact analyses. On the other hand, resources especially including pre-processed GCM outputs that can be quickly and easily utilised by non-specialists remain very limited, and only a few sources exist at all that provide free, online global datasets pre-compiled to deliver spatially averaged area-level estimates of multiple climate variables across administrative units (e.g. countries, states). Most importantly, existing initiatives either focus exclusively on the observational period such as the Weighted Climate Dataset 51 or rely on datasets such as WorldClim for future climate information, as AREAdata 52 , which has the added limitation of providing spatial averages based on an observational data source different from that used to generate the underlying future projections. In conclusion, a respective fully consistent and comprehensive resource to support robust climate change impact studies, covering both observations and future climate change, is still lacking. Here, we present a dataset of six climate variables, outlined in Table 1 , developed to address the limitations of existing climate (projection) datasets and to complement previous work by achieving several key objectives. Our dataset comprises (1) high-resolution (0.1°) daily climate projections covering 12 land-only domains between 60°N and 60°S (see Supplementary Material Table 1 for details), and (2) corresponding spatially averaged (population-weighted) area-level estimates derived from both the observational reference datasets and the generated projections at national (admin 0) and subnational (1–2) administrative unit levels. We used outputs from two observational datasets and six CMIP6 models based on two scenarios from the Shared Socioeconomic Pathways (SSP)- RCP framework, SSP2-4.5 andSSP5-8.5 53−55 , with all data sources outlined in Table 2 . A primary motivation for developing this resource was to support local and regional health-related tropical climate impact research by providing a more reliable, tailored resource for these often-neglected and understudied regions. To achieve this, we first aimed to reduce observational uncertainty and enhance the reliability of the reference climatology, with a strong focus on tropical study domains, including the Global South - particularly for precipitation, which is highly spatiotemporally variable and therefore generally more difficult to predict. We therefore use ERA5-Land reanalysis for all but one of the six climate variables, and supplement precipitation with CHIRPS, a widely used dataset shown to outperform others especially in tropical settings 16 , 56 , 57 . Secondly, our dataset represents a substantial improvement in spatial resolution compared to previous (cited above) statistically downscaled (quasi-)global datasets, which typically provide resolutions of 0.25° or coarser, making our generated projections particularly well suited for driving high-resolution impact models of various kinds. Third, we selected the Double Bias-Correction Constructed Analogues (DBCCA) statistical downscaling method due to its demonstrated strength in capturing both spatial patterns and location-specific distributions. It outperformed other methods by passing the greatest number of ClimDEX index (established indices to measure temporal and spatial patterns of temperature and precipitation extremes in the context of climate change) validation tests 32 , hence making it suitable for a broad range of applications, including various health-related temperature- and precipitation-sensitive climate impact assessments. Most importantly, we provide an unified open-access resource of readily usable information based on spatially aggregated climate data - both standard and population-weighted - with quasi-global coverage for 104 countries (administrative unit levels 0–2). These countries are currently the primary focus of the Vaccine Impact Modelling Consortium (VIMC) 58 and generally represent major hotspots for a variety of tropical infectious and communicable diseases. For the first time, we hereby provide a robust and consistent resource in a convenient, widely-used format that combines observational datasets with bias-corrected and downscaled climate projections, enabling non-specialists unfamiliar with the typical climate data formats (e.g., NetCDF) and processing methods to access information on both observations and corresponding consistent projections without additional preparation or heavy processing. Table 1 Selected and subsequently bias-corrected and downscaled climate variables for the historical (1985–2014) and future (2015–2100) periods. Final outputs are presented on a daily timescale and at a spatial resolution of 0.1°. Acronyms, units, and the upper and lower bounds used for data value validation are additionally provided, following Thrasher, et al. 43 . Variables Acronym Units Lower Bound Upper Bound Total precipitation pr mm 0 1040 Near-surface (2m) air temperature tas °C -73 67 Maximum near-surface (2m) air temperature tasmax °C -73 67 Minimum near-surface (2m) air temperature tasmin °C -73 67 Near-surface relative humidity hurs % 0 100 Near-surface specific humidity huss g/kg 0 40 Table 2 The two selected observational reference datasets, as well as the six selected Global Climate Models (GCMs) provided within the sixth phase of the Coupled Model Intercomparison Project (CMIP6). All datasets utilized in this analysis and presented in this manuscript are publicly available. The selected variables for both scenarios, SSP2-4.5 and SSP5-8.5, were available from all GCMs and were downloaded at their native resolution and on a daily temporal scale. The GCMs are based on different calendar types and provide varying native resolutions. The bias-corrected and downscaled output based on these data sources were standardized to a Gregorian calendar, and variables provided on a 0.1° regular latitude-longitude grid. Source Institute and Origin Version / Variant Label Original Calendar Type Native Resolution Main reference Observational Datasets (Reference Climatology) CHIRPS Climate Hazards Center (CHC) at UC Santa Barbara, USA 3.0 Gregorian (with leap years) 0.05° × 0.05° quasi-global (land-only, approximately 60°S-60°N) Funk, et al. 65 ERA5-Land European Centre for Medium-Range Weather Forecasts, Europe Accessed 2025 Gregorian (with leap years) 0.1° × 0.1° global (land-only) Muñoz-Sabater, et al. 62 Global Climate Models CanESM5 Canadian Centre for Climate Modelling and Analysis, Canada r1i1p1f1 365-day (no leap years) 500 km Swart, et al. 76 GFDL-ESM4 NOAA Geophysical Fluid Dynamics Laboratory, USA r1i1p1f1 365-day (no leap years) 100 km Dunne, et al. 77 MPI-ESM1-2-LR Max Planck Institute for Meteorology, Germany r1i1p1f1 Gregorian (with leap years) 250 km Mauritsen, et al. 78 MRI-ESM2-0 Meteorological Research Institute, Japan r1i1p1f1 Gregorian (with leap years) 100 km Yukimoto, et al. 79 TaiESM1 Research Centre for Environmental Changes, Taiwan r1i1p1f1 365-day (no leap years) 100 km Lee, et al. 80 UKESM1.0-LL UK Earth System Modelling project, UK r1i1p1f2* 360-day** 250 km Sellar, et al. 81 Note: * Since r1i1p1f1 was not available for download, the ensemble member r1i1p1f2 was used for UKESM1-0-LL ** UKESM1.0-LL uses a 360-day calendar consisting of 12 months, each with 30 days Methods Data acquisition. Historical and future climate model output were obtained from six GCMs participating in CMIP6. Table 2 summarizes the main characteristics of the six GCMs that were downscaled to create this dataset using the DBCCA method, alongside the observational datasets used as reference climatology. The data were accessed via WCRP’s distributed data archive developed and operated by the Earth System Grid Federation (ESGF), providing standardized, open access to a wide range of climate model outputs 59 . We assess two future (2015–2100) climate scenarios, based on the latest scenarios developed within the SSP-RCP framework as outlined in the IPCC sixth assessment report 1 , 60 . SSP2-4.5 represents a "middle-of-the-road" pathway with moderate challenges to both mitigation and adaptation, reflecting historical development trends. SSP5-8.5 describes a high-emissions, fossil fuel-driven future with high challenges to mitigation and low challenges to adaptation. For each GCM, we selected six variables, as shown in Table 1 , for their frequent use in climate impact assessments, particularly health-related tropical infectious disease modelling. For the reference climatology, we used ERA5-Land, provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) and obtained from the Copernicus Climate Change Service 61 . Reanalysis products like ERA5-Land use data assimilation schemes and corresponding models to merge observations with modelled forecasts, thereby producing consistent global gridded estimates across the historical period. ERA5-Land is based on rerunning the land component of ERA5, the fifth-generation ECMWF reanalysis and is limited to land surfaces 62 . Data for 2m air temperature (tas), 2m dew point temperature (dtas), and surface pressure (ps) were downloaded from ERA5-Land at native spatial resolution and hourly time steps. Daily precipitation data were obtained from the CHIRPS (Climate Hazards Group InfraRed Precipitation with Station data) dataset (version 3.0, global_daily, p05) 63 . CHIRPS is a quasi-global, satellite-based product providing nearly 40 years of rainfall estimates by combining satellite imagery with in-situ gauge data and CHC’s climatology (CHPclim) 64 , 65 . To generate spatially aggregated climate variables for administrative unit levels 0, 1 and 2, we used spatial boundary data from the Global Administrative Areas Database (GADM), version 4.1 66 . We use global, spatially downscaled time-series data of total population counts at 1-km resolution, including the base year 2000 and projections from 2010 to 2100, consistent with the SSPs, developed by the Socioeconomic Data and Applications Center 67 and obtained from the Havard Dataverse 68 . It is used to compute population-weighted spatial averages of all climate variables for all administrative unit levels and countries considered (full list provided in the Supplementary Material Table 2). For visualization, we clip all datasets to their land-only domain shapes by using the free vector and raster map data produced by Natural Earth based on their 10m cultural data with a resolution of 1:10m 69 . Data preparation. To derive daily information for all six variables shown in Table 1 for the observational reference, ERA5-Land hourly data were first aggregated to daily resolution following the convention of shifting accumulated variables backward by one hour to reflect their accumulation period, while retaining instantaneous variables at their original timestamps. To ensure the highest consistency, daily maximum (tasmax) and minimum 2m air temperature (tasmin) values were derived from the hourly tas time series. Building on the physical processes documentation of European Centre for Medium-Range Weather Forecasts (ECMWF) 70 , daily humidity-related variables were derived using Teten’s formula. Specifically, specific humidity (huss) was calculated based on dtas and ps, while relative humidity (hurs) was determined from dtas and tas, under the assumption of saturation over water for both calculations. Prior to applying the BC&D method, daily GCM outputs were regridded from their native resolutions to a regular 1° × 1° global latitude-longitude grid. CHIRPS data, originally at 0.05° resolution, were remapped to 0.1° to match ERA5-Land for the downscaling process. Additionally, both observational reference datasets were regridded to 1° resolution to allow bias-correction of the GCM outputs during the BC&D process. Generally, we applied bilinear regridding to continuous variables (e.g., temperature), while conservative remapping was used for variables like precipitation for its ability to conserve quantities, thereby not introducing or destroying water. However, it is important to note that all regridding methods can inevitably modify the statistical properties of the data - a trade-off that is unavoidable in order to achieve consistency across GCMs, as well as between GCMs and observational reference datasets. We also harmonized the datasets with respect to model grid types, grid orientation (e.g., north-south), coordinate reference systems (EPSG:4326), calendars, and standardized variable names and units (as depicted in Table 1 ). All datasets and model outputs, including ERA5-Land and CHIRPS, were converted to a consistent non-leap year calendar with 365 days per year before BC&D. For UKESM1-0-LL, which uses a 360-day calendar, to convert this model output to a 365-day calendar, we first mapped the data onto a regular non-leap year framework. Then, similar to the approach used in the LOCA statistically downscaled CMIP6 climate projections for North America 71 , the missing days per year were introduced by randomly selecting one day from evenly spaced temporal blocks across the year to interpolate the additional days (seven days in total). This approach minimizes artifacts in the annual cycle by distributing the interpolated days uniformly, thereby reducing their influence on statistical analyses. We select, clip and process the data based on 12 land-only domains. This definition follows the CORDEX framework (excluding Antarctica and the Arctic), which defines regional domains chosen to represent specific geographical regions, capture particular climate features, and often align with political or continental boundaries 72 . Details and boundaries of our defined 12 domains are provided in Supplementary Table 1, covering land-only outputs across our quasi-global extent (60°N − 60°S), tailored to and reflecting our focus on tropical areas as well as ensuring alignment with the CHIRPS observational domain. Bias-correction and downscaling process. We applied the DBCCA method 32 to bias-correct and downscale historical and future GCM outputs. This involved an initial application of the Bias-Corrected Constructed Analogues (BCCA) method, followed by a second-stage quantile mapping bias-correction. We use de-trended quantile mapping (DQM) in the first bias-correction step and quantile delta mapping (QDM) in the second, both with 50 quantiles. Bias-adjustment types were variable-dependent: additive for continuous variables (e.g., temperature), multiplicative for bounded variables (e.g., precipitation). Each variable was bias-corrected separately and the data was grouped by the month of the year before applying the adjustments separately to each group. The Constructed Analogue (CA) component represents a type of weather-typing approach and is based on the idea of using past weather situations as analogues for future weather conditions. The CA method identified 30 observational candidates as most suitable analogue days selected from a ± 45-day window around each target day based on climatological periods (30 years) to maintain seasonal alignment, with analogue similarity assessed via root mean squared error (RMSE), based on the observational data and GCM output regridded on a regular 1° × 1° global latitude-longitude grid. We applied least squares regression, specifically ridge regression, to compute analogue weights, which were used to linearly combine the respective high-resolution observational daily patterns to construct the downscaled fields. For precipitation, a square-root transformation was applied prior to analogue selection to ensure that the resulting downscaled values remained physically meaningful and did not include negative amounts. Furthermore, to address the well-known issue in GCMs - often referred to as the "drizzle bias problem", where models overestimate the frequency of low-intensity wet days - values below 1 mm/day (wet-day threshold) in GCMs were replaced with uniform random noise during the bias-correction and analogue construction steps, while after QDM respective values were explicitly set to 0 to better mirror the dry-day characteristics. Postprocessing. All datasets and models were converted to a standard Gregorian calendar via simple interpolation; leap day values (February 29) were generated by averaging data from February 28 and March 1 after the DBCCA procedure. Following Thrasher, et al. 43 , we also applied a quality control process to all downscaled output to check that values fell within a realistic range (Table 1 shows the bounding values used to evaluate each variable). We additionally checked if minimum temperature could exceed mean and maximum temperature after BC&D and we swapped the respective temperature values, if necessary. We prepared all data to meet the NetCDF Climate and Forecast (CF) Conventions. Any additional postprocessing and validation required for a particular application of these NetCDF data is left as an exercise for the end user. Spatial Aggregation. Based on the generated bias-corrected and downscaled climate projections, we derived spatially averaged and additionally population-weighted estimates for all variables at the country level (104 countries) and (where available from GADM) at administrative levels 1 and 2. A complete list of all countries is provided in the Supplementary Table 2, including information on the assigned domain used to generate the spatially averaged area-level estimates. If a country is fully embedded in more than one domain, the assignment to a specific domain is determined by a simple approximation: the distance between the centre of the country and the centre of each CORDEX domain is calculated, and the country is assigned to the domain with the shortest distance. In order to generate population-weighted estimates, the selected population data are first regridded to the same 0.1° × 0.1° global latitude-longitude grid to match the climate data. Gridded climate projections are then aggregated across regions defined by the chosen geographic boundaries using three approaches: (1) population-unweighted, (2) static population-weighted using the baseline population to still isolate the raw climate change signal, and (3) dynamically population-weighted, updated every ten years based on the previous population projections (e.g., 2035 values weighted by the 2030 population, 2048 values by the 2040 population). For each combination of GCM, scenario, and variable, three corresponding columns are hence produced: simple , static , and dynamic . This procedure is repeated for the selected observational reference datasets (using for the dynamic approach population estimates for 2000 for all years prior to 2000 and SSP2 for estimates in 2010 and 2020), producing a comprehensive, unified set of weather and climate data at the desired level of spatial granularity for easy integration into subsequent impact models. Data Records The complete dataset is available on Box and can be accessed by anyone from anywhere via a public link: https://imperialcollegelondon.box.com/s/y4f0ywor4gvcrdxkewlzerfxw4lik4ts . The downscaled (0.1°) daily data, derived from six GCMs using two observational datasets as reference climatology (Table 2 ), are publicly available for each domain. The data consist of all six climatological variables (Table 1 ) under two scenarios (SSP2-4.5 and SSP5-8.5), covering both the historical period (1981–2014) and the future period (2015–2100). The data are provided as compressed NetCDF files, with a total size of ~ 18.5 TB. File naming conventions for all variables and scenarios follow the following format for the compressed NetCDF files: domain _ variable _DBCCA_ model _ start _ end _ index _ experiment _compressed.nc Here, domain refers to the name of the CORDEX domain, such as South_America. variable indicates the downscaled variable, for example pr or tas and model specifies the downscaled GCM, for instance CanESM5. start and end denote the time period covered, with 1985 to 2014 for historical runs and 2015 to 2100 for future scenarios. index identifies the realization, such as r1i1p1f1, and experiment indicates the type of simulation, either historical or a future SSP scenario like SSP2-45 (“ssp245”). For further metadata, please refer to the information provided in each NetCDF file. The spatially aggregated datasets in Parquet format, derived for each of the six climate variables across all 104 countries and at administrative levels 0, 1, and 2, are made publicly accessible alongside the NetCDF data, with a total volume of approximately 350 GB. Each Parquet file corresponds to a specific administrative unit level and contains bias-corrected and downscaled climate projections for all six variables per CMIP6 GCM and scenario. The main values per variable are based on a standard area-level spatial aggregation ( simple ) procedure, and two additional columns provide population-weighted values ( static and dynamic ), as described in the previous section, leading to 18 columns per file in total. The files follow the same naming convention as the NETCDF files above, with additional information on the administrative unit included, but without the variable specification: domain _ country _v410_ adminunit _DBCCA_ model _ start _ end _ index _ experiment .parquet.gzip Additionally, we provide respective information derived from the observational reference datasets, ERA5-Land and CHIRPS, for their overlapping time period from 1981 to 2024, respectively. These files are also stored in Parquet format and named accordingly: domain _ country_ v410 _adminunit _ CHIRPSv3_ERA5Land_ start _ end _observation.parquet.gzip All spatially averaged files follow the official GADM administrative unit identifiers (GIDs) 66 -based naming structure, allowing direct alignment with the corresponding GADM (version 4.10) administrative boundaries and facilitating reproducible linkage to the external dataset. The corresponding naming and identifier of each admin unit was read directly from the GADM geopackage used in the spatial averaging process, allowing the two datasets to be linked unambiguously. The GID begins with the three-letter ISO 3166-1 alpha-3 country code. If subdivisions exist, they are identified by numbers from 1 to n, where n is the number of subdivisions at the respective administrative unit level. Numeric codes are assigned within each higher-level subdivision and are concatenated with the identifier of the preceding level, originally using a dot as the delimiter, but here all periods (“.”) are replaced by underscores (“_”) to avoid processing errors in some standard GIS systems. Each GID also includes a version suffix, appended after an underscore, which is retained in the file names. For each geographical and administrative unit, we store the number of pixels used in the spatial aggregation as metadata within the Parquet files, encoded through attributes. Additionally, descriptive information is provided for each variable (e.g., pr) alongside the respective aggregation procedure (e.g., simple). For example, for an admin-1 unit (identified with 4) in Bolivia, the following attributes are shown: The administrative unit is identified by admin_name = BOL_v410_4_1, with pixels = 449,448 indicating the number of grid cells used for spatial aggregation. The aggregation methodology is documented in the suffix attribute, which specifies standard area-level aggregation and population-weighting conventions, while the variables attribute lists all included climate variables together with their precise naming and units. Technical Validation Here, we present the validation of the generated bias-corrected and downscaled climate projections for one example domain, South America, focusing on pr and tas. Complete results based on the evaluation of historical climate simulations against the reference climatology for all variables and domains are provided in the Supplementary Material. Here, we additionally present, as an illustrative example to assess physical credibility and to perform a sanity check of the sign and magnitude of the climate change signal of our generated dataset, a comparison between the bias-corrected and downscaled future projections and the raw model projections for each GCM across South America. Comparison of downscaled and GCM data against the climatological reference. To assess the quality of the bias-corrected and downscaled data several statistical and graphical methods were used. The high-resolution climate projections and the raw GCM outputs were evaluated against the climatological reference by drawing comparisons between model outputs and observations over the period 1985–2014, hence using 12-month climatological time series. Three metrics were employed: the Pearson correlation coefficient to assess the temporal correlation and thus the similarity of the climatological seasonal cycle shape; the bias to quantify systematic, directional differences and determine whether the raw and DBCCA-adjusted simulations systematically overestimate or underestimate precipitation or temperature; and the root mean square error (RMSE) to assess the overall magnitude of errors. In addition to generating high-resolution data, we hence assessed the performance of DBCCA in reducing biases and errors in GCM simulations, based on monthly climatologies. Figure 1 Temporal correlation (a), bias (b), and RMSE (c) of monthly climatological average temperature between ERA5-Land and raw GCMs (top row, labelled Raw) as well as ERA5-Land and bias-corrected and downscaled (BC&D) climate projections (bottom row, labelled BC&D) averaged over the period 1985–2014. Each column represents a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, MRI-ESM2-0, TaiESM1, and UKESM1.0-LL. Median values across the domain are reported below each panel, with the 5th and 95th percentiles shown in brackets. Here, we report annual-average values of all metrics derived from this analysis. Consequently, the bias should be interpreted as an indicator of overall annual over- or underestimation of precipitation, keeping in mind that positive and negative monthly biases may compensate when averaged. We hence recommend relying on RMSE when comparing raw GCMs with bias-corrected and downscaled climate projections to assess the typical magnitude of errors. However, averaging temporal correlation and RMSE over the year can also mask variations in seasonal performance, as models may reproduce some months well and others poorly. We therefore strongly encourage users to additionally evaluate model performance using climatological monthly means and daily climatologies before application of the data, especially when focusing on a specific region, in addition to the annual-average metrics presented here. Fig. 1-2 show a comparison between reference climatology and either the historical raw or DBCCA-adjusted GCM outputs, illustrating (a) temporal correlation, (b) bias and (c) RMSE between ERA5-Land (for tas) or CHIRPS (for pr) and the historical climate projections (first row), as well as the bias-corrected and downscaled data (second row), for monthly climatologies of tas and pr averaged over the period from 1985 to 2014. Median values across the domain are reported below each panel, with the 5th and 95th percentiles indicated in brackets. Please refer to Supplementary Table 3 for corresponding statistics (median, percentiles) for all variables across all domains. Fig. 3 shows a more detailed depiction of the spatial patterns of multi-year monthly climatological average biases in tas and pr for all bias-corrected and downscaled climate projections, relative to ERA5-Land (tas) or CHIRPS (pr). Figure 1 and 2 clearly show that the bias-corrected and downscaled data show overall higher correlations, as well as lower bias and RMSE, across all variables and GCMs, when compared to the raw historical climate projections (considering climatological averages over the period from 1985 to 2014). For temperature, the raw GCM outputs exhibit substantial annually averaged biases calculated based on climatological monthly means, ranging from − 9.23°C (minimum; CanESM5) to + 16.39°C (maximum; UKESM1.0-LL), and respective high RMSE values, ranging from 0.11°C (TaiESM1) to 16.39°C (UKESM1.0-LL). After applying bias-correction and downscaling using DBCCA, these errors are greatly reduced. The downscaled data shows biases ranging from − 0.02°C (TaiESM1/MRI-ESM2-0) to 0.10°C (UKESM1.0-LL), with RMSE values ranging from 0.00°C (all GCMs) to a maximum of 0.10°C (TaiESM1). Respective annually averaged temporal correlations assessed based on monthly climatologies of tas over the period from 1985 to 2014 (i.e., the representation of the climatological seasonal cycle) with observations over South America are substantially higher for the bias-corrected and downscaled data compared to the raw GCM outputs. Figure 3 provides a detailed depiction of the spatial patterns of the remaining annually averaged monthly climatological tas bias, showing that residual biases are generally more pronounced in southwestern South America, particularly in Andes-influenced regions for some models, such as GFDL-ESM4 and UKESM1.0-LL. Especially for precipitation, the raw GCM outputs exhibit substantial biases, ranging from − 315.25 mm to + 754.37 mm (both observed in CanESM5), and high RMSE values, ranging from 1.75 mm (UKESM1.0-LL) to 754.37 mm (UKESM1.0-LL). In comparison, the bias-corrected and downscaled data shows by far lower biases ranging from − 18.64 mm (TaiESM1) to 3.67 mm (UKESM1.0-LL), with RMSE values ranging from 0.23 mm (MPI-ESM1-2-LR) to 18.64 mm (TaiESM1). Respective temporal correlations over South America are substantially higher in the bias-corrected and downscaled data compared to the raw GCM outputs. Figure 3 also shows a tendency for increased residual annually averaged monthly climatological pr biases in coastal and mountain-influenced areas of South America, particularly along the mid-eastern coastal regions of the domain. Further details on the (annually averaged) statistics across all domains by considering monthly climatologies of tas and pr based on the period from 1985 to 2014 can be found in Supplementary Table 3. Overall, the raw GCM data show a large bias and RMSE as well as a weaker temporal correlation for all variables compared to the downscaled data across all domains for tas. For pr, some domains exhibit patterns similar to those observed for South America, e.g., SEA, with a pronounced increase in all evaluation metrics. However, although the range of biases (from the 5th to 95th percentiles) is substantially reduced after bias-correction and downscaling, the median of bias across many domains is often slightly larger in absolute terms in the DBCCA-adjusted data than in the raw GCMs, when considering annual averages derived from climatological monthly means. Nevertheless, the bias-corrected and downscaled data are climatologically more accurate than the raw GCM output in terms of temporal correlation (reflecting an improved representation of the climatological seasonal cycle shape) and most importantly, the RMSE, indicating a smaller overall error magnitude in the bias-corrected and downscaled climate projections in comparison to the raw GCM output (hence here, as mentioned, annually averaged climatological bias patterns should be mostly interpreted by their sign, showing whether the model overestimates or underestimates values over the climatological year as a whole). Consequently, we conclude that the comparison between the downscaled and GCMs clearly shows the advantage of the bias-correction and downscaling method in removing systematic errors from GCMs and developing high-resolution climate datasets to drive impact models. Daily climatologies and distributions of continent-level projections and observations. Initial model biases varied among GCMs and were generally inconsistent throughout the year, but discrepancies between the historical climate simulations and the reference climatology were substantially reduced after bias-correction and downscaling in South America, as shown in Figs. 4 and 5 for tas and pr, respectively. The figures compare spatially averaged historical raw GCM outputs (a) and DBCCA-adjusted projections (b) over South America against the observational reference (ERA5-Land for tas, CHIRPS for pr). Daily climatologies (left column) averaged over the period from 1985 to 2014 and respective probability density functions (PDFs) (right column) are shown for each GCM, in comparison to ERA5-Land / CHIRPS (with grey shading indicating the observed ± one standard deviation range for the observational reference). It becomes evident in Fig. 4 that residual biases after DBCCA adjustment were minimal for tas, with resulting model distributions closely matching observations, hence demonstrating strong alignment in climatological statistics over the period 1985–2014. Certain GCMs, such as MRI-ESM2-0 and UKESM1.0-LL, initially exhibited a warm bias during the austral summer and spring, and a cold bias in winter. There existed greater inter-model variability during the warm season. Biases also varied across models, with some, such as MPI-ESM1-2-LR, showing mean temperatures near the centre to lower end of the observed climatological range, while others, particularly during the hottest months, such as CanESM5, strongly exceed the upper end of the observed range. Comparing the PDFs based on the raw GCMs, some models exhibit slight bimodal tas distributions (e.g., MPI-ESM1-2-LR), whereas others show less pronounced bimodality (e.g., CanESM5), with clear deviations from the observational reference. The generally warm biases of certain GCMs, particularly CanESM5, UKESM1.0-LL, and MRI-ESM2-0, are clearly evident. The observed biases were substantially reduced following bias-correction and downscaling and the historical simulations adjusted using DBCCA align closely with the observed tas distributions. Regarding daily climatologies of pr shown in Fig. 5 , there is generally a range of initial model biases across the GCMs, with some models having mean values that fall within or below the low end of the observed climatological range for most of the year (e.g., MPI-ESM1-2-LR, CanESM5, TaiESM1), while others peak at or surpass the high end of the observed range for much of the year (e.g., MRI-ESM2-0). These variations in deviations are also evident, for example, when the historical model climatology indicates the lowest precipitation values - and thus the climatologically driest period - occurring earlier (e.g., MPI-ESM1-2-LR) or later (e.g., GFDL-ESM4) in the year compared to the observed climatology, depending on the model. Model biases throughout the year remain but are considerably reduced after bias-correction and downscaling. While the climate projections with respect to their climatological statistics still do not perfectly match observations, the deviations are smaller, and the mean of the projections overlap substantially with the one-standard-deviation ranges of the observational reference. Regarding the PDFs, the differences in the shape of the distributions are much larger than for tas, and the overall variability deviates clearly from the observational reference across models. The distributions from some GCMs, such as TaiESM1 and MPI-ESM1-2-LR, are sharply bimodal, whereas the observed precipitation distribution is less so. While the DBCCA-adjusted historical simulations align more closely with the reference climatology, stronger deviations remain compared to tas, with a slight shift of the distribution toward lower precipitation values across all models. Daily Climate Extremes. Furthermore, extreme indices are used to assess the performance of the bias-corrected and downscaled data in capturing extreme events, such as heavy precipitation and hot days. The indices are based on definitions from the Expert Team on Climate Change Detection and Indices (ETCCDI) 73 and generated by using xclim, an operational Python library that provides a framework for implementing and customizing a wide range of climate-related indicators. For impact applications, such as assessing climate change impacts in a health context, it is often important to evaluate the annual frequency of impact-relevant daily climate extremes as well as their timing throughout the year; therefore, it is critical to assess how well these characteristics are reproduced in bias-corrected and downscaled climate datasets relative to the reference climatology. We thus encourage end users to further validate our dataset according to the specific requirements of their application, using customized definitions of extremes relevant to, for example, specific health applications or disease-modelling exercises, and to conduct corresponding seasonality analyses. Here, we present the mean differences in the annual number of heavy precipitation days and hot days. We define heavy precipitation days as days with daily pr exceeding 10 mm. Hot days are defined as days with tas above 30°C, reflecting the general definition of summer, warm or hot days while also accounting for the geographical context of the presented domain and intended applications. The number of hot days is well reproduced by all bias-corrected and downscaled climate projections, as shown in Fig. 6 (a), presenting the average difference in the annual number of hot days between the GCMs and the observational reference data over the period 1985–2014. Overall, the DBCCA-adjusted data provide an accurate representation of hot days across South America, with differences generally below +/-1.5 days. Based on the observational reference data (ERA5-Land), the rounded mean number of hot days across the domain is 3.65 per year (median: 0), ranging from 0 to 125.57days. Similarly, Fig. 6 (b) shows the average difference in the annual number of heavy precipitation days between the bias-corrected and downscaled climate projections and the observational reference data for the period 1985–2014. Overall, the generated historical climate simulations capture heavy precipitation days across South America accurately, with differences typically within ± 2.5 days. According to the observational reference data (CHIRPS), the rounded mean number of heavy precipitation days across the domain is 55.8 year (median: 53.6), ranging from 0 to 319.57 days. Further details on the evaluation of extreme events across all domains can be found in the Supplementary Material. Overall, the bias-corrected and downscaled climate projections capture most effectively the climatological annual number of hot and heavy precipitation days as shown in the comparisons between the models and the observational reference datasets across each domain (depicted in Supplementary Material Table 4 across all domains). These comparisons demonstrate that the DBCCA algorithm generally produces climate projections that are well suited for analysing not only changes in mean climate conditions but also shifts in extreme events. Future Temperature and Precipitation Changes. Following the health checks and validation of the bias-corrected and downscaled outputs with respect to the agreement between historical climate projection data and observations, including extreme weather indicators, we here now evaluate the credibility, sign, and magnitude of the climate change signal across South America. In principle, projected changes from the bias-corrected and downscaled data should be physically credible, fall within the expected range of climate change effects, and generally be consistent with those from the raw, unadjusted GCM outputs, while naturally being modified by incorporating greater spatial detail and, in particular, reduced biases in the mean state of the models. Following the IPCC convention of traditionally using 20-year periods, we here compare the future climate of 2081–2100 with the historical period 1995–2014. Figure 7 shows the comparison of the climate change signal based on the multi-year average annual projected changes in tas per GCM by contrasting the raw [1° x 1° spatial resolution] and bias-corrected, downscaled CMIP6 models under both scenarios. The comparisons show that, in general, the sign of change is consistent, and the major spatial patterns of the climate change signal evident in the raw GCM outputs are reproduced in the bias-corrected and downscaled CMIP6 models. In flatter regions of the continent, the downscaled data in particular further refine the spatial structure by representing local variability more distinctly than the underlying coarse-resolution GCMs, while preserving the overall climate change patterns. However, while spatial details are generally more discernible, the magnitude of change is also clearly modified in some regions of the continent. This effect is particularly evident in topographically complex areas, such as the mountainous regions of the Andes. These regions are poorly represented in raw GCM outputs, especially in coarse-resolution models such as CanESM5, and are only partially captured in relatively higher‐resolution models such as GFDL-ESM4. In contrast, the bias-corrected and downscaled projections resolve these regions clearly across all GCMs; however, it also becomes evident that they also now comparably exhibit a reduced average annual warming signal. Similarly, Fig. 8 highlights that, for most of the continent, the climate change signal for multi-year average annual projected pr in both the sign and magnitude of change is reproduced, while the downscaled, bias-corrected data again clearly refines the spatial structure and provides greater spatial detail. This is particularly evident e.g., in the representation of strong multi-year average dried conditions in the northern regions of South America, which are depicted in varying regional locations but consistently cover large extents under both scenario conditions and across almost all GCMs (to a lesser extent in MRI-ESM2-0). However, in the Andes-dominated regions, particularly the northwestern coastal areas, the downscaled projections show a noticeable alteration or even absence of the climate change signal, especially in the MPI-ESM1-2-LR and UKESM1-0-LL models, under both scenario conditions. The bias-corrected and downscaled projections generally do not reproduce the strong projected wetting signal in parts of the coastal areas, with only a faint indication of wetter conditions appearing in a few scattered locations. We conclude that the DBCCA-adjusted projections are in general credible in terms of the physics, as well as the sign and magnitude of the climate change signal, being consistent with expected patterns across South America. We judge that the differences between raw GCMs and processed projections mainly reflect the correction of underlying model biases and the increased spatial detail introduced by the downscaling procedure. In some sparse locations, however, we cannot rule out that the bias-corrected and downscaled projections might miss or too strongly reduce the signal that appeared in the raw GCMs. We therefore encourage end users to carefully examine the magnitude and sign of change in our bias-corrected and downscaled climate projections for a particular study location or area across all domains, including South America, before applying the data in further analyses or impact studies. Conclusions and Limitations. In summary, our generated climate projections successfully reproduce observed historical conditions across all domains, as demonstrated by their representation of monthly and daily climatologies, as well as extreme events such as hot and heavy precipitation days, based on the climatological period from 1985 to 2014. Nevertheless, we want to point out that, although we used comprehensive, high-resolution observational datasets to downscale and validate the GCM outputs, these datasets may still introduce additional uncertainties into both historical and future climate projections, as errors and biases in the reference climatology can propagate through the bias-correction and downscaling processes. This was mitigated in part by relying on observational references that have been evaluated and found to perform well - specifically, CHIRPS for total precipitation - particularly in tropical regions, including the Global South 16 , 56 , 57 . Furthermore, the DBCCA algorithm, applied to the daily time series, prioritizes bias reduction across the whole shape of the distribution for each variable. As a result, a trade-off arises between improving the representation of extremes and preserving the mean response of the variable, which must be considered before applying the resulting data in a specific context. We also note the assumptions inherent to statistical downscaling methods, such as stationarity, alongside these uncertainties in the reference datasets. Despite these potential sources of uncertainty, we are confident that our bias-corrected and downscaled high-resolution climate dataset provide reliable inputs for global, regional, and local impact assessment studies, offering substantially improved accuracy compared to raw GCM outputs. Regarding the spatially aggregated climate variables, we want to point out that the specific country or administrative unit, together with the provided metadata on the number of pixels, should be carefully considered by end users, particularly for regions that rely on only a small number of pixels. In such cases, the reliability of the spatially aggregated estimates might be reduced due to the limited number of pixels, their location within the land-sea mask of the underlying datasets (e.g., for small and remote islands), and the effective resolution of each data source. The effective resolution - the smallest feature that can be realistically represented - is often substantially larger (typically by a factor ≥ 2) in model-based datasets than the nominal grid spacing, which is commonly expressed in terms of grid pixels. In this regard, we also want to highlight that some spatially averaged climate variables for certain administrative units - especially island or coastal areas - contain no data or only data for CHIRPS-derived variables (i.e., pr). This behaviour is expected and arises from differences in land-sea masks and spatial land coverage between datasets, with CHIRPS providing coverage in some coastal or island areas where ERA5-Land does not. Furthermore, spatial aggregation was performed by averaging values across all grid cells touching each administrative unit or country, consistent with common practice in climate science and accounting for the effective resolution of the underlying datasets. Consequently, when summaries across multiple administrative units are required, we recommend recomputing the spatial aggregation using the corresponding combined boundaries rather than relying on precomputed area-level estimates. Furthermore, while administrative unit identifiers (GIDs) were directly extracted from the official GADM geopackage (version 4.1), we noted inconsistencies in GID version suffixes in certain countries. Although GADM version 4.1 would nominally imply a uniform suffix of _1, several administrative units include alternative suffixes (e.g., _2) within the official geopackage. These discrepancies are inherent to the GADM data and are therefore reflected in our naming of the corresponding Parquet files. For the 104 countries, spatially averaged (population-weighted) estimates were generated for administrative levels 0, 1 and 2 where available, but GADM version 4.1 does not provide all levels for some countries (e.g., Belize), so only available levels are included (see Supplementary Material Table 2 for details). Concluding, our dataset, comprising quasi-global, daily, high-resolution, and spatially aggregated climate variables derived from downscaled CMIP6 models, offers a unified and consistent resource for assessing future climate change and variability, supports high-resolution impact assessment models, and provides accessible climate information at administrative unit levels - particularly benefiting researcher, non-specialists and stakeholders working in tropical regions, including those in the Global South. Usage Notes The compressed NetCDF files are written following the CF Conventions. NetCDF data can be analysed in standard programming languages like R (e.g., using packages like ncdf4, raster or terra) or in Python (e.g., using xarray or netcdf4). Additionally, we recommend using the Climate Data Operators (CDO) tool 74 for simple dataset operations and visualizations of (uncompressed) NetCDF data, including its Python interface, which acts as a wrapper for the CDO command-line binary. For less experienced users, we note that NetCDF data can also be read and used in GIS-oriented applications, such as the open-source QGIS platform. Please also refer to metadata.yaml provided alongside the NetCDF data for additional information. In addition to the pixel-based NetCDF dataset, we provide spatially aggregated and population-weighted data in an accessible, user-friendly format. These data, aggregated at national and sub-national levels (up to administrative unit level 2), can be easily selected and are customizable by e.g., the applied weighting schemes, temporal resolution, timeframe and administrative level, making them suitable for direct use in climate impact assessments. This flexibility enhances the replicability of impact studies, promotes transparent data practices, and facilitates seamless integration with other domain-specific datasets, such as those used in health-related impact research. Again, there are common packages available in standard programming languages such as R and Python for reading data in Parquet format (e.g., the pandas library can utilize pyarrow as a backend to efficiently read Parquet files in Python). Please also refer to readme.pdf provided alongside the data for additional information. In particular, after downloading the dataset, it is important to verify that the number of files in each folder matches the numbers specified in the readme.pdf. The dataset is therefore designed to support the growing climate impact research field, which increasingly encompasses disciplines such as epidemiology and public health. We hence anticipate that this resource will be particularly valuable for infectious disease modelers, epidemiologists, and other practitioners outside the traditional climate science community. When using the generated NetCDF data or the spatially aggregated information provided in the Parquet files, users should cite this publication. Users should also always acknowledge the underlying data sources used to construct this dataset, particularly ERA5-Land, CHIRPS, and the CMIP6 models, by citing their respective data records and associated publications. Declarations Code and Data Availability The DBCCA code used to downscale the CMIP6 GCMs follows the methodology described in Werner and Cannon 32 as well as in the University of Toronto Climate Downscaling Workflow Guidebook (UTCDW) 75 , with example code available at the GitHub repository https://github.com/mikemorris12/UTCDW_Guidebook (released under the MIT License), with the respective algorithm setup here adapted and refined as described to generate the bias-corrected and downscaled climate projections across all GCMs, variables and scenarios for all 12 domains. Python code for downloading and preparing the observational reference datasets, validating all input datasets, and aggregating both the reference data and the bias-corrected and downscaled climate projections is also available at https://github.com/SallyAJ/cmip6-ref-aggregation.git (released under the MIT License). The complete dataset is available on Box under : https://imperialcollegelondon.box.com/s/y4f0ywor4gvcrdxkewlzerfxw4lik4ts . Acknowledgements This work was carried out as part of the Vaccine Impact Modelling Consortium (www.vaccineimpact.org), but the views expressed are those of the authors and not necessarily those of the Consortium or its funders. The funders were given the opportunity to review this paper prior to publication, but the final decision on the content of the publication was taken by the authors. This work was primarily supported by the Wellcome Trust via the Vaccine Impact Modelling Consortium [Grant Number 226727_Z_22_Z], with additional support from the Bill & Melinda Gates Foundation [Grant Number INV-034281], previously (OPP1157270/INV-009125), and Gavi, the Vaccine Alliance. Under the grant conditions of the Foundation, a Creative Commons Attribution 4.0 Generic License has already been assigned to the Author Accepted Manuscript version that might arise from this submission. We also acknowledge funding provided by the Jameel Institute (supported by a philanthropic donation from Community Jameel) and from the MRC Centre for Global Infectious Disease Analysis (reference MR/X020258/1), funded by the UK Medical Research Council (MRC). The latter UK-funded award is carried out in the frame of the Global Health EDCTP3 Joint Undertaking. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors would also like to thank Chris De La Force for providing technical support related to high-performance computing and cluster management. Author contributions S.J. performed the data collection, wrangling, and processing, including the implementation of the bias-correction and downscaling algorithm on CMIP6 model outputs, as well as the spatial aggregation of climate variables at country and administrative unit levels. S.J. and C.W. conceived the research idea and analytical approach, supported and guided by discussions with I.D., P.W., R.T., and N.F.. W.H. and K.G. supported discussions on the logistics of computing, data storage, and distribution, while all authors contributed to various aspects of computing, storage, and data dissemination. S.J. developed and prepared the original draft, and all authors contributed to reviewing and editing the manuscript. N.F. supervised the research. Competing interests The authors declare that no competing interests exist. References Intergovernmental Panel on Climate Change (IPCC). Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. (Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2021). Eyring, V. et al. Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geosci Model Dev 9, 1937–1958 (2016). https://doi.org:10.5194/gmd-9-1937-2016 Vogel, E. et al. An evaluation framework for downscaling and bias correction in climate change impact studies. J. Hydrol. 622, 129693 (2023). https://doi.org:10.1016/j.jhydrol.2023.129693 Maraun, D. & Widmann, M. Statistical Downscaling and Bias Correction for Climate Research . (Cambridge University Press, 2017). Cannon, A. J., Piani, C. & Sippel, S. in Climate Extremes and Their Implications for Impact and Risk Assessment (eds Jana Sillmann, Sebastian Sippel, & Simone Russo) Ch. 5, 77–104 (Elsevier, 2020). Teutschbein, C. & Seibert, J. Bias correction of regional climate model simulations for hydrological climate-change impact studies: Review and evaluation of different methods. J. Hydrol. 456–457, 12–29 (2012). https://doi.org:10.1016/j.jhydrol.2012.05.052 Menapace, A. et al. Review of bias correction methods for climate model outputs in hydrology. J. Hydrol. 660, 133213 (2025). https://doi.org: https://doi.org/10.1016/j.jhydrol.2025.133213 Gebrechorkos, S. H., Hulsmann, S. & Bernhofer, C. Statistically downscaled climate dataset for East Africa. Sci. Data 6, 31 (2019). https://doi.org:10.1038/s41597-019-0038-1 Maraun, D. et al. Precipitation downscaling under climate change: Recent developments to bridge the gap between dynamical models and the end user. Rev. Geophys. 48, RG3003 (2010). https://doi.org:10.1029/2009rg000314 Keller, A. A., Garner, K. L., Rao, N., Knipping, E. & Thomas, J. Downscaling approaches of climate change projections for watershed modeling: Review of theoretical and practical considerations. PLOS Water 1, e0000046 (2022). https://doi.org:10.1371/journal.pwat.0000046 Chow, K. K. C., Sankaré, H., Diaconescu, E. P., Murdock, T. Q. & Cannon, A. J. Bias-adjusted and downscaled humidex projections for heat preparedness and adaptation in Canada. Geosci. Data J. 11, 680–698 (2024). https://doi.org:10.1002/gdj3.241 Jacobeit, J., Hertig, E., Seubert, S. & Lutz, K. Statistical downscaling for climate change projections in the Mediterranean region: methods and results. Reg Environ Change 14, 1891–1906 (2014). https://doi.org:10.1007/s10113-014-0605-0 Tarek, M., Brissette, F. & Arsenault, R. Uncertainty of gridded precipitation and temperature reference datasets in climate change impact studies. Hydrol. Earth Syst. Sci. 25, 3331–3350 (2021). https://doi.org:10.5194/hess-25-3331-2021 Tarek, M., Brissette, F. P. & Arsenault, R. Large-Scale Analysis of Global Gridded Precipitation and Temperature Datasets for Climate Change Impact Studies. J. Hydrometeorol. 21, 2623–2640 (2020). https://doi.org:10.1175/jhm-d-20-0100.1 Jahn, S., Fraser, K., Gaythorpe, K. A. M., Wainwright, C. M. & Ferguson, N. M. Evaluating the role of observational uncertainty in climate impact assessments: Temperature-driven yellow fever risk in South America. PLOS Climate 4, e0000601 (2025). https://doi.org:10.1371/journal.pclm.0000601 Jahn, S., Gaythorpe, K. A. M., Wainwright, C. M. & Ferguson, N. M. Evaluation of the Performance and Utility of Global Gridded Precipitation Products for Health Applications and Impact Assessments in South America. Geohealth 9, e2024GH001260 (2025). https://doi.org:10.1029/2024GH001260 Ludwig, R., Schmid, J. & Gampe, D. Impact of Reference Dataset Selection on RCM Evaluation, Bias Correction, and Resulting Climate Change Signals of Precipitation. J. Hydrometeorol. 20, 1813–1828 (2019). https://doi.org:10.1175/jhm-d-18-0108.1 Gutowski, W. J. et al. WCRP COordinated Regional Downscaling EXperiment (CORDEX): a diagnostic MIP for CMIP6. Geosci Model Dev 9, 4087–4095 (2016). https://doi.org:10.5194/gmd-9-4087-2016 Maraun, D. Bias Correcting Climate Change Simulations - a Critical Review. Current Climate Change Reports 2, 211–220 (2016). https://doi.org:10.1007/s40641-016-0050-x Yakubu, F., Bohner, J., Schickhoff, U., Scholten, T. & Hasson, S. U. Global Bias-Corrected CORDEX Datasets at Half Degree Resolution. Sci Data 12, 1781 (2025). https://doi.org:10.1038/s41597-025-06200-4 Prein, A. F. et al. A review on regional convection-permitting climate modeling: Demonstrations, prospects, and challenges. Rev. Geophys. 53, 323–361 (2015). https://doi.org:10.1002/2014RG000475 Giorgi, F. Thirty Years of Regional Climate Modeling: Where Are We and Where Are We Going next? J. Geophys. Res.: Atmos. 124, Pages 5696–5723 (2019). https://doi.org:10.1029/2018jd030094 Giorgi, F., Jones, C. & Asrar, G. R. Addressing climate information needs at the regional level: The CORDEX framework. WMO Bulletin 58, 175–183 (2009). Chapman, S. et al. Evaluation of Dynamically Downscaled CMIP6-CCAM Models Over Australia. Earth's Future 11, e2023EF003548 (2023). https://doi.org:10.1029/2023ef003548 Kendon, E. J. et al. Enhanced future changes in wet and dry extremes over Africa at convection-permitting scale. Nat. Commun. 10, 1794 (2019). https://doi.org:10.1038/s41467-019-09776-9 Senior, C. et al. Technical guidelines for using CP4-Africa simulation data , < https://doi.org/10.5281/zenodo.4316466%3E (2020). Vosper, S. B. et al. A Pan-African Convection-Permitting Regional Climate Simulation with the Met Office Unified Model: CP4-Africa. J. Clim. 31, 3485–3508 (2018). https://doi.org:10.1175/jcli-d-17-0503.1 Gutiérrez, J. M. et al. An intercomparison of a large ensemble of statistical downscaling methods over Europe: Results from the VALUE perfect predictor cross-validation experiment. Int. J. Climatol. 39, 3750–3785 (2018). https://doi.org:10.1002/joc.5462 Pierce, D. W., Cayan, D. R. & Thrasher, B. L. Statistical Downscaling Using Localized Constructed Analogs (LOCA). J. Hydrometeorol. 15, 2558–2585 (2014). https://doi.org:10.1175/jhm-d-14-0082.1 Maraun, D. Bias Correction, Quantile Mapping, and Downscaling: Revisiting the Inflation Issue. J. Clim. 26, 2137–2143 (2013). https://doi.org:10.1175/jcli-d-12-00821.1 Hnilica, J., Hanel, M. & Puš, V. Multisite bias correction of precipitation data from regional climate models. Int. J. Clim. 37, 2934–2946 (2016). https://doi.org:10.1002/joc.4890 Werner, A. T. & Cannon, A. J. Hydrologic extremes – an intercomparison of multiple gridded statistical downscaling methods. Hydrol. Earth Syst. Sci. 20, 1483–1508 (2016). https://doi.org:10.5194/hess-20-1483-2016 Gaythorpe, K. A., Hamlet, A., Cibrelus, L., Garske, T. & Ferguson, N. M. The effect of climate change on yellow fever disease burden in Africa. Elife 9, e55619 (2020). https://doi.org:10.7554/eLife.55619 Sargent, K., Mollard, J., Henley, S. F. & Bollasina, M. A. Predicting Transmission Suitability of Mosquito-Borne Diseases under Climate Change to Underpin Decision Making. Int. J. Environ. Res. Public Health 19, 13656 (2022). https://doi.org:10.3390/ijerph192013656 Ryan, S. J., Carlson, C. J., Mordecai, E. A. & Johnson, L. R. Global expansion and redistribution of Aedes-borne virus transmission risk with climate change. PLoS Negl. Trop. Dis. 13, e0007213 (2019). https://doi.org:10.1371/journal.pntd.0007213 Ryan, S. J. et al. Warming temperatures could expose more than 1.3 billion new people to Zika virus risk by 2050. Glob. Chan. Biol. 27, 84–93 (2021). https://doi.org:10.1111/gcb.15384 Fick, S. E. & Hijmans, R. J. WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Int. J. Climatol. 37, 4302–4315 (2017). https://doi.org:10.1002/joc.5086 Fick, S. E. & Hijmans, R. J. Downscaling future and past climate data from GCMs , %3Chttps://www.worldclim.org/data/downscaling.html%3E (2024). Farooq, Z. et al. Impact of climate and Aedes albopictus establishment on dengue and chikungunya outbreaks in Europe: a time-to-event analysis. Lancet Planet Health 9, e374-e383 (2025). https://doi.org:10.1016/S2542-5196(25)00059-2 Colon-Gonzalez, F. J. et al. Projecting the future incidence and burden of dengue in Southeast Asia. Nat Commun 14, 5439 (2023). https://doi.org:10.1038/s41467-023-41017-y Lange, S. Trend-preserving bias adjustment and statistical downscaling with ISIMIP3BASD (v1.0). Geosci Model Dev 12, 3055–3070 (2019). https://doi.org:10.5194/gmd-12-3055-2019 Thrasher, B., Maurer, E. P., McKellar, C. & Duffy, P. B. Technical Note: Bias correcting climate model simulated daily temperature extremes with quantile mapping. Hydrol. Earth Syst. Sci. 16, 3309–3314 (2012). https://doi.org:10.5194/hess-16-3309-2012 Thrasher, B. et al. NASA Global Daily Downscaled Projections, CMIP6. Sci. Data 9, 262 (2022). https://doi.org:10.1038/s41597-022-01393-4 Iwamura, T., Guzman-Holst, A. & Murray, K. A. Accelerating invasion potential of disease vector Aedes aegypti under climate change. Nat Commun 11, 2130 (2020). https://doi.org:10.1038/s41467-020-16010-4 Symons, T. L. et al. Projected impacts of climate change on malaria in Africa. Nature (2026). https://doi.org:10.1038/s41586-025-10015-z Sheffield, J., Goteti, G. & Wood, E. F. Development of a 50-year high-resolution global dataset of meteorological forcings for land surface modeling. J. Clim. 19, 3088–3111 (2006). https://doi.org: https://doi.org/10.1175/JCLI3790.1 Hassler, B. & Lauer, A. Comparison of Reanalysis and Observational Precipitation Datasets Including ERA5 and WFDE5. Atmosphere 12, 1462 (2021). https://doi.org:10.3390/atmos12111462 Gebrechorkos, S. et al. A high-resolution daily global dataset of statistically downscaled CMIP6 models for climate impact analyses. Sci. Data 10, 611 (2023). https://doi.org:10.1038/s41597-023-02528-x Gergel, D. R. et al. Global Downscaled Projections for Climate Impacts Research (GDPCIR): preserving quantile trends for modeling future climate impacts. Geosci Model Dev 17, 191–227 (2024). https://doi.org:10.5194/gmd-17-191-2024 Lavers, D. A., Simmons, A., Vamborg, F. & Rodwell, M. J. An evaluation of ERA5 precipitation for climate monitoring. Q. J. R. Meteorol. Soc. 148, 3152–3165 (2022). https://doi.org:10.1002/qj.4351 Gortan, M., Testa, L., Fagiolo, G. & Lamperti, F. A unified dataset for pre-processed climate indicators weighted by gridded economic activity. Sci. Data 11, 533 (2024). https://doi.org:10.1038/s41597-024-03304-1 Smith, T. P., Stemkovski, M., Koontz, A. & Pearse, W. D. AREAdata: A worldwide climate dataset averaged across spatial units at different scales through time. Data Brief 43, 108438 (2022). https://doi.org:10.1016/j.dib.2022.108438 O'Neill, B. C. et al. The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6. Geosci Model Dev 9, 3461–3482 (2016). https://doi.org:10.5194/gmd-9-3461-2016 O’Neill, B. C. et al. The roads ahead: Narratives for shared socioeconomic pathways describing world futures in the 21st century. Glob. Environ. Change 42, 169–180 (2017). https://doi.org:10.1016/j.gloenvcha.2015.01.004 Riahi, K. et al. The Shared Socioeconomic Pathways and their energy, land use, and greenhouse gas emissions implications: An overview. Glob. Environ. Change 42, 153–168 (2017). https://doi.org:10.1016/j.gloenvcha.2016.05.009 Gebrechorkos, S. H., Hülsmann, S. & Bernhofer, C. Evaluation of multiple climate data sources for managing environmental resources in East Africa. Hydrol. Earth Syst. Sci. 22, 4547–4564 (2018). https://doi.org:10.5194/hess-22-4547-2018 Beck, H. E. et al. Global-scale evaluation of 22 precipitation datasets using gauge observations and hydrological modeling. Hydrol. Earth Syst. Sci. 21, 6201–6217 (2017). https://doi.org:10.5194/hess-21-6201-2017 VIMC (Vaccine Impact Modelling Consortium). Vaccine Impact Modelling Consortium , < https://www.vaccineimpact.org/%3E (2024). Earth System Grid Federation (ESGF) User Support Working Team. ESGF User Support: User guide , %3Chttps://esgf.github.io/esgf-user-support/user_guide.html%3E (2019). Intergovernmental Panel on Climate Change (IPCC). Climate Change 2022: Impacts, Adaptation and Vulnerability. Working Group II contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. (Cambridge, UK and New York, NY, USA, 2022). Copernicus Climate Change Service (C3S). ERA5-Land hourly data from 1950 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS). %3Chttps://cds.climate.copernicus.eu/datasets/reanalysis-era5-land?tab=download%3E (2019). Muñoz-Sabater, J. et al. ERA5-Land: a state-of-the-art global reanalysis dataset for land applications. Earth Syst. Sci. Data 13, 4349–4383 (2021). https://doi.org:10.5194/essd-13-4349-2021 Climate Hazards Center - University of California Santa Barbara. Climate Hazards Center Infrared Precipitation with Stations version 3. CHIRPS3 Data Repository , %3Chttps://esgf.github.io/esgf-user-support/user_guide.html%3E (2025). Funk, C. C. et al. The climate hazards infrared precipitation with stations–a new environmental record for monitoring extremes. Sci. Data 2, 150066 (2015). https://doi.org:10.1038/sdata.2015.66 Funk, C. C. et al. A quasi-global precipitation time series for drought monitoring. Report No. 832, 12 (Reston, VA, 2014). Hijmans, R. et al. GADM database of Global Administrative Areas, version 4.1. , %3Chttps://gadm.org/%3E (2022). Gao, J. & National Center for Atmospheric Research (NCAR). Downscaling global spatial population projections from 1/8-degree to 1-km grid cells. NCAR Technical Note (2017). https://doi.org:10.5065/D60Z721H Gao, J. & Socioeconomic Data and Applications Center (SEDAC). Global 1-km Downscaled Population Grids, SSP-Consistent Projections and Base Year, v1.01 (2000–2100) (2020). https://doi.org: https://doi.org/10.7910/DVN/TLJ99B Natural Earth. Natural Earth - Free vector and raster map data , %3Chttps://www.naturalearthdata.com/about/%3E (2025). (ECMWF), E. C. f. M.-R. W. F. IFS documentation , %3Chttps://www.ecmwf.int/en/publications/ifs-documentation%3E (2025). Pierce, D. W. LOCA Statistical Downscaling (Localized Constructed Analogs) , %3Chttps://loca.ucsd.edu/loca-calendar/%3E (2025). World Climate Research Programme (WCRP). CORDEX domain description - Download domain description (update 23/10/15) , %3Chttps://cordex.org/domains/cordex-domain-description/%3E (2025). Expert Team on Climate Change Detection and Indices (ETCCDI). ETCCDI Climate Change Indices , < https://etccdi.pacificclimate.org/index.shtml%3E (2020). Schulzweida, U. & Max Planck Institute (MPI) for Meteorology. CDO (Climate Data Operators) Version 2.5.2 - Documentation: CDO User Guide , < https://code.mpimet.mpg.de/projects/cdo/wiki/Cdo#Documentation%3E (2025). Morris, M.,;, Kushner, P. J. & Smith, K. L. The University of Toronto Climate Downscaling Workflow: Tools and Resources for Climate Change Impact Analysis. J. open source educ. 7, 243 (2024). https://doi.org:10.21105/jose.00243 Swart, N. C. et al. The Canadian Earth System Model version 5 (CanESM5.0.3). Geosci Model Dev 12, 4823–4873 (2019). https://doi.org:10.5194/gmd-12-4823-2019 Dunne, J. P. et al. The GFDL Earth System Model Version 4.1 (GFDL-ESM 4.1): Overall Coupled Model Description and Simulation Characteristics. J. Adv. Model. Earth Syst. 12 (2020). https://doi.org:10.1029/2019ms002015 Mauritsen, T. et al. Developments in the MPI-M Earth System Model version 1.2 (MPI-ESM1.2) and Its Response to Increasing CO(2). J. Adv. Model Earth. Syst. 11, 998–1038 (2019). https://doi.org:10.1029/2018MS001400 Yukimoto, S. et al. The Meteorological Research Institute Earth System Model Version 2.0, MRI-ESM2.0: Description and Basic Evaluation of the Physical Component. J. Meteor. Soc. Japan. Ser. II 97, 931–965 (2019). https://doi.org:10.2151/jmsj.2019-051 Lee, W.-L. et al. Taiwan Earth System Model Version 1: description and evaluation of mean state. Geosci Model Dev 13, 3887–3904 (2020). https://doi.org:10.5194/gmd-13-3887-2020 Sellar, A. A. et al. UKESM1: Description and Evaluation of the U.K. Earth System Model. J. Adv. Model. Earth Syst. 11, 4513–4558 (2019). https://doi.org:10.1029/2019ms001739 Additional Declarations No competing interests reported. Supplementary Files SciDataSupplement.pdf Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9508034","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"data-descriptor","associatedPublications":[],"authors":[{"id":628448791,"identity":"ba8ecc23-30da-40d1-aa02-0f18f5fdcf05","order_by":0,"name":"Sally Jahn","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA70lEQVRIiWNgGAWjYHCCBGYGBhseCJPBAkwzNhDWkgbTIkGUFgaglsMwNhFa5BsYHn4u3HFeRnd287EPD2okGPjbD7BJzsCjxeAAQ7L0zDO3eczuHEuekXBMgkHiTAKb5AZ8WoAekOZtA2q5kWPMkMAGdNgNBjbJB/gdlvybt+0cUEv+Z4aEfxIM8oS0MBxgSAPacgBkCzNDYpsEgwFIC16HHWZIs+ZtSwb5xZghsU+Cx/BMYrMlPu/Lt/ck3+Zts7M3u938mPHHNxs5ueOHD97swecwZp4ECAMSIww8REQk+wEULaNgFIyCUTAKMAAAnDRGoHrKb3wAAAAASUVORK5CYII=","orcid":"","institution":"Imperial College London","correspondingAuthor":true,"prefix":"","firstName":"Sally","middleName":"","lastName":"Jahn","suffix":""},{"id":628448792,"identity":"4f729279-0d75-44d7-9769-e5e43a0c5ece","order_by":1,"name":"Katy A M Gaythorpe","email":"","orcid":"","institution":"Imperial College London","correspondingAuthor":false,"prefix":"","firstName":"Katy","middleName":"A M","lastName":"Gaythorpe","suffix":""},{"id":628448793,"identity":"90f36a3b-14b1-4573-92d6-b28053815ef0","order_by":2,"name":"Ilaria Dorigatti","email":"","orcid":"","institution":"Imperial College London","correspondingAuthor":false,"prefix":"","firstName":"Ilaria","middleName":"","lastName":"Dorigatti","suffix":""},{"id":628448794,"identity":"11b48bd0-48a9-4aa0-b1d0-c38ba3863160","order_by":3,"name":"Peter Winskill","email":"","orcid":"","institution":"Imperial College London","correspondingAuthor":false,"prefix":"","firstName":"Peter","middleName":"","lastName":"Winskill","suffix":""},{"id":628448795,"identity":"ccca9ba4-ba14-4bd3-bc9a-f85ada627482","order_by":4,"name":"Wes Hinsley","email":"","orcid":"","institution":"Imperial College London","correspondingAuthor":false,"prefix":"","firstName":"Wes","middleName":"","lastName":"Hinsley","suffix":""},{"id":628448796,"identity":"ee0ada79-78b0-48e8-bfae-7beecfa8d74f","order_by":5,"name":"Caroline M Wainwright","email":"","orcid":"","institution":"University of Leeds","correspondingAuthor":false,"prefix":"","firstName":"Caroline","middleName":"M","lastName":"Wainwright","suffix":""},{"id":628448797,"identity":"7c969696-36f4-4aca-a8c4-4be8b45577c1","order_by":6,"name":"Ralf Toumi","email":"","orcid":"","institution":"Imperial College London","correspondingAuthor":false,"prefix":"","firstName":"Ralf","middleName":"","lastName":"Toumi","suffix":""},{"id":628448798,"identity":"17f29c92-64b3-4c22-bee9-15281eaf6046","order_by":7,"name":"Neil M Ferguson","email":"","orcid":"","institution":"Imperial College London","correspondingAuthor":false,"prefix":"","firstName":"Neil","middleName":"M","lastName":"Ferguson","suffix":""}],"badges":[],"createdAt":"2026-04-23 14:38:41","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9508034/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9508034/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108181740,"identity":"feb5540b-e9e6-45af-84bd-4abb71ab27bb","added_by":"auto","created_at":"2026-04-30 08:58:52","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":5985121,"visible":true,"origin":"","legend":"\u003cp\u003eTemporal correlation (a), bias (b), and RMSE (c) of monthly climatological average temperature between ERA5-Land and raw GCMs (top row, labelled Raw) as well as ERA5-Land and bias-corrected and downscaled (BC\u0026amp;D) climate projections (bottom row, labelled BC\u0026amp;D) averaged over the period 1985-2014. Each column represents a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, MRI-ESM2-0, TaiESM1, and UKESM1.0-LL. Median values across the domain are reported below each panel, with the 5\u003csup\u003eth\u003c/sup\u003e and 95\u003csup\u003eth\u003c/sup\u003e percentiles shown in brackets.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/ecc1aca2f5ef761939090987.jpeg"},{"id":108076752,"identity":"2b13c4d0-792c-42c5-afe7-44d940675d9f","added_by":"auto","created_at":"2026-04-29 07:00:07","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":6117936,"visible":true,"origin":"","legend":"\u003cp\u003eTemporal correlation (a), bias (b), and RMSE (c) of monthly climatological average precipitation between CHIRPS and raw GCMs (top row, labelled Raw) as well as CHIRPS and bias-corrected and downscaled (BC\u0026amp;D) climate projections (bottom row, labelled BC\u0026amp;D) averaged over the period 1985-2014. Each column represents a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, MRI-ESM2-0, TaiESM1, and UKESM1.0-LL. Median values across the domain are reported below each panel, with the 5\u003csup\u003eth\u003c/sup\u003e and 95\u003csup\u003eth\u003c/sup\u003e percentiles shown in brackets.\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/aa31fe6f126d9946bd48b006.jpeg"},{"id":108181840,"identity":"3e8a1564-348c-46ef-b5fb-0cc87e499561","added_by":"auto","created_at":"2026-04-30 08:58:58","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":862580,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial patterns of multi-year monthly climatological average annual biases in temperature (tas) (a) and precipitation (pr) (b) for all bias-corrected and downscaled climate projections, relative to ERA5-Land (tas) or CHIRPS (pr), averaged over 1985-2014 across the South American study domain. Each column corresponds to a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, MRI-ESM2-0, TaiESM1, and UKESM1.0-LL.\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/6db90ef4aa77c42080bee7d7.jpeg"},{"id":108181678,"identity":"c98106cf-ad95-441c-b199-5c3f26f18cd5","added_by":"auto","created_at":"2026-04-30 08:58:49","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2142769,"visible":true,"origin":"","legend":"\u003cp\u003eDaily mean climatologies (left column) and probability density functions (PDFs) (right column) of multi-year 2m surface air temperature (tas) for historical raw (spatial resolution: 1°) GCM outputs (a) and bias-corrected, downscaled (spatial resolution: 0.1°) climate projections (b). Values are temporally averaged over 1985-2014 and spatially averaged across the entire South American domain. Each line represents a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, MRI-ESM2-0, TaiESM1, and UKESM1.0-LL. For reference, daily climatologies - shading in grey to indicate the observed ± one standard deviation range - and PDFs from observational data are also shown, using ERA5-Land for tas (spatially and temporally averaged at 0.1° resolution over the same historical period). To allow comparisons, all calendars exclude leap years.\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/1c976814ce2ad2db104d923a.jpeg"},{"id":108076755,"identity":"05b42e31-53d6-4554-89d6-d2a6677b1d95","added_by":"auto","created_at":"2026-04-29 07:00:07","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2394716,"visible":true,"origin":"","legend":"\u003cp\u003eDaily mean climatologies (left column) and probability density functions (PDFs) (right column) of multi-year total precipitation (pr) for historical raw (spatial resolution: 1°) GCM outputs (a) and bias-corrected, downscaled (spatial resolution: 0.1°) climate projections (b). Values are temporally averaged over 1985-2014 and spatially averaged across the entire South American domain. Each line represents a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, MRI-ESM2-0, TaiESM1, and UKESM1.0-LL. For reference, daily climatologies - shading in grey to indicate the observed ± one standard deviation range - and PDFs from observational data are also shown, using CHIRPS for pr (spatially and temporally averaged at 0.1° resolution over the same historical period). To allow comparisons, all calendars exclude leap years.\u003c/p\u003e","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/e7dbc855573576f27a92cb11.jpeg"},{"id":108182632,"identity":"ab3caa5b-9422-4bbf-b536-4131558b856a","added_by":"auto","created_at":"2026-04-30 08:59:28","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1233269,"visible":true,"origin":"","legend":"\u003cp\u003eThe difference in the average annual number of hot days [days/year] (a) and heavy precipitation days [days/year] (b), between historical bias-corrected and downscaled climate projections and observational reference data. Hot days were defined as days with daily 2m surface air temperature (tas, source: ERA5-Land) exceeding 30°C. Heavy precipitation days were defined as days with daily total precipitation (pr, source: CHIRPS) exceeding 10 mm. The data have a spatial resolution of 0.1° and were computed from daily values over the period 1985-2014. Each column shows a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, and MRI-ESM2-0, TaiESM1, UKESM1.0-LL. Pink and light green colours indicate overestimation and underestimation of warm days, respectively. Green and brown colours indicate overestimation and underestimation of heavy precipitation days, respectively.\u003c/p\u003e","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/59abf2a42a28dfeeb3b19d67.jpeg"},{"id":108076758,"identity":"12571fe5-9ce5-4259-8e65-2edfbf613a61","added_by":"auto","created_at":"2026-04-29 07:00:07","extension":"jpeg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":5303862,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial patterns of multi-year average annual projected changes in 2m surface air temperature (tas) [°C] from raw (top row, labelled Raw) and bias-corrected and downscaled (bottom row, labelled BC\u0026amp;D) CMIP6 models under SSP2-4.5 (a) and SSP5-8.5 (b) for 2081-2100 compared to 1995-2014 across the South American study domain. Each column corresponds to a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, MRI-ESM2-0, TaiESM1, and UKESM1.0-LL.\u003c/p\u003e","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/a2606728382809d50c3a20d2.jpeg"},{"id":108182027,"identity":"9dbbe248-383f-41b1-bb68-186382ad92f0","added_by":"auto","created_at":"2026-04-30 08:59:05","extension":"jpeg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":5346195,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial patterns of multi-year average annual projected changes in total precipitation (pr) [mm] from raw (top row, labelled Raw) and bias-corrected and downscaled (bottom row, labelled BC\u0026amp;D) CMIP6 models under SSP2-4.5 (a) and SSP5-8.5 (b) for 2081-2100 compared to 1985-2014 across the South American study domain. Each column corresponds to a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, MRI-ESM2-0, TaiESM1, and UKESM1.0-LL.\u003c/p\u003e","description":"","filename":"floatimage8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/45387e4216fc74ed0a93645b.jpeg"},{"id":108490885,"identity":"63b64578-a977-4d2c-8e0d-939796953ca8","added_by":"auto","created_at":"2026-05-05 09:49:29","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":29895710,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/b6444402-e187-4585-aae8-c9bedf2cae49.pdf"},{"id":108181873,"identity":"d5b505b3-58af-416b-b743-04b612daa9e4","added_by":"auto","created_at":"2026-04-30 08:58:58","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":640569,"visible":true,"origin":"","legend":"","description":"","filename":"SciDataSupplement.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9508034/v1/def4f1bca8c238aee371b0d7.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Quasi-global, land-only, high-resolution and spatially averaged climate variables from downscaled CMIP6 models for climate impact research","fulltext":[{"header":"Background \u0026 Summary","content":"\u003cp\u003eGlobal Climate Models (GCMs), such as those provided by the Coupled Model Intercomparison Project (CMIP)\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, are highly complex tools used to simulate the Earth system and are widely applied in climate impact research. GCMs differ in their construction, which impacts the representation of physical processes, spatial resolution, climate sensitivity, and their intended scope for further applications. GCMs generally capture large-scale climate patterns, but, due to their relatively coarse spatial resolutions (about 1\u0026ndash;3\u0026deg;), rely heavily on parameterizations, simplified representations of natural, often sub-grid-scale processes that are too complex or infeasible to simulate directly. Parameterizations, simplified physical and thermodynamic schemes, coarse spatial resolution, and incomplete representations of key climate system processes in GCMs are major sources of uncertainty in climate (change) modelling. These limitations lead to systematic model errors (biases) and to the poor, or even absent, representation of fine-scale features that are critical for accurate regional and local climate impact assessments\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. Therefore, before GCM outputs can be effectively applied at local or regional scales, it is essential to correct model biases\u003csup\u003e\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e, and to generate climate information at finer spatial resolutions than those provided by GCMs via appropriate bias-correction and downscaling (BC\u0026amp;D) methods\u003csup\u003e\u003cspan additionalcitationids=\"CR9 CR10 CR11\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Observational datasets serve as reference climatology, but variations in data sources, quality control, generation methods, and spatiotemporal resolution introduce an additional source of uncertainty in climate impact research\u003csup\u003e\u003cspan additionalcitationids=\"CR14 CR15 CR16\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Any deficiencies in the observational reference are often transferred to future climate projections. Hence, the choice of GCMs, reference observational datasets, and respective BC\u0026amp;D methods can lead to substantial differences in the fine-scale spatiotemporal patterns of the bias-corrected and downscaled climate projections. It is therefore critical to carefully consider these choices when selecting or generating robust high-resolution climate projections to support impact assessments in key sectors such as agriculture, water resources, and health.\u003c/p\u003e \u003cp\u003eDownscaling approaches are typically categorized into two main types: dynamical and statistical. Dynamical downscaling uses outputs from GCMs as boundary conditions for driving limited-area, high-resolution Regional Climate Models (RCMs) to generate high-resolution climate projections. This approach is exemplified by the Coordinated Regional Climate Downscaling Experiment (CORDEX)\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e that was established and implemented to generate high-resolution dynamical downscaled outputs for the fifth phase of CMIP (CMIP5). However, RCMs are sensitive to the boundary conditions provided by their driving GCMs and remain subject to substantial errors, resulting in a strong dependence on the choice of GCMs, and regional biases that must be corrected for accurate climate impact assessments\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Moreover, dynamical downscaling is computationally expensive, requiring significant data storage and processing capacity, which limits the number of available models, simulation runs, and generated products\u003csup\u003e\u003cspan additionalcitationids=\"CR22\" citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e and are often limited in spatial extent\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. For example, while CP4-Africa, a pan-African convection-permitting regional climate simulation, is one of the only initiatives that explicitly simulates convection without a respective parameterization and provides output at a higher spatial resolution than other regional climate downscaling exercises (4.5 km), it is currently based on a single driving model and emissions scenario (using an idealized future climate under Representative Concentration Pathway (RCP) 8.5), limited to the African continent, and offers only reduced temporal coverage based on two 10-year time periods\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. In contrast, statistical downscaling uses hydrometeorological observations to adjust model biases and refine spatial resolution. Statistical BC\u0026amp;D methods range from simple approaches - such as delta change methods that correct only the mean of the variable of interest - to more detailed techniques that adjust biases across the full distribution of values, such as Quantile Delta Mapping (QDM)\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. Quantile Mapping (QM) methods have been widely adopted in climate impact studies, especially at the global scale, due to their relatively low computational cost compared to alternative techniques\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. A simple approach to generating downscaled climate information involves first simply interpolating coarse-resolution model outputs to the finer resolution of gridded observational datasets, followed by the application of a bias-correction method such as QDM independently to each grid cell. However, this cell-by-cell correction can disrupt the spatial covariance structure of climate variables, potentially leading to unrealistic spatial patterns\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. To address these limitations, more sophisticated statistical downscaling methods have been developed, e.g., weather typing approaches like analogue-based methods\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eOver recent decades, a variety of statistically downscaled climate projection datasets have been generated for a wide range of applications, utilizing a broad spectrum of BC\u0026amp;D approaches that span varying levels of methodological complexity, each with its own strengths, limitations, and scope of application. Recent climate impact assessments, e.g., those focused on vector-borne diseases (VBDs), have still often relied on projections generated using relatively simple BC\u0026amp;D methods\u003csup\u003e\u003cspan additionalcitationids=\"CR34 CR35\" citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e, such as linear scaling and delta change techniques, e.g., by utilizing the WorldClim dataset\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e,\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. However, WorldClim\u0026rsquo;s methods for generating projections are effective primarily at conveying broad-scale climate change signals, and its outputs are limited to monthly climatological means averaged over 20-year periods, making it insufficient for modern impact studies which are increasingly focused on assessing changes in seasonality and extreme events. The Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) has also developed downscaled and bias-corrected climate projections from CMIP6 models specifically targeting at impact assessment across a variety of sectors and was already applied in recent VBD outbreak research\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e,\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e, but it has a relatively coarse spatial resolution (0.5\u0026deg;)\u003csup\u003e41\u003c/sup\u003e. The ISMIP bias-correction algorithm was also applied to dynamically downscaled output; for example, one version was used to produce global bias-corrected daily datasets derived from CORDEX at half-degree resolution\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Another widely used dataset is NASA\u0026rsquo;s Earth Exchange (NEX) Global Daily Downscaled Projections (GDDP) dataset\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e, employing a more advanced BC\u0026amp;D method known as Bias-Correction and Spatial Disaggregation (BCSD). However, despite its widespread adoption\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e,\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e, BCSD has notable limitations, as it corrects only the mean and variance, leading to inadequate preservation of trends especially in the distribution tails. Additionally, the dataset relies on the Global Meteorological Forcing Dataset\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e as its reference, a reanalysis product that is no longer maintained and has declined in widespread use\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. A recent high-resolution, daily global dataset of downscaled CMIP6 models\u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e, was developed using the statistical Bias-Correction Constructed Analogues with Quantile Mapping Reordering (BCCAQ) method, which demonstrates superior skill in representing the behaviour and statistical characteristics of hydrologic extremes, and is hence particularly well suited for hydrological assessments. Furthermore, Gergel, et al. \u003csup\u003e49\u003c/sup\u003e published a dataset for climate impact research based on a novel spatial BC\u0026amp;D approach to generate global downscaled climate projections designed to better preserve trends in the distribution tails. The method is highly sensitive to the choice of reference dataset, and ERA5 reanalysis was used despite its less reliable performance in tropical regions, particularly for precipitation estimates, for which it is primarily recommended for extratropical monitoring\u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eHence, the practitioner\u0026rsquo;s dilemma is no longer primarily the lack of observational data sources or bias-corrected and downscaled climate projections, but rather how to select and evaluate an appropriate dataset for a specific application. In practice, products are often chosen based on availability, familiarity with the data provider, and, most importantly, convenience of format, as most data on weather, climate and climate change are typically provided in specialized formats (e.g., NetCDF) and often require substantial processing before they can be used for downstream impact analyses. On the other hand, resources especially including pre-processed GCM outputs that can be quickly and easily utilised by non-specialists remain very limited, and only a few sources exist at all that provide free, online global datasets pre-compiled to deliver spatially averaged area-level estimates of multiple climate variables across administrative units (e.g. countries, states). Most importantly, existing initiatives either focus exclusively on the observational period such as the Weighted Climate Dataset\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e or rely on datasets such as WorldClim for future climate information, as AREAdata\u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e, which has the added limitation of providing spatial averages based on an observational data source different from that used to generate the underlying future projections. In conclusion, a respective fully consistent and comprehensive resource to support robust climate change impact studies, covering both observations and future climate change, is still lacking.\u003c/p\u003e \u003cp\u003eHere, we present a dataset of six climate variables, outlined in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, developed to address the limitations of existing climate (projection) datasets and to complement previous work by achieving several key objectives. Our dataset comprises (1) high-resolution (0.1\u0026deg;) daily climate projections covering 12 land-only domains between 60\u0026deg;N and 60\u0026deg;S (see Supplementary Material Table\u0026nbsp;1 for details), and (2) corresponding spatially averaged (population-weighted) area-level estimates derived from both the observational reference datasets and the generated projections at national (admin 0) and subnational (1\u0026ndash;2) administrative unit levels. We used outputs from two observational datasets and six CMIP6 models based on two scenarios from the Shared Socioeconomic Pathways (SSP)- RCP framework, SSP2-4.5 andSSP5-8.5\u003csup\u003e53\u0026minus;55\u003c/sup\u003e, with all data sources outlined in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. A primary motivation for developing this resource was to support local and regional health-related tropical climate impact research by providing a more reliable, tailored resource for these often-neglected and understudied regions. To achieve this, we first aimed to reduce observational uncertainty and enhance the reliability of the reference climatology, with a strong focus on tropical study domains, including the Global South - particularly for precipitation, which is highly spatiotemporally variable and therefore generally more difficult to predict. We therefore use ERA5-Land reanalysis for all but one of the six climate variables, and supplement precipitation with CHIRPS, a widely used dataset shown to outperform others especially in tropical settings\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e,\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e. Secondly, our dataset represents a substantial improvement in spatial resolution compared to previous (cited above) statistically downscaled (quasi-)global datasets, which typically provide resolutions of 0.25\u0026deg; or coarser, making our generated projections particularly well suited for driving high-resolution impact models of various kinds. Third, we selected the Double Bias-Correction Constructed Analogues (DBCCA) statistical downscaling method due to its demonstrated strength in capturing both spatial patterns and location-specific distributions. It outperformed other methods by passing the greatest number of ClimDEX index (established indices to measure temporal and spatial patterns of temperature and precipitation extremes in the context of climate change) validation tests\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e, hence making it suitable for a broad range of applications, including various health-related temperature- and precipitation-sensitive climate impact assessments. Most importantly, we provide an unified open-access resource of readily usable information based on spatially aggregated climate data - both standard and population-weighted - with quasi-global coverage for 104 countries (administrative unit levels 0\u0026ndash;2). These countries are currently the primary focus of the Vaccine Impact Modelling Consortium (VIMC)\u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e and generally represent major hotspots for a variety of tropical infectious and communicable diseases. For the first time, we hereby provide a robust and consistent resource in a convenient, widely-used format that combines observational datasets with bias-corrected and downscaled climate projections, enabling non-specialists unfamiliar with the typical climate data formats (e.g., NetCDF) and processing methods to access information on both observations and corresponding consistent projections without additional preparation or heavy processing.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cem\u003eSelected and subsequently bias-corrected and downscaled climate variables for the historical (1985\u0026ndash;2014) and future (2015\u0026ndash;2100) periods. Final outputs are presented on a daily timescale and at a spatial resolution of 0.1\u0026deg;. Acronyms, units, and the upper and lower bounds used for data value validation are additionally provided, following Thrasher, et al.\u003c/em\u003e \u003csup\u003e\u003cem\u003e43\u003c/em\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAcronym\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUnits\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLower Bound\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eUpper Bound\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal precipitation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003epr\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003emm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1040\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNear-surface (2m) air temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003etas\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026deg;C\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e67\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMaximum near-surface (2m) air temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003etasmax\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026deg;C\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e67\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMinimum near-surface (2m) air temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003etasmin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026deg;C\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e67\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNear-surface relative humidity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ehurs\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNear-surface specific humidity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ehuss\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eg/kg\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cem\u003eThe two selected observational reference datasets, as well as the six selected Global Climate Models (GCMs) provided within the sixth phase of the Coupled Model Intercomparison Project (CMIP6). All datasets utilized in this analysis and presented in this manuscript are publicly available. The selected variables for both scenarios, SSP2-4.5 and SSP5-8.5, were available from all GCMs and were downloaded at their native resolution and on a daily temporal scale. The GCMs are based on different calendar types and provide varying native resolutions. The bias-corrected and downscaled output based on these data sources were standardized to a Gregorian calendar, and variables provided on a 0.1\u0026deg; regular latitude-longitude grid.\u003c/em\u003e\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSource\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInstitute and Origin\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eVersion / Variant Label\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOriginal Calendar Type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNative Resolution\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMain reference\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"5\" nameend=\"c5\" namest=\"c1\"\u003e \u003cp\u003eObservational Datasets (Reference Climatology)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCHIRPS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eClimate Hazards Center (CHC) at UC Santa Barbara, USA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGregorian (with leap years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.05\u0026deg; \u0026times; 0.05\u0026deg; quasi-global (land-only, approximately 60\u0026deg;S-60\u0026deg;N)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFunk, et al. \u003csup\u003e65\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eERA5-Land\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEuropean Centre for Medium-Range Weather Forecasts, Europe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAccessed 2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGregorian (with leap years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.1\u0026deg; \u0026times; 0.1\u0026deg; global (land-only)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMu\u0026ntilde;oz-Sabater, et al. \u003csup\u003e62\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"5\" nameend=\"c5\" namest=\"c1\"\u003e \u003cp\u003eGlobal Climate Models\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCanESM5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCanadian Centre for Climate Modelling and Analysis, Canada\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003er1i1p1f1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e365-day (no leap years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e500 km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSwart, et al. \u003csup\u003e76\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGFDL-ESM4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNOAA Geophysical Fluid Dynamics Laboratory, USA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003er1i1p1f1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e365-day (no leap years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e100 km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eDunne, et al. \u003csup\u003e77\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMPI-ESM1-2-LR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMax Planck Institute for Meteorology, Germany\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003er1i1p1f1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGregorian (with leap years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e250 km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMauritsen, et al. \u003csup\u003e78\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMRI-ESM2-0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMeteorological Research Institute, Japan\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003er1i1p1f1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGregorian (with leap years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e100 km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYukimoto, et al. \u003csup\u003e79\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTaiESM1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eResearch Centre for Environmental Changes, Taiwan\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003er1i1p1f1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e365-day (no leap years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e100 km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLee, et al. \u003csup\u003e80\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUKESM1.0-LL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUK Earth System Modelling project, UK\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003er1i1p1f2*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e360-day**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e250 km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSellar, et al. \u003csup\u003e81\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNote:\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e* Since r1i1p1f1 was not available for download, the ensemble member r1i1p1f2 was used for UKESM1-0-LL\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e** UKESM1.0-LL uses a 360-day calendar consisting of 12 months, each with 30 days\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e \u003cb\u003eData acquisition.\u003c/b\u003e Historical and future climate model output were obtained from six GCMs participating in CMIP6. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e summarizes the main characteristics of the six GCMs that were downscaled to create this dataset using the DBCCA method, alongside the observational datasets used as reference climatology. The data were accessed via WCRP\u0026rsquo;s distributed data archive developed and operated by the Earth System Grid Federation (ESGF), providing standardized, open access to a wide range of climate model outputs\u003csup\u003e\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u003c/sup\u003e. We assess two future (2015\u0026ndash;2100) climate scenarios, based on the latest scenarios developed within the SSP-RCP framework as outlined in the IPCC sixth assessment report\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e\u003c/sup\u003e. SSP2-4.5 represents a \"middle-of-the-road\" pathway with moderate challenges to both mitigation and adaptation, reflecting historical development trends. SSP5-8.5 describes a high-emissions, fossil fuel-driven future with high challenges to mitigation and low challenges to adaptation. For each GCM, we selected six variables, as shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, for their frequent use in climate impact assessments, particularly health-related tropical infectious disease modelling.\u003c/p\u003e \u003cp\u003eFor the reference climatology, we used ERA5-Land, provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) and obtained from the Copernicus Climate Change Service\u003csup\u003e\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e. Reanalysis products like ERA5-Land use data assimilation schemes and corresponding models to merge observations with modelled forecasts, thereby producing consistent global gridded estimates across the historical period. ERA5-Land is based on rerunning the land component of ERA5, the fifth-generation ECMWF reanalysis and is limited to land surfaces\u003csup\u003e\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u003c/sup\u003e. Data for 2m air temperature (tas), 2m dew point temperature (dtas), and surface pressure (ps) were downloaded from ERA5-Land at native spatial resolution and hourly time steps. Daily precipitation data were obtained from the CHIRPS (Climate Hazards Group InfraRed Precipitation with Station data) dataset (version 3.0, global_daily, p05)\u003csup\u003e\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e\u003c/sup\u003e. CHIRPS is a quasi-global, satellite-based product providing nearly 40 years of rainfall estimates by combining satellite imagery with in-situ gauge data and CHC\u0026rsquo;s climatology (CHPclim)\u003csup\u003e\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e,\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eTo generate spatially aggregated climate variables for administrative unit levels 0, 1 and 2, we used spatial boundary data from the Global Administrative Areas Database (GADM), version 4.1\u003csup\u003e66\u003c/sup\u003e. We use global, spatially downscaled time-series data of total population counts at 1-km resolution, including the base year 2000 and projections from 2010 to 2100, consistent with the SSPs, developed by the Socioeconomic Data and Applications Center\u003csup\u003e\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e\u003c/sup\u003e and obtained from the Havard Dataverse\u003csup\u003e\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e\u003c/sup\u003e. It is used to compute population-weighted spatial averages of all climate variables for all administrative unit levels and countries considered (full list provided in the Supplementary Material Table\u0026nbsp;2). For visualization, we clip all datasets to their land-only domain shapes by using the free vector and raster map data produced by Natural Earth based on their 10m cultural data with a resolution of 1:10m\u003csup\u003e\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003cb\u003eData preparation.\u003c/b\u003e To derive daily information for all six variables shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e for the observational reference, ERA5-Land hourly data were first aggregated to daily resolution following the convention of shifting accumulated variables backward by one hour to reflect their accumulation period, while retaining instantaneous variables at their original timestamps. To ensure the highest consistency, daily maximum (tasmax) and minimum 2m air temperature (tasmin) values were derived from the hourly tas time series. Building on the physical processes documentation of European Centre for Medium-Range Weather Forecasts (ECMWF)\u003csup\u003e\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e\u003c/sup\u003e, daily humidity-related variables were derived using Teten\u0026rsquo;s formula. Specifically, specific humidity (huss) was calculated based on dtas and ps, while relative humidity (hurs) was determined from dtas and tas, under the assumption of saturation over water for both calculations.\u003c/p\u003e \u003cp\u003ePrior to applying the BC\u0026amp;D method, daily GCM outputs were regridded from their native resolutions to a regular 1\u0026deg; \u0026times; 1\u0026deg; global latitude-longitude grid. CHIRPS data, originally at 0.05\u0026deg; resolution, were remapped to 0.1\u0026deg; to match ERA5-Land for the downscaling process. Additionally, both observational reference datasets were regridded to 1\u0026deg; resolution to allow bias-correction of the GCM outputs during the BC\u0026amp;D process. Generally, we applied bilinear regridding to continuous variables (e.g., temperature), while conservative remapping was used for variables like precipitation for its ability to conserve quantities, thereby not introducing or destroying water. However, it is important to note that all regridding methods can inevitably modify the statistical properties of the data - a trade-off that is unavoidable in order to achieve consistency across GCMs, as well as between GCMs and observational reference datasets.\u003c/p\u003e \u003cp\u003eWe also harmonized the datasets with respect to model grid types, grid orientation (e.g., north-south), coordinate reference systems (EPSG:4326), calendars, and standardized variable names and units (as depicted in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). All datasets and model outputs, including ERA5-Land and CHIRPS, were converted to a consistent non-leap year calendar with 365 days per year before BC\u0026amp;D. For UKESM1-0-LL, which uses a 360-day calendar, to convert this model output to a 365-day calendar, we first mapped the data onto a regular non-leap year framework. Then, similar to the approach used in the LOCA statistically downscaled CMIP6 climate projections for North America\u003csup\u003e\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e\u003c/sup\u003e, the missing days per year were introduced by randomly selecting one day from evenly spaced temporal blocks across the year to interpolate the additional days (seven days in total). This approach minimizes artifacts in the annual cycle by distributing the interpolated days uniformly, thereby reducing their influence on statistical analyses.\u003c/p\u003e \u003cp\u003eWe select, clip and process the data based on 12 land-only domains. This definition follows the CORDEX framework (excluding Antarctica and the Arctic), which defines regional domains chosen to represent specific geographical regions, capture particular climate features, and often align with political or continental boundaries\u003csup\u003e\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e\u003c/sup\u003e. Details and boundaries of our defined 12 domains are provided in Supplementary Table\u0026nbsp;1, covering land-only outputs across our quasi-global extent (60\u0026deg;N\u0026thinsp;\u0026minus;\u0026thinsp;60\u0026deg;S), tailored to and reflecting our focus on tropical areas as well as ensuring alignment with the CHIRPS observational domain.\u003c/p\u003e \u003cp\u003e \u003cb\u003eBias-correction and downscaling process.\u003c/b\u003e We applied the DBCCA method\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e to bias-correct and downscale historical and future GCM outputs. This involved an initial application of the Bias-Corrected Constructed Analogues (BCCA) method, followed by a second-stage quantile mapping bias-correction. We use de-trended quantile mapping (DQM) in the first bias-correction step and quantile delta mapping (QDM) in the second, both with 50 quantiles. Bias-adjustment types were variable-dependent: additive for continuous variables (e.g., temperature), multiplicative for bounded variables (e.g., precipitation). Each variable was bias-corrected separately and the data was grouped by the month of the year before applying the adjustments separately to each group. The Constructed Analogue (CA) component represents a type of weather-typing approach and is based on the idea of using past weather situations as analogues for future weather conditions. The CA method identified 30 observational candidates as most suitable analogue days selected from a\u0026thinsp;\u0026plusmn;\u0026thinsp;45-day window around each target day based on climatological periods (30 years) to maintain seasonal alignment, with analogue similarity assessed via root mean squared error (RMSE), based on the observational data and GCM output regridded on a regular 1\u0026deg; \u0026times; 1\u0026deg; global latitude-longitude grid. We applied least squares regression, specifically ridge regression, to compute analogue weights, which were used to linearly combine the respective high-resolution observational daily patterns to construct the downscaled fields. For precipitation, a square-root transformation was applied prior to analogue selection to ensure that the resulting downscaled values remained physically meaningful and did not include negative amounts. Furthermore, to address the well-known issue in GCMs - often referred to as the \"drizzle bias problem\", where models overestimate the frequency of low-intensity wet days - values below 1 mm/day (wet-day threshold) in GCMs were replaced with uniform random noise during the bias-correction and analogue construction steps, while after QDM respective values were explicitly set to 0 to better mirror the dry-day characteristics.\u003c/p\u003e \u003cp\u003e \u003cb\u003ePostprocessing.\u003c/b\u003e All datasets and models were converted to a standard Gregorian calendar via simple interpolation; leap day values (February 29) were generated by averaging data from February 28 and March 1 after the DBCCA procedure. Following Thrasher, et al. \u003csup\u003e43\u003c/sup\u003e, we also applied a quality control process to all downscaled output to check that values fell within a realistic range (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows the bounding values used to evaluate each variable). We additionally checked if minimum temperature could exceed mean and maximum temperature after BC\u0026amp;D and we swapped the respective temperature values, if necessary. We prepared all data to meet the NetCDF Climate and Forecast (CF) Conventions. Any additional postprocessing and validation required for a particular application of these NetCDF data is left as an exercise for the end user.\u003c/p\u003e \u003cp\u003e \u003cb\u003eSpatial Aggregation.\u003c/b\u003e Based on the generated bias-corrected and downscaled climate projections, we derived spatially averaged and additionally population-weighted estimates for all variables at the country level (104 countries) and (where available from GADM) at administrative levels 1 and 2. A complete list of all countries is provided in the Supplementary Table\u0026nbsp;2, including information on the assigned domain used to generate the spatially averaged area-level estimates. If a country is fully embedded in more than one domain, the assignment to a specific domain is determined by a simple approximation: the distance between the centre of the country and the centre of each CORDEX domain is calculated, and the country is assigned to the domain with the shortest distance. In order to generate population-weighted estimates, the selected population data are first regridded to the same 0.1\u0026deg; \u0026times; 0.1\u0026deg; global latitude-longitude grid to match the climate data. Gridded climate projections are then aggregated across regions defined by the chosen geographic boundaries using three approaches: (1) population-unweighted, (2) static population-weighted using the baseline population to still isolate the raw climate change signal, and (3) dynamically population-weighted, updated every ten years based on the previous population projections (e.g., 2035 values weighted by the 2030 population, 2048 values by the 2040 population). For each combination of GCM, scenario, and variable, three corresponding columns are hence produced: \u003cem\u003esimple\u003c/em\u003e, \u003cem\u003estatic\u003c/em\u003e, and \u003cem\u003edynamic\u003c/em\u003e. This procedure is repeated for the selected observational reference datasets (using for the \u003cem\u003edynamic\u003c/em\u003e approach population estimates for 2000 for all years prior to 2000 and SSP2 for estimates in 2010 and 2020), producing a comprehensive, unified set of weather and climate data at the desired level of spatial granularity for easy integration into subsequent impact models.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eData Records\u003c/h2\u003e \u003cp\u003eThe complete dataset is available on Box and can be accessed by anyone from anywhere via a public link: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://imperialcollegelondon.box.com/s/y4f0ywor4gvcrdxkewlzerfxw4lik4ts\u003c/span\u003e\u003cspan address=\"https://imperialcollegelondon.box.com/s/y4f0ywor4gvcrdxkewlzerfxw4lik4ts\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The downscaled (0.1\u0026deg;) daily data, derived from six GCMs using two observational datasets as reference climatology (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), are publicly available for each domain. The data consist of all six climatological variables (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) under two scenarios (SSP2-4.5 and SSP5-8.5), covering both the historical period (1981\u0026ndash;2014) and the future period (2015\u0026ndash;2100). The data are provided as compressed NetCDF files, with a total size of ~\u0026thinsp;18.5 TB. File naming conventions for all variables and scenarios follow the following format for the compressed NetCDF files:\u003c/p\u003e \u003cp\u003e \u003cem\u003edomain\u003c/em\u003e_\u003cem\u003evariable\u003c/em\u003e_DBCCA_\u003cem\u003emodel\u003c/em\u003e_\u003cem\u003estart\u003c/em\u003e_\u003cem\u003eend\u003c/em\u003e_\u003cem\u003eindex\u003c/em\u003e_\u003cem\u003eexperiment\u003c/em\u003e_compressed.nc\u003c/p\u003e \u003cp\u003eHere, \u003cem\u003edomain\u003c/em\u003e refers to the name of the CORDEX domain, such as South_America. \u003cem\u003evariable\u003c/em\u003e indicates the downscaled variable, for example pr or tas and \u003cem\u003emodel\u003c/em\u003e specifies the downscaled GCM, for instance CanESM5. \u003cem\u003estart\u003c/em\u003e and \u003cem\u003eend\u003c/em\u003e denote the time period covered, with 1985 to 2014 for historical runs and 2015 to 2100 for future scenarios. \u003cem\u003eindex\u003c/em\u003e identifies the realization, such as r1i1p1f1, and \u003cem\u003eexperiment\u003c/em\u003e indicates the type of simulation, either historical or a future SSP scenario like SSP2-45 (\u0026ldquo;ssp245\u0026rdquo;). For further metadata, please refer to the information provided in each NetCDF file.\u003c/p\u003e \u003cp\u003eThe spatially aggregated datasets in Parquet format, derived for each of the six climate variables across all 104 countries and at administrative levels 0, 1, and 2, are made publicly accessible alongside the NetCDF data, with a total volume of approximately 350 GB. Each Parquet file corresponds to a specific administrative unit level and contains bias-corrected and downscaled climate projections for all six variables per CMIP6 GCM and scenario. The main values per variable are based on a standard area-level spatial aggregation (\u003cem\u003esimple\u003c/em\u003e) procedure, and two additional columns provide population-weighted values (\u003cem\u003estatic\u003c/em\u003e and \u003cem\u003edynamic\u003c/em\u003e), as described in the previous section, leading to 18 columns per file in total. The files follow the same naming convention as the NETCDF files above, with additional information on the administrative unit included, but without the variable specification:\u003c/p\u003e \u003cp\u003e \u003cem\u003edomain\u003c/em\u003e_\u003cem\u003ecountry\u003c/em\u003e_v410_\u003cem\u003eadminunit\u003c/em\u003e_DBCCA_\u003cem\u003emodel\u003c/em\u003e_\u003cem\u003estart\u003c/em\u003e_\u003cem\u003eend\u003c/em\u003e_\u003cem\u003eindex\u003c/em\u003e_\u003cem\u003eexperiment\u003c/em\u003e.parquet.gzip\u003c/p\u003e \u003cp\u003eAdditionally, we provide respective information derived from the observational reference datasets, ERA5-Land and CHIRPS, for their overlapping time period from 1981 to 2024, respectively. These files are also stored in Parquet format and named accordingly:\u003c/p\u003e \u003cp\u003e \u003cem\u003edomain\u003c/em\u003e_\u003cem\u003ecountry_\u003c/em\u003ev410\u003cem\u003e_adminunit\u003c/em\u003e_ CHIRPSv3_ERA5Land_\u003cem\u003estart\u003c/em\u003e_\u003cem\u003eend\u003c/em\u003e_observation.parquet.gzip\u003c/p\u003e \u003cp\u003eAll spatially averaged files follow the official GADM administrative unit identifiers (GIDs)\u003csup\u003e\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e\u003c/sup\u003e -based naming structure, allowing direct alignment with the corresponding GADM (version 4.10) administrative boundaries and facilitating reproducible linkage to the external dataset. The corresponding naming and identifier of each admin unit was read directly from the GADM geopackage used in the spatial averaging process, allowing the two datasets to be linked unambiguously. The GID begins with the three-letter ISO 3166-1 alpha-3 country code. If subdivisions exist, they are identified by numbers from 1 to n, where n is the number of subdivisions at the respective administrative unit level. Numeric codes are assigned within each higher-level subdivision and are concatenated with the identifier of the preceding level, originally using a dot as the delimiter, but here all periods (\u0026ldquo;.\u0026rdquo;) are replaced by underscores (\u0026ldquo;_\u0026rdquo;) to avoid processing errors in some standard GIS systems. Each GID also includes a version suffix, appended after an underscore, which is retained in the file names. For each geographical and administrative unit, we store the number of pixels used in the spatial aggregation as metadata within the Parquet files, encoded through attributes. Additionally, descriptive information is provided for each variable (e.g., pr) alongside the respective aggregation procedure (e.g., simple). For example, for an admin-1 unit (identified with 4) in Bolivia, the following attributes are shown: The administrative unit is identified by \u003cem\u003eadmin_name\u003c/em\u003e\u0026thinsp;=\u0026thinsp;BOL_v410_4_1, with \u003cem\u003epixels\u003c/em\u003e\u0026thinsp;=\u0026thinsp;449,448 indicating the number of grid cells used for spatial aggregation. The aggregation methodology is documented in the \u003cem\u003esuffix\u003c/em\u003e attribute, which specifies standard area-level aggregation and population-weighting conventions, while the \u003cem\u003evariables\u003c/em\u003e attribute lists all included climate variables together with their precise naming and units.\u003c/p\u003e \u003c/div\u003e"},{"header":"Technical Validation","content":"\u003cp\u003eHere, we present the validation of the generated bias-corrected and downscaled climate projections for one example domain, South America, focusing on pr and tas. Complete results based on the evaluation of historical climate simulations against the reference climatology for all variables and domains are provided in the Supplementary Material. Here, we additionally present, as an illustrative example to assess physical credibility and to perform a sanity check of the sign and magnitude of the climate change signal of our generated dataset, a comparison between the bias-corrected and downscaled future projections and the raw model projections for each GCM across South America.\u003c/p\u003e \u003cp\u003e \u003cb\u003eComparison of downscaled and GCM data against the climatological reference.\u003c/b\u003e To assess the quality of the bias-corrected and downscaled data several statistical and graphical methods were used. The high-resolution climate projections and the raw GCM outputs were evaluated against the climatological reference by drawing comparisons between model outputs and observations over the period 1985\u0026ndash;2014, hence using 12-month climatological time series. Three metrics were employed: the Pearson correlation coefficient to assess the temporal correlation and thus the similarity of the climatological seasonal cycle shape; the bias to quantify systematic, directional differences and determine whether the raw and DBCCA-adjusted simulations systematically overestimate or underestimate precipitation or temperature; and the root mean square error (RMSE) to assess the overall magnitude of errors. In addition to generating high-resolution data, we hence assessed the performance of DBCCA in reducing biases and errors in GCM simulations, based on monthly climatologies.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eFigure\u0026nbsp;1\u003c/b\u003e Temporal correlation (a), bias (b), and RMSE (c) of monthly climatological average temperature between ERA5-Land and raw GCMs (top row, labelled Raw) as well as ERA5-Land and bias-corrected and downscaled (BC\u0026amp;D) climate projections (bottom row, labelled BC\u0026amp;D) averaged over the period 1985\u0026ndash;2014. Each column represents a different GCM: CanESM5, GFDL-ESM4, MPI-ESM1-2-LR, MRI-ESM2-0, TaiESM1, and UKESM1.0-LL. Median values across the domain are reported below each panel, with the 5th and 95th percentiles shown in brackets.\u003c/p\u003e \u003cp\u003eHere, we report annual-average values of all metrics derived from this analysis. Consequently, the bias should be interpreted as an indicator of overall annual over- or underestimation of precipitation, keeping in mind that positive and negative monthly biases may compensate when averaged. We hence recommend relying on RMSE when comparing raw GCMs with bias-corrected and downscaled climate projections to assess the typical magnitude of errors. However, averaging temporal correlation and RMSE over the year can also mask variations in seasonal performance, as models may reproduce some months well and others poorly. We therefore strongly encourage users to additionally evaluate model performance using climatological monthly means and daily climatologies before application of the data, especially when focusing on a specific region, in addition to the annual-average metrics presented here.\u003c/p\u003e \u003cp\u003eFig. 1-2 show a comparison between reference climatology and either the historical raw or DBCCA-adjusted GCM outputs, illustrating (a) temporal correlation, (b) bias and (c) RMSE between ERA5-Land (for tas) or CHIRPS (for pr) and the historical climate projections (first row), as well as the bias-corrected and downscaled data (second row), for monthly climatologies of tas and pr averaged over the period from 1985 to 2014. Median values across the domain are reported below each panel, with the 5th and 95th percentiles indicated in brackets. Please refer to Supplementary Table 3 for corresponding statistics (median, percentiles) for all variables across all domains. Fig. 3 shows a more detailed depiction of the spatial patterns of multi-year monthly climatological average biases in tas and pr for all bias-corrected and downscaled climate projections, relative to ERA5-Land (tas) or CHIRPS (pr).\u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;1 and 2 clearly show that the bias-corrected and downscaled data show overall higher correlations, as well as lower bias and RMSE, across all variables and GCMs, when compared to the raw historical climate projections (considering climatological averages over the period from 1985 to 2014).\u003c/p\u003e \u003cp\u003eFor temperature, the raw GCM outputs exhibit substantial annually averaged biases calculated based on climatological monthly means, ranging from \u0026minus;\u0026thinsp;9.23\u0026deg;C (minimum; CanESM5) to +\u0026thinsp;16.39\u0026deg;C (maximum; UKESM1.0-LL), and respective high RMSE values, ranging from 0.11\u0026deg;C (TaiESM1) to 16.39\u0026deg;C (UKESM1.0-LL). After applying bias-correction and downscaling using DBCCA, these errors are greatly reduced. The downscaled data shows biases ranging from \u0026minus;\u0026thinsp;0.02\u0026deg;C (TaiESM1/MRI-ESM2-0) to 0.10\u0026deg;C (UKESM1.0-LL), with RMSE values ranging from 0.00\u0026deg;C (all GCMs) to a maximum of 0.10\u0026deg;C (TaiESM1). Respective annually averaged temporal correlations assessed based on monthly climatologies of tas over the period from 1985 to 2014 (i.e., the representation of the climatological seasonal cycle) with observations over South America are substantially higher for the bias-corrected and downscaled data compared to the raw GCM outputs. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e provides a detailed depiction of the spatial patterns of the remaining annually averaged monthly climatological tas bias, showing that residual biases are generally more pronounced in southwestern South America, particularly in Andes-influenced regions for some models, such as GFDL-ESM4 and UKESM1.0-LL.\u003c/p\u003e \u003cp\u003eEspecially for precipitation, the raw GCM outputs exhibit substantial biases, ranging from \u0026minus;\u0026thinsp;315.25 mm to +\u0026thinsp;754.37 mm (both observed in CanESM5), and high RMSE values, ranging from 1.75 mm (UKESM1.0-LL) to 754.37 mm (UKESM1.0-LL). In comparison, the bias-corrected and downscaled data shows by far lower biases ranging from \u0026minus;\u0026thinsp;18.64 mm (TaiESM1) to 3.67 mm (UKESM1.0-LL), with RMSE values ranging from 0.23 mm (MPI-ESM1-2-LR) to 18.64 mm (TaiESM1). Respective temporal correlations over South America are substantially higher in the bias-corrected and downscaled data compared to the raw GCM outputs. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e also shows a tendency for increased residual annually averaged monthly climatological pr biases in coastal and mountain-influenced areas of South America, particularly along the mid-eastern coastal regions of the domain.\u003c/p\u003e \u003cp\u003eFurther details on the (annually averaged) statistics across all domains by considering monthly climatologies of tas and pr based on the period from 1985 to 2014 can be found in Supplementary Table\u0026nbsp;3. Overall, the raw GCM data show a large bias and RMSE as well as a weaker temporal correlation for all variables compared to the downscaled data across all domains for tas. For pr, some domains exhibit patterns similar to those observed for South America, e.g., SEA, with a pronounced increase in all evaluation metrics. However, although the range of biases (from the 5th to 95th percentiles) is substantially reduced after bias-correction and downscaling, the median of bias across many domains is often slightly larger in absolute terms in the DBCCA-adjusted data than in the raw GCMs, when considering annual averages derived from climatological monthly means. Nevertheless, the bias-corrected and downscaled data are climatologically more accurate than the raw GCM output in terms of temporal correlation (reflecting an improved representation of the climatological seasonal cycle shape) and most importantly, the RMSE, indicating a smaller overall error magnitude in the bias-corrected and downscaled climate projections in comparison to the raw GCM output (hence here, as mentioned, annually averaged climatological bias patterns should be mostly interpreted by their sign, showing whether the model overestimates or underestimates values over the climatological year as a whole). Consequently, we conclude that the comparison between the downscaled and GCMs clearly shows the advantage of the bias-correction and downscaling method in removing systematic errors from GCMs and developing high-resolution climate datasets to drive impact models.\u003c/p\u003e \u003cp\u003e \u003cb\u003eDaily climatologies and distributions of continent-level projections and observations.\u003c/b\u003e Initial model biases varied among GCMs and were generally inconsistent throughout the year, but discrepancies between the historical climate simulations and the reference climatology were substantially reduced after bias-correction and downscaling in South America, as shown in Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e for tas and pr, respectively. The figures compare spatially averaged historical raw GCM outputs (a) and DBCCA-adjusted projections (b) over South America against the observational reference (ERA5-Land for tas, CHIRPS for pr). Daily climatologies (left column) averaged over the period from 1985 to 2014 and respective probability density functions (PDFs) (right column) are shown for each GCM, in comparison to ERA5-Land / CHIRPS (with grey shading indicating the observed\u0026thinsp;\u0026plusmn;\u0026thinsp;one standard deviation range for the observational reference).\u003c/p\u003e \u003cp\u003eIt becomes evident in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e that residual biases after DBCCA adjustment were minimal for tas, with resulting model distributions closely matching observations, hence demonstrating strong alignment in climatological statistics over the period 1985\u0026ndash;2014. Certain GCMs, such as MRI-ESM2-0 and UKESM1.0-LL, initially exhibited a warm bias during the austral summer and spring, and a cold bias in winter. There existed greater inter-model variability during the warm season. Biases also varied across models, with some, such as MPI-ESM1-2-LR, showing mean temperatures near the centre to lower end of the observed climatological range, while others, particularly during the hottest months, such as CanESM5, strongly exceed the upper end of the observed range. Comparing the PDFs based on the raw GCMs, some models exhibit slight bimodal tas distributions (e.g., MPI-ESM1-2-LR), whereas others show less pronounced bimodality (e.g., CanESM5), with clear deviations from the observational reference. The generally warm biases of certain GCMs, particularly CanESM5, UKESM1.0-LL, and MRI-ESM2-0, are clearly evident. The observed biases were substantially reduced following bias-correction and downscaling and the historical simulations adjusted using DBCCA align closely with the observed tas distributions.\u003c/p\u003e \u003cp\u003eRegarding daily climatologies of pr shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e, there is generally a range of initial model biases across the GCMs, with some models having mean values that fall within or below the low end of the observed climatological range for most of the year (e.g., MPI-ESM1-2-LR, CanESM5, TaiESM1), while others peak at or surpass the high end of the observed range for much of the year (e.g., MRI-ESM2-0). These variations in deviations are also evident, for example, when the historical model climatology indicates the lowest precipitation values - and thus the climatologically driest period - occurring earlier (e.g., MPI-ESM1-2-LR) or later (e.g., GFDL-ESM4) in the year compared to the observed climatology, depending on the model. Model biases throughout the year remain but are considerably reduced after bias-correction and downscaling. While the climate projections with respect to their climatological statistics still do not perfectly match observations, the deviations are smaller, and the mean of the projections overlap substantially with the one-standard-deviation ranges of the observational reference. Regarding the PDFs, the differences in the shape of the distributions are much larger than for tas, and the overall variability deviates clearly from the observational reference across models. The distributions from some GCMs, such as TaiESM1 and MPI-ESM1-2-LR, are sharply bimodal, whereas the observed precipitation distribution is less so. While the DBCCA-adjusted historical simulations align more closely with the reference climatology, stronger deviations remain compared to tas, with a slight shift of the distribution toward lower precipitation values across all models.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eDaily Climate Extremes.\u003c/b\u003e Furthermore, extreme indices are used to assess the performance of the bias-corrected and downscaled data in capturing extreme events, such as heavy precipitation and hot days. The indices are based on definitions from the Expert Team on Climate Change Detection and Indices (ETCCDI)\u003csup\u003e\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e\u003c/sup\u003e and generated by using xclim, an operational Python library that provides a framework for implementing and customizing a wide range of climate-related indicators. For impact applications, such as assessing climate change impacts in a health context, it is often important to evaluate the annual frequency of impact-relevant daily climate extremes as well as their timing throughout the year; therefore, it is critical to assess how well these characteristics are reproduced in bias-corrected and downscaled climate datasets relative to the reference climatology. We thus encourage end users to further validate our dataset according to the specific requirements of their application, using customized definitions of extremes relevant to, for example, specific health applications or disease-modelling exercises, and to conduct corresponding seasonality analyses. Here, we present the mean differences in the annual number of heavy precipitation days and hot days. We define heavy precipitation days as days with daily pr exceeding 10 mm. Hot days are defined as days with tas above 30\u0026deg;C, reflecting the general definition of summer, warm or hot days while also accounting for the geographical context of the presented domain and intended applications.\u003c/p\u003e \u003cp\u003eThe number of hot days is well reproduced by all bias-corrected and downscaled climate projections, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e6\u003c/span\u003e (a), presenting the average difference in the annual number of hot days between the GCMs and the observational reference data over the period 1985\u0026ndash;2014. Overall, the DBCCA-adjusted data provide an accurate representation of hot days across South America, with differences generally below +/-1.5 days. Based on the observational reference data (ERA5-Land), the rounded mean number of hot days across the domain is 3.65 per year (median: 0), ranging from 0 to 125.57days. Similarly, Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e6\u003c/span\u003e (b) shows the average difference in the annual number of heavy precipitation days between the bias-corrected and downscaled climate projections and the observational reference data for the period 1985\u0026ndash;2014. Overall, the generated historical climate simulations capture heavy precipitation days across South America accurately, with differences typically within \u0026plusmn;\u0026thinsp;2.5 days. According to the observational reference data (CHIRPS), the rounded mean number of heavy precipitation days across the domain is 55.8 year (median: 53.6), ranging from 0 to 319.57 days. Further details on the evaluation of extreme events across all domains can be found in the Supplementary Material. Overall, the bias-corrected and downscaled climate projections capture most effectively the climatological annual number of hot and heavy precipitation days as shown in the comparisons between the models and the observational reference datasets across each domain (depicted in Supplementary Material Table\u0026nbsp;4 across all domains). These comparisons demonstrate that the DBCCA algorithm generally produces climate projections that are well suited for analysing not only changes in mean climate conditions but also shifts in extreme events.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eFuture Temperature and Precipitation Changes.\u003c/b\u003e Following the health checks and validation of the bias-corrected and downscaled outputs with respect to the agreement between historical climate projection data and observations, including extreme weather indicators, we here now evaluate the credibility, sign, and magnitude of the climate change signal across South America. In principle, projected changes from the bias-corrected and downscaled data should be physically credible, fall within the expected range of climate change effects, and generally be consistent with those from the raw, unadjusted GCM outputs, while naturally being modified by incorporating greater spatial detail and, in particular, reduced biases in the mean state of the models. Following the IPCC convention of traditionally using 20-year periods, we here compare the future climate of 2081\u0026ndash;2100 with the historical period 1995\u0026ndash;2014.\u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e7\u003c/span\u003e shows the comparison of the climate change signal based on the multi-year average annual projected changes in tas per GCM by contrasting the raw [1\u0026deg; x 1\u0026deg; spatial resolution] and bias-corrected, downscaled CMIP6 models under both scenarios. The comparisons show that, in general, the sign of change is consistent, and the major spatial patterns of the climate change signal evident in the raw GCM outputs are reproduced in the bias-corrected and downscaled CMIP6 models. In flatter regions of the continent, the downscaled data in particular further refine the spatial structure by representing local variability more distinctly than the underlying coarse-resolution GCMs, while preserving the overall climate change patterns. However, while spatial details are generally more discernible, the magnitude of change is also clearly modified in some regions of the continent. This effect is particularly evident in topographically complex areas, such as the mountainous regions of the Andes. These regions are poorly represented in raw GCM outputs, especially in coarse-resolution models such as CanESM5, and are only partially captured in relatively higher‐resolution models such as GFDL-ESM4. In contrast, the bias-corrected and downscaled projections resolve these regions clearly across all GCMs; however, it also becomes evident that they also now comparably exhibit a reduced average annual warming signal.\u003c/p\u003e \u003cp\u003eSimilarly, Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e8\u003c/span\u003e highlights that, for most of the continent, the climate change signal for multi-year average annual projected pr in both the sign and magnitude of change is reproduced, while the downscaled, bias-corrected data again clearly refines the spatial structure and provides greater spatial detail. This is particularly evident e.g., in the representation of strong multi-year average dried conditions in the northern regions of South America, which are depicted in varying regional locations but consistently cover large extents under both scenario conditions and across almost all GCMs (to a lesser extent in MRI-ESM2-0). However, in the Andes-dominated regions, particularly the northwestern coastal areas, the downscaled projections show a noticeable alteration or even absence of the climate change signal, especially in the MPI-ESM1-2-LR and UKESM1-0-LL models, under both scenario conditions. The bias-corrected and downscaled projections generally do not reproduce the strong projected wetting signal in parts of the coastal areas, with only a faint indication of wetter conditions appearing in a few scattered locations.\u003c/p\u003e \u003cp\u003eWe conclude that the DBCCA-adjusted projections are in general credible in terms of the physics, as well as the sign and magnitude of the climate change signal, being consistent with expected patterns across South America. We judge that the differences between raw GCMs and processed projections mainly reflect the correction of underlying model biases and the increased spatial detail introduced by the downscaling procedure. In some sparse locations, however, we cannot rule out that the bias-corrected and downscaled projections might miss or too strongly reduce the signal that appeared in the raw GCMs. We therefore encourage end users to carefully examine the magnitude and sign of change in our bias-corrected and downscaled climate projections for a particular study location or area across all domains, including South America, before applying the data in further analyses or impact studies.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eConclusions and Limitations.\u003c/b\u003e In summary, our generated climate projections successfully reproduce observed historical conditions across all domains, as demonstrated by their representation of monthly and daily climatologies, as well as extreme events such as hot and heavy precipitation days, based on the climatological period from 1985 to 2014. Nevertheless, we want to point out that, although we used comprehensive, high-resolution observational datasets to downscale and validate the GCM outputs, these datasets may still introduce additional uncertainties into both historical and future climate projections, as errors and biases in the reference climatology can propagate through the bias-correction and downscaling processes. This was mitigated in part by relying on observational references that have been evaluated and found to perform well - specifically, CHIRPS for total precipitation - particularly in tropical regions, including the Global South\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e,\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e. Furthermore, the DBCCA algorithm, applied to the daily time series, prioritizes bias reduction across the whole shape of the distribution for each variable. As a result, a trade-off arises between improving the representation of extremes and preserving the mean response of the variable, which must be considered before applying the resulting data in a specific context. We also note the assumptions inherent to statistical downscaling methods, such as stationarity, alongside these uncertainties in the reference datasets. Despite these potential sources of uncertainty, we are confident that our bias-corrected and downscaled high-resolution climate dataset provide reliable inputs for global, regional, and local impact assessment studies, offering substantially improved accuracy compared to raw GCM outputs.\u003c/p\u003e \u003cp\u003eRegarding the spatially aggregated climate variables, we want to point out that the specific country or administrative unit, together with the provided metadata on the number of pixels, should be carefully considered by end users, particularly for regions that rely on only a small number of pixels. In such cases, the reliability of the spatially aggregated estimates might be reduced due to the limited number of pixels, their location within the land-sea mask of the underlying datasets (e.g., for small and remote islands), and the effective resolution of each data source. The effective resolution - the smallest feature that can be realistically represented - is often substantially larger (typically by a factor\u0026thinsp;\u0026ge;\u0026thinsp;2) in model-based datasets than the nominal grid spacing, which is commonly expressed in terms of grid pixels. In this regard, we also want to highlight that some spatially averaged climate variables for certain administrative units - especially island or coastal areas - contain no data or only data for CHIRPS-derived variables (i.e., pr). This behaviour is expected and arises from differences in land-sea masks and spatial land coverage between datasets, with CHIRPS providing coverage in some coastal or island areas where ERA5-Land does not. Furthermore, spatial aggregation was performed by averaging values across all grid cells touching each administrative unit or country, consistent with common practice in climate science and accounting for the effective resolution of the underlying datasets. Consequently, when summaries across multiple administrative units are required, we recommend recomputing the spatial aggregation using the corresponding combined boundaries rather than relying on precomputed area-level estimates. Furthermore, while administrative unit identifiers (GIDs) were directly extracted from the official GADM geopackage (version 4.1), we noted inconsistencies in GID version suffixes in certain countries. Although GADM version 4.1 would nominally imply a uniform suffix of _1, several administrative units include alternative suffixes (e.g., _2) within the official geopackage. These discrepancies are inherent to the GADM data and are therefore reflected in our naming of the corresponding Parquet files. For the 104 countries, spatially averaged (population-weighted) estimates were generated for administrative levels 0, 1 and 2 where available, but GADM version 4.1 does not provide all levels for some countries (e.g., Belize), so only available levels are included (see Supplementary Material Table\u0026nbsp;2 for details).\u003c/p\u003e \u003cp\u003eConcluding, our dataset, comprising quasi-global, daily, high-resolution, and spatially aggregated climate variables derived from downscaled CMIP6 models, offers a unified and consistent resource for assessing future climate change and variability, supports high-resolution impact assessment models, and provides accessible climate information at administrative unit levels - particularly benefiting researcher, non-specialists and stakeholders working in tropical regions, including those in the Global South.\u003c/p\u003e"},{"header":"Usage Notes","content":"\u003cp\u003eThe compressed NetCDF files are written following the CF Conventions. NetCDF data can be analysed in standard programming languages like R (e.g., using packages like ncdf4, raster or terra) or in Python (e.g., using xarray or netcdf4). Additionally, we recommend using the Climate Data Operators (CDO) tool\u003csup\u003e\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e\u003c/sup\u003e for simple dataset operations and visualizations of (uncompressed) NetCDF data, including its Python interface, which acts as a wrapper for the CDO command-line binary. For less experienced users, we note that NetCDF data can also be read and used in GIS-oriented applications, such as the open-source QGIS platform. Please also refer to metadata.yaml provided alongside the NetCDF data for additional information.\u003c/p\u003e \u003cp\u003eIn addition to the pixel-based NetCDF dataset, we provide spatially aggregated and population-weighted data in an accessible, user-friendly format. These data, aggregated at national and sub-national levels (up to administrative unit level 2), can be easily selected and are customizable by e.g., the applied weighting schemes, temporal resolution, timeframe and administrative level, making them suitable for direct use in climate impact assessments. This flexibility enhances the replicability of impact studies, promotes transparent data practices, and facilitates seamless integration with other domain-specific datasets, such as those used in health-related impact research. Again, there are common packages available in standard programming languages such as R and Python for reading data in Parquet format (e.g., the pandas library can utilize pyarrow as a backend to efficiently read Parquet files in Python). Please also refer to readme.pdf provided alongside the data for additional information. In particular, after downloading the dataset, it is important to verify that the number of files in each folder matches the numbers specified in the readme.pdf.\u003c/p\u003e \u003cp\u003eThe dataset is therefore designed to support the growing climate impact research field, which increasingly encompasses disciplines such as epidemiology and public health. We hence anticipate that this resource will be particularly valuable for infectious disease modelers, epidemiologists, and other practitioners outside the traditional climate science community. When using the generated NetCDF data or the spatially aggregated information provided in the Parquet files, users should cite this publication. Users should also always acknowledge the underlying data sources used to construct this dataset, particularly ERA5-Land, CHIRPS, and the CMIP6 models, by citing their respective data records and associated publications.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch3\u003eCode and Data Availability\u003c/h3\u003e\n\u003cp\u003eThe DBCCA code used to downscale the CMIP6 GCMs follows the methodology described in Werner and Cannon \u003csup\u003e32\u003c/sup\u003e as well as in the University of Toronto Climate Downscaling Workflow Guidebook (UTCDW)\u003csup\u003e75\u003c/sup\u003e, with example code available at the GitHub repository https://github.com/mikemorris12/UTCDW_Guidebook (released under the MIT License), with the respective algorithm setup here adapted and refined as described to generate the bias-corrected and downscaled climate projections across all GCMs, variables and scenarios for all 12 domains. Python code for downloading and preparing the observational reference datasets, validating all input datasets, and aggregating both the reference data and the bias-corrected and downscaled climate projections is also available at https://github.com/SallyAJ/cmip6-ref-aggregation.git (released under the MIT License). The complete dataset is available on Box under : \u0026nbsp;https://imperialcollegelondon.box.com/s/y4f0ywor4gvcrdxkewlzerfxw4lik4ts .\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003eAcknowledgements\u003c/h3\u003e\n\u003cp\u003eThis work was carried out as part of the Vaccine Impact Modelling Consortium (www.vaccineimpact.org), but the views expressed are those of the authors and not necessarily those of the Consortium or its funders. The funders were given the opportunity to review this paper prior to publication, but the final decision on the content of the publication was taken by the authors. This work was primarily supported by the Wellcome Trust via the Vaccine Impact Modelling Consortium [Grant Number 226727_Z_22_Z], with additional support from the Bill \u0026amp; Melinda Gates Foundation [Grant Number INV-034281], previously (OPP1157270/INV-009125), and Gavi, the Vaccine Alliance. Under the grant conditions of the Foundation, a Creative Commons Attribution 4.0 Generic License has already been assigned to the Author Accepted Manuscript version that might arise from this submission. We also acknowledge funding provided by the Jameel Institute (supported by a philanthropic donation from Community Jameel) and from the MRC Centre for Global Infectious Disease Analysis (reference MR/X020258/1), funded by the UK Medical Research Council (MRC). The latter UK-funded award is carried out in the frame of the Global Health EDCTP3 Joint Undertaking. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors would also like to thank Chris De La Force for providing technical support related to high-performance computing and cluster management.\u003c/p\u003e\n\u003ch3\u003eAuthor contributions\u003c/h3\u003e\n\u003cp\u003eS.J. performed the data collection, wrangling, and processing, including the implementation of the bias-correction and downscaling algorithm on CMIP6 model outputs, as well as the spatial aggregation of climate variables at country and administrative unit levels. S.J. and C.W. conceived the research idea and analytical approach, supported and guided by discussions with I.D., P.W., R.T., and N.F.. W.H. and K.G. supported discussions on the logistics of computing, data storage, and distribution, while all authors contributed to various aspects of computing, storage, and data dissemination. S.J. developed and prepared the original draft, and all authors contributed to reviewing and editing the manuscript. N.F. supervised the research.\u003c/p\u003e\n\u003ch3\u003eCompeting interests\u003c/h3\u003e\n\u003cp\u003eThe authors declare that no competing interests exist.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eIntergovernmental Panel on Climate Change (IPCC). Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. (Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEyring, V. \u003cem\u003eet al.\u003c/em\u003e Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. \u003cem\u003eGeosci Model Dev\u003c/em\u003e 9, 1937\u0026ndash;1958 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/gmd-9-1937-2016\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/gmd-9-1937-2016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVogel, E. \u003cem\u003eet al.\u003c/em\u003e An evaluation framework for downscaling and bias correction in climate change impact studies. \u003cem\u003eJ. Hydrol.\u003c/em\u003e 622, 129693 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.jhydrol.2023.129693\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.jhydrol.2023.129693\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaraun, D. \u0026amp; Widmann, M. \u003cem\u003eStatistical Downscaling and Bias Correction for Climate Research\u003c/em\u003e. (Cambridge University Press, 2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCannon, A. J., Piani, C. \u0026amp; Sippel, S. in \u003cem\u003eClimate Extremes and Their Implications for Impact and Risk Assessment\u003c/em\u003e (eds Jana Sillmann, Sebastian Sippel, \u0026amp; Simone Russo) Ch. 5, 77\u0026ndash;104 (Elsevier, 2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeutschbein, C. \u0026amp; Seibert, J. Bias correction of regional climate model simulations for hydrological climate-change impact studies: Review and evaluation of different methods. \u003cem\u003eJ. Hydrol.\u003c/em\u003e 456\u0026ndash;457, 12\u0026ndash;29 (2012). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.jhydrol.2012.05.052\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.jhydrol.2012.05.052\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMenapace, A. \u003cem\u003eet al.\u003c/em\u003e Review of bias correction methods for climate model outputs in hydrology. \u003cem\u003eJ. Hydrol.\u003c/em\u003e 660, 133213 (2025). https://doi.org:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jhydrol.2025.133213\u003c/span\u003e\u003cspan address=\"10.1016/j.jhydrol.2025.133213\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGebrechorkos, S. H., Hulsmann, S. \u0026amp; Bernhofer, C. Statistically downscaled climate dataset for East Africa. \u003cem\u003eSci. Data\u003c/em\u003e 6, 31 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41597-019-0038-1\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41597-019-0038-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaraun, D. \u003cem\u003eet al.\u003c/em\u003e Precipitation downscaling under climate change: Recent developments to bridge the gap between dynamical models and the end user. \u003cem\u003eRev. Geophys.\u003c/em\u003e 48, RG3003 (2010). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1029/2009rg000314\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1029/2009rg000314\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKeller, A. A., Garner, K. L., Rao, N., Knipping, E. \u0026amp; Thomas, J. Downscaling approaches of climate change projections for watershed modeling: Review of theoretical and practical considerations. \u003cem\u003ePLOS Water\u003c/em\u003e 1, e0000046 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1371/journal.pwat.0000046\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1371/journal.pwat.0000046\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChow, K. K. C., Sankar\u0026eacute;, H., Diaconescu, E. P., Murdock, T. Q. \u0026amp; Cannon, A. J. Bias-adjusted and downscaled humidex projections for heat preparedness and adaptation in Canada. \u003cem\u003eGeosci. Data J.\u003c/em\u003e 11, 680\u0026ndash;698 (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1002/gdj3.241\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1002/gdj3.241\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJacobeit, J., Hertig, E., Seubert, S. \u0026amp; Lutz, K. Statistical downscaling for climate change projections in the Mediterranean region: methods and results. \u003cem\u003eReg Environ Change\u003c/em\u003e 14, 1891\u0026ndash;1906 (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1007/s10113-014-0605-0\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1007/s10113-014-0605-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTarek, M., Brissette, F. \u0026amp; Arsenault, R. Uncertainty of gridded precipitation and temperature reference datasets in climate change impact studies. \u003cem\u003eHydrol. Earth Syst. Sci.\u003c/em\u003e 25, 3331\u0026ndash;3350 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/hess-25-3331-2021\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/hess-25-3331-2021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTarek, M., Brissette, F. P. \u0026amp; Arsenault, R. Large-Scale Analysis of Global Gridded Precipitation and Temperature Datasets for Climate Change Impact Studies. \u003cem\u003eJ. Hydrometeorol.\u003c/em\u003e 21, 2623\u0026ndash;2640 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1175/jhm-d-20-0100.1\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1175/jhm-d-20-0100.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJahn, S., Fraser, K., Gaythorpe, K. A. M., Wainwright, C. M. \u0026amp; Ferguson, N. M. Evaluating the role of observational uncertainty in climate impact assessments: Temperature-driven yellow fever risk in South America. \u003cem\u003ePLOS Climate\u003c/em\u003e 4, e0000601 (2025). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1371/journal.pclm.0000601\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1371/journal.pclm.0000601\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJahn, S., Gaythorpe, K. A. M., Wainwright, C. M. \u0026amp; Ferguson, N. M. Evaluation of the Performance and Utility of Global Gridded Precipitation Products for Health Applications and Impact Assessments in South America. \u003cem\u003eGeohealth\u003c/em\u003e 9, e2024GH001260 (2025). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1029/2024GH001260\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1029/2024GH001260\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLudwig, R., Schmid, J. \u0026amp; Gampe, D. Impact of Reference Dataset Selection on RCM Evaluation, Bias Correction, and Resulting Climate Change Signals of Precipitation. \u003cem\u003eJ. Hydrometeorol.\u003c/em\u003e 20, 1813\u0026ndash;1828 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1175/jhm-d-18-0108.1\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1175/jhm-d-18-0108.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGutowski, W. J. \u003cem\u003eet al.\u003c/em\u003e WCRP COordinated Regional Downscaling EXperiment (CORDEX): a diagnostic MIP for CMIP6. \u003cem\u003eGeosci Model Dev\u003c/em\u003e 9, 4087\u0026ndash;4095 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/gmd-9-4087-2016\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/gmd-9-4087-2016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaraun, D. Bias Correcting Climate Change Simulations - a Critical Review. \u003cem\u003eCurrent Climate Change Reports\u003c/em\u003e 2, 211\u0026ndash;220 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1007/s40641-016-0050-x\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1007/s40641-016-0050-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYakubu, F., Bohner, J., Schickhoff, U., Scholten, T. \u0026amp; Hasson, S. U. Global Bias-Corrected CORDEX Datasets at Half Degree Resolution. \u003cem\u003eSci Data\u003c/em\u003e 12, 1781 (2025). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41597-025-06200-4\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41597-025-06200-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrein, A. F. \u003cem\u003eet al.\u003c/em\u003e A review on regional convection-permitting climate modeling: Demonstrations, prospects, and challenges. \u003cem\u003eRev. Geophys.\u003c/em\u003e 53, 323\u0026ndash;361 (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1002/2014RG000475\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1002/2014RG000475\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGiorgi, F. Thirty Years of Regional Climate Modeling: Where Are We and Where Are We Going next? \u003cem\u003eJ. Geophys. Res.: Atmos.\u003c/em\u003e 124, Pages 5696\u0026ndash;5723 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1029/2018jd030094\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1029/2018jd030094\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGiorgi, F., Jones, C. \u0026amp; Asrar, G. R. Addressing climate information needs at the regional level: The CORDEX framework. \u003cem\u003eWMO Bulletin\u003c/em\u003e 58, 175\u0026ndash;183 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChapman, S. \u003cem\u003eet al.\u003c/em\u003e Evaluation of Dynamically Downscaled CMIP6-CCAM Models Over Australia. \u003cem\u003eEarth's Future\u003c/em\u003e 11, e2023EF003548 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1029/2023ef003548\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1029/2023ef003548\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKendon, E. J. \u003cem\u003eet al.\u003c/em\u003e Enhanced future changes in wet and dry extremes over Africa at convection-permitting scale. \u003cem\u003eNat. Commun.\u003c/em\u003e 10, 1794 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41467-019-09776-9\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41467-019-09776-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSenior, C. \u003cem\u003eet al. Technical guidelines for using CP4-Africa simulation data\u003c/em\u003e, \u0026lt;\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5281/zenodo.4316466%3E\u003c/span\u003e\u003cspan address=\"10.5281/zenodo.4316466%3E\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVosper, S. B. \u003cem\u003eet al.\u003c/em\u003e A Pan-African Convection-Permitting Regional Climate Simulation with the Met Office Unified Model: CP4-Africa. \u003cem\u003eJ. Clim.\u003c/em\u003e 31, 3485\u0026ndash;3508 (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1175/jcli-d-17-0503.1\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1175/jcli-d-17-0503.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuti\u0026eacute;rrez, J. M. \u003cem\u003eet al.\u003c/em\u003e An intercomparison of a large ensemble of statistical downscaling methods over Europe: Results from the VALUE perfect predictor cross-validation experiment. \u003cem\u003eInt. J. Climatol.\u003c/em\u003e 39, 3750\u0026ndash;3785 (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1002/joc.5462\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1002/joc.5462\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePierce, D. W., Cayan, D. R. \u0026amp; Thrasher, B. L. Statistical Downscaling Using Localized Constructed Analogs (LOCA). \u003cem\u003eJ. Hydrometeorol.\u003c/em\u003e 15, 2558\u0026ndash;2585 (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1175/jhm-d-14-0082.1\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1175/jhm-d-14-0082.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaraun, D. Bias Correction, Quantile Mapping, and Downscaling: Revisiting the Inflation Issue. \u003cem\u003eJ. Clim.\u003c/em\u003e 26, 2137\u0026ndash;2143 (2013). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1175/jcli-d-12-00821.1\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1175/jcli-d-12-00821.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHnilica, J., Hanel, M. \u0026amp; Puš, V. Multisite bias correction of precipitation data from regional climate models. \u003cem\u003eInt. J. Clim.\u003c/em\u003e 37, 2934\u0026ndash;2946 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1002/joc.4890\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1002/joc.4890\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWerner, A. T. \u0026amp; Cannon, A. J. Hydrologic extremes \u0026ndash; an intercomparison of multiple gridded statistical downscaling methods. \u003cem\u003eHydrol. Earth Syst. Sci.\u003c/em\u003e 20, 1483\u0026ndash;1508 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/hess-20-1483-2016\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/hess-20-1483-2016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGaythorpe, K. A., Hamlet, A., Cibrelus, L., Garske, T. \u0026amp; Ferguson, N. M. The effect of climate change on yellow fever disease burden in Africa. \u003cem\u003eElife\u003c/em\u003e 9, e55619 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.7554/eLife.55619\u003c/span\u003e\u003cspan address=\"https://doi.org:10.7554/eLife.55619\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSargent, K., Mollard, J., Henley, S. F. \u0026amp; Bollasina, M. A. Predicting Transmission Suitability of Mosquito-Borne Diseases under Climate Change to Underpin Decision Making. \u003cem\u003eInt. J. Environ. Res. Public Health\u003c/em\u003e 19, 13656 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.3390/ijerph192013656\u003c/span\u003e\u003cspan address=\"https://doi.org:10.3390/ijerph192013656\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRyan, S. J., Carlson, C. J., Mordecai, E. A. \u0026amp; Johnson, L. R. Global expansion and redistribution of Aedes-borne virus transmission risk with climate change. \u003cem\u003ePLoS Negl. Trop. Dis.\u003c/em\u003e 13, e0007213 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1371/journal.pntd.0007213\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1371/journal.pntd.0007213\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRyan, S. J. \u003cem\u003eet al.\u003c/em\u003e Warming temperatures could expose more than 1.3 billion new people to Zika virus risk by 2050. \u003cem\u003eGlob. Chan. Biol.\u003c/em\u003e 27, 84\u0026ndash;93 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1111/gcb.15384\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1111/gcb.15384\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFick, S. E. \u0026amp; Hijmans, R. J. WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. \u003cem\u003eInt. J. Climatol.\u003c/em\u003e 37, 4302\u0026ndash;4315 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1002/joc.5086\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1002/joc.5086\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFick, S. E. \u0026amp; Hijmans, R. J. \u003cem\u003eDownscaling future and past climate data from GCMs\u003c/em\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e%3Chttps://www.worldclim.org/data/downscaling.html%3E\u003c/span\u003e\u003cspan address=\"http://%3Chttps://www.worldclim.org/data/downscaling.html%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFarooq, Z. \u003cem\u003eet al.\u003c/em\u003e Impact of climate and Aedes albopictus establishment on dengue and chikungunya outbreaks in Europe: a time-to-event analysis. \u003cem\u003eLancet Planet Health\u003c/em\u003e 9, e374-e383 (2025). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/S2542-5196(25)00059-2\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/S2542-5196(25)00059-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eColon-Gonzalez, F. J. \u003cem\u003eet al.\u003c/em\u003e Projecting the future incidence and burden of dengue in Southeast Asia. \u003cem\u003eNat Commun\u003c/em\u003e 14, 5439 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41467-023-41017-y\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41467-023-41017-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLange, S. Trend-preserving bias adjustment and statistical downscaling with ISIMIP3BASD (v1.0). \u003cem\u003eGeosci Model Dev\u003c/em\u003e 12, 3055\u0026ndash;3070 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/gmd-12-3055-2019\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/gmd-12-3055-2019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThrasher, B., Maurer, E. P., McKellar, C. \u0026amp; Duffy, P. B. Technical Note: Bias correcting climate model simulated daily temperature extremes with quantile mapping. \u003cem\u003eHydrol. Earth Syst. Sci.\u003c/em\u003e 16, 3309\u0026ndash;3314 (2012). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/hess-16-3309-2012\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/hess-16-3309-2012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThrasher, B. \u003cem\u003eet al.\u003c/em\u003e NASA Global Daily Downscaled Projections, CMIP6. \u003cem\u003eSci. Data\u003c/em\u003e 9, 262 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41597-022-01393-4\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41597-022-01393-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIwamura, T., Guzman-Holst, A. \u0026amp; Murray, K. A. Accelerating invasion potential of disease vector Aedes aegypti under climate change. \u003cem\u003eNat Commun\u003c/em\u003e 11, 2130 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41467-020-16010-4\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41467-020-16010-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSymons, T. L. \u003cem\u003eet al.\u003c/em\u003e Projected impacts of climate change on malaria in Africa. \u003cem\u003eNature\u003c/em\u003e (2026). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41586-025-10015-z\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41586-025-10015-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSheffield, J., Goteti, G. \u0026amp; Wood, E. F. Development of a 50-year high-resolution global dataset of meteorological forcings for land surface modeling. \u003cem\u003eJ. Clim.\u003c/em\u003e 19, 3088\u0026ndash;3111 (2006). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:\u003c/span\u003e\u003cspan address=\"https://doi.org:\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1175/JCLI3790.1\u003c/span\u003e\u003cspan address=\"10.1175/JCLI3790.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHassler, B. \u0026amp; Lauer, A. Comparison of Reanalysis and Observational Precipitation Datasets Including ERA5 and WFDE5. \u003cem\u003eAtmosphere\u003c/em\u003e 12, 1462 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.3390/atmos12111462\u003c/span\u003e\u003cspan address=\"https://doi.org:10.3390/atmos12111462\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGebrechorkos, S. \u003cem\u003eet al.\u003c/em\u003e A high-resolution daily global dataset of statistically downscaled CMIP6 models for climate impact analyses. \u003cem\u003eSci. Data\u003c/em\u003e 10, 611 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41597-023-02528-x\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41597-023-02528-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGergel, D. R. \u003cem\u003eet al.\u003c/em\u003e Global Downscaled Projections for Climate Impacts Research (GDPCIR): preserving quantile trends for modeling future climate impacts. \u003cem\u003eGeosci Model Dev\u003c/em\u003e 17, 191\u0026ndash;227 (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/gmd-17-191-2024\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/gmd-17-191-2024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLavers, D. A., Simmons, A., Vamborg, F. \u0026amp; Rodwell, M. J. An evaluation of ERA5 precipitation for climate monitoring. \u003cem\u003eQ. J. R. Meteorol. Soc.\u003c/em\u003e 148, 3152\u0026ndash;3165 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1002/qj.4351\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1002/qj.4351\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGortan, M., Testa, L., Fagiolo, G. \u0026amp; Lamperti, F. A unified dataset for pre-processed climate indicators weighted by gridded economic activity. \u003cem\u003eSci. Data\u003c/em\u003e 11, 533 (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41597-024-03304-1\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41597-024-03304-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmith, T. P., Stemkovski, M., Koontz, A. \u0026amp; Pearse, W. D. AREAdata: A worldwide climate dataset averaged across spatial units at different scales through time. \u003cem\u003eData Brief\u003c/em\u003e 43, 108438 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.dib.2022.108438\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.dib.2022.108438\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO'Neill, B. C. \u003cem\u003eet al.\u003c/em\u003e The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6. \u003cem\u003eGeosci Model Dev\u003c/em\u003e 9, 3461\u0026ndash;3482 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/gmd-9-3461-2016\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/gmd-9-3461-2016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO\u0026rsquo;Neill, B. C. \u003cem\u003eet al.\u003c/em\u003e The roads ahead: Narratives for shared socioeconomic pathways describing world futures in the 21st century. \u003cem\u003eGlob. Environ. Change\u003c/em\u003e 42, 169\u0026ndash;180 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.gloenvcha.2015.01.004\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.gloenvcha.2015.01.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRiahi, K. \u003cem\u003eet al.\u003c/em\u003e The Shared Socioeconomic Pathways and their energy, land use, and greenhouse gas emissions implications: An overview. \u003cem\u003eGlob. Environ. Change\u003c/em\u003e 42, 153\u0026ndash;168 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.gloenvcha.2016.05.009\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.gloenvcha.2016.05.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGebrechorkos, S. H., H\u0026uuml;lsmann, S. \u0026amp; Bernhofer, C. Evaluation of multiple climate data sources for managing environmental resources in East Africa. \u003cem\u003eHydrol. Earth Syst. Sci.\u003c/em\u003e 22, 4547\u0026ndash;4564 (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/hess-22-4547-2018\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/hess-22-4547-2018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeck, H. E. \u003cem\u003eet al.\u003c/em\u003e Global-scale evaluation of 22 precipitation datasets using gauge observations and hydrological modeling. \u003cem\u003eHydrol. Earth Syst. Sci.\u003c/em\u003e 21, 6201\u0026ndash;6217 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/hess-21-6201-2017\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/hess-21-6201-2017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVIMC (Vaccine Impact Modelling Consortium). \u003cem\u003eVaccine Impact Modelling Consortium\u003c/em\u003e, \u0026lt;\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.vaccineimpact.org/%3E\u003c/span\u003e\u003cspan address=\"https://www.vaccineimpact.org/%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEarth System Grid Federation (ESGF) User Support Working Team. \u003cem\u003eESGF User Support: User guide\u003c/em\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e%3Chttps://esgf.github.io/esgf-user-support/user_guide.html%3E\u003c/span\u003e\u003cspan address=\"http://%3Chttps://esgf.github.io/esgf-user-support/user_guide.html%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIntergovernmental Panel on Climate Change (IPCC). Climate Change 2022: Impacts, Adaptation and Vulnerability. Working Group II contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. (Cambridge, UK and New York, NY, USA, 2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCopernicus Climate Change Service (C3S). \u003cem\u003eERA5-Land hourly data from 1950 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS).\u003c/em\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e%3Chttps://cds.climate.copernicus.eu/datasets/reanalysis-era5-land?tab=download%3E\u003c/span\u003e\u003cspan address=\"http://%3Chttps://cds.climate.copernicus.eu/datasets/reanalysis-era5-land?tab=download%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMu\u0026ntilde;oz-Sabater, J. \u003cem\u003eet al.\u003c/em\u003e ERA5-Land: a state-of-the-art global reanalysis dataset for land applications. \u003cem\u003eEarth Syst. Sci. Data\u003c/em\u003e 13, 4349\u0026ndash;4383 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/essd-13-4349-2021\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/essd-13-4349-2021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eClimate Hazards Center - University of California Santa Barbara. \u003cem\u003eClimate Hazards Center Infrared Precipitation with Stations version 3. CHIRPS3 Data Repository\u003c/em\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e%3Chttps://esgf.github.io/esgf-user-support/user_guide.html%3E\u003c/span\u003e\u003cspan address=\"http://%3Chttps://esgf.github.io/esgf-user-support/user_guide.html%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFunk, C. C. \u003cem\u003eet al.\u003c/em\u003e The climate hazards infrared precipitation with stations\u0026ndash;a new environmental record for monitoring extremes. \u003cem\u003eSci. Data\u003c/em\u003e 2, 150066 (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/sdata.2015.66\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/sdata.2015.66\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFunk, C. C. \u003cem\u003eet al.\u003c/em\u003e A quasi-global precipitation time series for drought monitoring. Report No. 832, 12 (Reston, VA, 2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHijmans, R. \u003cem\u003eet al. GADM database of Global Administrative Areas, version 4.1.\u003c/em\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e%3Chttps://gadm.org/%3E\u003c/span\u003e\u003cspan address=\"http://%3Chttps://gadm.org/%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGao, J. \u0026amp; National Center for Atmospheric Research (NCAR). Downscaling global spatial population projections from 1/8-degree to 1-km grid cells. \u003cem\u003eNCAR Technical Note\u003c/em\u003e (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5065/D60Z721H\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5065/D60Z721H\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGao, J. \u0026amp; Socioeconomic Data and Applications Center (SEDAC). Global 1-km Downscaled Population Grids, SSP-Consistent Projections and Base Year, v1.01 (2000\u0026ndash;2100) (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:\u003c/span\u003e\u003cspan address=\"https://doi.org:\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7910/DVN/TLJ99B\u003c/span\u003e\u003cspan address=\"10.7910/DVN/TLJ99B\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNatural Earth. \u003cem\u003eNatural Earth - Free vector and raster map data\u003c/em\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e%3Chttps://www.naturalearthdata.com/about/%3E\u003c/span\u003e\u003cspan address=\"http://%3Chttps://www.naturalearthdata.com/about/%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e(ECMWF), E. C. f. M.-R. W. F. \u003cem\u003eIFS documentation\u003c/em\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e%3Chttps://www.ecmwf.int/en/publications/ifs-documentation%3E\u003c/span\u003e\u003cspan address=\"http://%3Chttps://www.ecmwf.int/en/publications/ifs-documentation%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePierce, D. W. \u003cem\u003eLOCA Statistical Downscaling (Localized Constructed Analogs)\u003c/em\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e%3Chttps://loca.ucsd.edu/loca-calendar/%3E\u003c/span\u003e\u003cspan address=\"http://%3Chttps://loca.ucsd.edu/loca-calendar/%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorld Climate Research Programme (WCRP). \u003cem\u003eCORDEX domain description - Download domain description (update 23/10/15)\u003c/em\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e%3Chttps://cordex.org/domains/cordex-domain-description/%3E\u003c/span\u003e\u003cspan address=\"http://%3Chttps://cordex.org/domains/cordex-domain-description/%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eExpert Team on Climate Change Detection and Indices (ETCCDI). \u003cem\u003eETCCDI Climate Change Indices\u003c/em\u003e, \u0026lt;\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://etccdi.pacificclimate.org/index.shtml%3E\u003c/span\u003e\u003cspan address=\"https://etccdi.pacificclimate.org/index.shtml%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchulzweida, U. \u0026amp; Max Planck Institute (MPI) for Meteorology. \u003cem\u003eCDO (Climate Data Operators) Version 2.5.2 - Documentation: CDO User Guide\u003c/em\u003e, \u0026lt;\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://code.mpimet.mpg.de/projects/cdo/wiki/Cdo#Documentation%3E\u003c/span\u003e\u003cspan address=\"https://code.mpimet.mpg.de/projects/cdo/wiki/Cdo#Documentation%3E\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorris, M.,;, Kushner, P. J. \u0026amp; Smith, K. L. The University of Toronto Climate Downscaling Workflow: Tools and Resources for Climate Change Impact Analysis. \u003cem\u003eJ. open source educ.\u003c/em\u003e 7, 243 (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.21105/jose.00243\u003c/span\u003e\u003cspan address=\"https://doi.org:10.21105/jose.00243\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSwart, N. C. \u003cem\u003eet al.\u003c/em\u003e The Canadian Earth System Model version 5 (CanESM5.0.3). \u003cem\u003eGeosci Model Dev\u003c/em\u003e 12, 4823\u0026ndash;4873 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/gmd-12-4823-2019\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/gmd-12-4823-2019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDunne, J. P. \u003cem\u003eet al.\u003c/em\u003e The GFDL Earth System Model Version 4.1 (GFDL-ESM 4.1): Overall Coupled Model Description and Simulation Characteristics. \u003cem\u003eJ. Adv. Model. Earth Syst.\u003c/em\u003e 12 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1029/2019ms002015\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1029/2019ms002015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMauritsen, T. \u003cem\u003eet al.\u003c/em\u003e Developments in the MPI-M Earth System Model version 1.2 (MPI-ESM1.2) and Its Response to Increasing CO(2). \u003cem\u003eJ. Adv. Model Earth. Syst.\u003c/em\u003e 11, 998\u0026ndash;1038 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1029/2018MS001400\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1029/2018MS001400\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYukimoto, S. \u003cem\u003eet al.\u003c/em\u003e The Meteorological Research Institute Earth System Model Version 2.0, MRI-ESM2.0: Description and Basic Evaluation of the Physical Component. \u003cem\u003eJ. Meteor. Soc. Japan. Ser. II\u003c/em\u003e 97, 931\u0026ndash;965 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.2151/jmsj.2019-051\u003c/span\u003e\u003cspan address=\"https://doi.org:10.2151/jmsj.2019-051\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee, W.-L. \u003cem\u003eet al.\u003c/em\u003e Taiwan Earth System Model Version 1: description and evaluation of mean state. \u003cem\u003eGeosci Model Dev\u003c/em\u003e 13, 3887\u0026ndash;3904 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.5194/gmd-13-3887-2020\u003c/span\u003e\u003cspan address=\"https://doi.org:10.5194/gmd-13-3887-2020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSellar, A. A. \u003cem\u003eet al.\u003c/em\u003e UKESM1: Description and Evaluation of the U.K. Earth System Model. \u003cem\u003eJ. Adv. Model. Earth Syst.\u003c/em\u003e 11, 4513\u0026ndash;4558 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1029/2019ms001739\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1029/2019ms001739\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9508034/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9508034/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eGlobal climate models (GCMs) are essential tools for understanding the climate system and projecting its evolution under different scenarios. However, differences in model construction introduce uncertainties, and GCMs have coarse resolution and inherent biases, limiting their effectiveness for informing local or regional adaptation and mitigation planning. We apply the statistical Double Bias-Corrected Constructed Analogues (DBCCA) method to generate bias-corrected and downscaled climate projections at daily and 0.1\u0026deg; spatial resolution for 1985\u0026ndash;2100, with a specific focus on supporting tropical health-related impact research. Our quasi-global, high-resolution projections are based on six GCMs from Phase 6 of the Coupled Model Intercomparison Project (CMIP6) under two emission scenarios (SSP2-4.5 and SSP5-8.5), covering 12 land-only domains between 60\u0026deg;N and 60\u0026deg;S. Moreover, many impact researchers rely on accessible data aggregated to administrative units, often weighted by population, rather than gridded data, as these align more directly with policy- and decision-making. Based on our projections, we also provide user-ready (population-weighted) spatially aggregated climate variables at administrative unit levels (0\u0026ndash;2) for 104 countries prioritized for tropical disease research.\u003c/p\u003e","manuscriptTitle":"Quasi-global, land-only, high-resolution and spatially averaged climate variables from downscaled CMIP6 models for climate impact research","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-29 06:59:55","doi":"10.21203/rs.3.rs-9508034/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"33ebca6d-67f7-4db9-acae-b6fd82f6c441","owner":[],"postedDate":"April 29th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"288460781951109469782915994476327568607","date":"2026-05-12T10:40:22+00:00","index":38,"fulltext":""},{"type":"reviewerAgreed","content":"170332719656339582251732114257609517500","date":"2026-05-12T06:33:57+00:00","index":36,"fulltext":""},{"type":"reviewersInvited","content":"14","date":"2026-05-06T12:43:20+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-05-04T09:53:11+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-05-01T07:31:56+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-05-01T07:31:23+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-05-06T12:53:33+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-29 06:59:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9508034","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9508034","identity":"rs-9508034","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}