Diversity of Mesoamerican Midsummer Drought

Diversity of Mesoamerican Midsummer Drought | 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PDF Research Article Diversity of Mesoamerican Midsummer Drought Zijie Zhao, Yanxuan Du This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7652378/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract The Mesoamerican midsummer drought (MSD) is a distinctive precipitation feature characterized by a mid-season rainfall reduction within the boreal summer wet season. Despite its socioeconomic significance, most previous studies have emphasized its canonical bimodal structure, leaving the diversity of MSD expressions less explored. Using ERA5 reanalysis (1979–2019), we objectively identify MSD events for each grid cells over the domain and classify them into four distinct clusters via K-means analysis. These clusters reveal diverse temporal structures, intensities, and spatial preferences, spanning southern Mexico, Central America, the Caribbean basin, and adjacent oceans. Composite analyses of sea surface temperature (SST), cloud fraction, winds, and moisture flux convergence (MFC) indicate that low-level circulations—particularly the Caribbean and Chocó low-level jets, along with eastern Pacific winds—play a central role in shaping MSD diversity. In contrast, eastern Pacific SST anomalies exhibit only weak and inconsistent associations, suggesting a secondary role of SST–cloud feedbacks. Decomposition of MFC further highlights the combined zonal and meridional moisture transport as the primary driver of bimodality, with meridional fluxes being especially important for Caribbean MSD events. An evaluation of 33 CMIP6 models shows that while most capture the overall MSD frequency, they underperform in reproducing asymmetric precipitation structures, particularly those over the Caribbean. These results emphasize the need to incorporate MSD diversity into model evaluation frameworks to improve regional precipitation projections. Our findings provide a new perspective on the mechanisms underlying MSD variability and establish a foundation for more reliable seasonal prediction and climate change assessments in the Intra-Americas Seas region. Midsummer Drought Precipitation Drought CMIP6 Cluster Analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 1. Introduction A large portion of the Intra-Americas Seas region—including Central America, southern Mexico, the Caribbean, and adjacent oceanic areas—exhibits a bimodal annual precipitation cycle, typically marked by a relative “trough” within the canonical May–October rainy season (Durán-Quesada et al., 2020 , García-Franco et al., 2023 , Magaña et al., 1999, Mosiño and García, 1966 ). This feature, commonly referred to as the Midsummer Drought (MSD) (Magaña et al., 1999) or veranillo in local terminology (Alfaro, 2002 , Alfaro, 2014 , Dilley, 1996 ), exerts substantial socioeconomic impacts. In particular, it has been linked to agricultural losses (de Sousa et al., 2018 , Eakin et al., 2018 , Hellin et al., 2017 , Pons et al., 2017 ) and to notable shifts in its timing and spatial extent over recent decades (Anderson et al., 2019 , Corrales-Suastegui et al., 2020 , Rauscher et al., 2008 ). Given its regional importance, it is critical to advance our understanding of MSD characteristics and underlying mechanisms for improved seasonal precipitation prediction. A canonical seasonal precipitation cycle associated with the MSD is characterized by two distinct rainfall peaks: the first between May and July and the second between August and October, separated by a relative reduction in precipitation (Alfaro, 2002 , Alfaro and Hidalgo, 2017 , García-Franco et al., 2023 , Magaña et al., 1999, Maurer et al., 2022 ). However, both the bimodal structure and the timing and intensity of the MSD exhibit pronounced spatiotemporal variability. Strong MSD signatures have been identified along the Pacific side of Central America, the Yucatán Peninsula, parts of the Pacific coast of southern Mexico, and across several Caribbean islands, including Cuba and Haiti (Alfaro, 2002 , Alfaro, 2014 , Gamble et al., 2008 , Magaña et al., 1999, Perdigón-Morales et al., 2018 ). However, the occurrence of evident MSD signals along the Caribbean side of Central America remains less certain While some studies report only weak evidence of MSD in this region, others suggest that portions of the Caribbean coast may display a bimodal annual rainfall distribution (Karnauskas et al., 2013 , Maldonado et al., 2016 ). In the latter cases, however, the second peak often occurs in November, extending beyond the canonical May–October rainy season (Zhao and Zhang, 2021 ). Spatial variability is further evident in the increasing duration and intensity of the MSD toward the southeast along the Pacific side of southern Mexico and Central America (Anderson et al., 2019 , Zhao et al., 2020 ). Projections for the 21st century suggest substantial changes in MSD characteristics—including their timing, intensity, and spatial extent—indicating a potential emergence of enhanced spatial variability in MSD under future climate conditions. (Corrales‐Suastegui et al., 2020, Rauscher et al., 2008 ). This pronounced spatiotemporal variability underscores the complexity of characterizing the MSD, suggesting that a pure climatological definition based on canonical bimodal precipitation may be insufficient to capture its full range of expressions. The complexity of the MSD is further reflected in extensive efforts to elucidate its underlying mechanisms in literatures. Since the initial hypothesis that the two rainfall peaks of the MSD correspond to the double passage of the Intertropical Convergence Zone (ITCZ) (Hastenrath, 1976 ), numerous mechanisms have been proposed to explain the origin of its bimodal precipitation structure. Among them, the SST–cloud radiative feedback mechanism proposed by Magaña et al. (1999) emphasizes the role of Eastern Pacific (EP) sea surface temperature (SST) variability and associated cloud feedbacks. In this framework, early-summer SST warming enhances cloudiness through strong radiative effects, producing the first precipitation peak. The resulting increase in cloud cover reduces shortwave radiation while strengthened low-level winds enhance latent heat flux, together leading to SST cooling and suppressed convection, consistent with the midsummer reduction in rainfall. Subsequently, weakened convection allows greater solar insolation, that warms the SST again and supports a second rainfall peak. This mechanism was later refined by Magaña and Caetano ( 2005 ), who documented a bimodal SST structure resembling the MSD during the boreal summer of 2001. The SST–cloud radiative feedback implies that seasonal forcing alone can sustain recurrent MSD events, whereas their interannual diversity arises primarily from remote forcings. However, the correspondence between SST and MSD remains ambiguous: it is less evident in reanalysis products and climate model simulations, and SST variability does not fully resolve the observed precipitation seasonality associated with the MSD (García-Franco et al., 2023 ). Beyond eastern Pacific SST, the MSD has also been linked to low-level pressure gyres and their associated wind circulations (Amador, 2008 ). Seasonal variations of the North Atlantic Subtropical High (NASH) modulate the strength of the Caribbean Low-Level Jet (CLLJ) (Mapes et al., 2005 ), giving rise to its bimodal annual cycle (Wang, 2007 , Wang and Lee, 2007 ). The CLLJ is considered a key driver of moisture transport from the Caribbean into the canonical MSD region and plays a central role in regulating convective activity during MSD evolution through moisture accumulation and dissipation (Corrales-Suastegui et al., 2020 , Gamble et al., 2008 , Hidalgo et al., 2015 , Martinez et al., 2019 , Small et al., 2007 ). In addition to its role in shaping the year-to-year occurrence of the MSD, the CLLJ has been shown to influence precipitation variability across multiple time scales, including decadal (Cerón et al., 2021b ) and intraseasonal (Perdigón-Morales et al., 2021 ) variations over Central America. While the CLLJ plays a central role in modulating regional circulation, other low-level wind systems also contribute to precipitation variability across Mesoamerica and the Caribbean, most notably the eastern Pacific trade winds and the Chocó Low-Level Jet (CHLLJ). Variations in the eastern Pacific trade winds have been recognized as indicators of MSD timing and duration, as they modulate surface latent heat fluxes and may influence the development of low-level convergence associated with the first precipitation peak. The CHLLJ, in turn, serves as an important conveyor of moisture from the Pacific into Central America and Colombia, intensifying eastward during the rainy season (Cerón et al., 2021a, Gallego et al., 2019 , Sierra et al., 2021 ). Its strong moisture transport capacity favors extreme rainfall in regions such as Lloró, one of the rainiest localities worldwide (Poveda and Mesa, 2000 ). The CHLLJ further interacts across multiple scales with both the ITCZ and the CLLJ, jointly modulating the development of mesoscale convective systems during the rainy season (Durán-Quesada et al., 2020 , Loaiza Ceron et al., 2020 , Mejía et al., 2021 , Small et al., 2007 ). Although its contribution to the MSD has received relatively little attention, seasonal variations in the CHLLJ are likely linked to moisture availability during the MSD over Central America (Durán-Quesada et al., 2020 , Durán-Quesada et al., 2010 , Gallego et al., 2019 ). Collectively, these low-level wind systems enhance the spatiotemporal variability of the MSD and represent a key source of its regional diversity. Although multiple theories have been proposed to explain the mechanisms underlying the MSD, most have emphasized its canonical bimodal structure in seasonal precipitation distribution. Yet, previous studies have revealed substantial spatiotemporal variability, suggesting diverse characteristics, drivers, and physical backgrounds of the MSD. A better understanding of this diversity is essential to extend the concept of the MSD beyond its canonical framework and to improve fine-scale predictions that require explicit characterization of MSD features at the grid-cell level. This study aims to investigate the general aspects of MSD diversity, including its characterization, variability, and potential mechanisms. Section 2 describes the data and methods, with particular attention to the identification of MSD events and their diversity. Section 3 presents the main results, followed by a discussion in Section 4 . 2. Data and Method 2.1.Data In this study, we analyze precipitation from the ERA5 reanalysis (Hersbach et al., 2020 ) for the period 1979–2019 over the domain [0–30°N, 60–120°W] (Fig. 1 a). ERA5 is chosen because it provides a long-term (41 years), fine-scale dataset that reliably represents precipitation over both land and ocean, allowing for the construction of a robust climatology. Previous evaluations show that ERA5 reproduces the seasonal cycle and interannual variability of precipitation over Central America and Mexico reasonably well (Centella-Artola et al., 2020 , Morales-Velázquez et al., 2021 ), and it captures the mean state and timing of the MSD in close agreement with satellite-based and gauge-based products such as TRMM and CHIRPS (García-Franco et al., 2023 ). Additional diagnostics—including cloud fraction, SST, horizontal winds at different pressure levels, and specific humidity—are also obtained from ERA5. These fields have been extensively validated and show improved consistency with observations compared to earlier reanalyses. Taken together, these assessments indicate that ERA5 provides a suitable reference for characterizing the fine-scale features and dynamics of the MSD, which are often not fully captured by observations alone. Although small-scale biases remain, their influence is limited in this study, which focuses on larger-scale spatiotemporal climatology and intraseasonal variability. 2.2.MSD detection In this study, we use an event-detection method consisting of two steps to identify the MSD and its associated metrics (Zhao et al., 2020 ) (Fig. 1 b). We first calculate the seasonally varying precipitation climatology in each grid and locate two local maxima: one between May 15th and July 15th, and the other between August 15th and October 15th. Next, we perform two independent linear regressions for each grid: one from the first day of the year to the first maximum, and the other from the second maximum to the end of the year. If the fitted linear trend from the first regression is significantly positive and the trend from the second regression is significantly negative, the climatological existence of the MSD in that particular grid is confirmed (Fig. 1 c). Subsequently, we employ the aforementioned criteria to evaluate the precipitation time series of each year in that grid and determine the yearly occurrence of the MSD. For every MSD event, some metrics can be determined, such as the onset date that corresponds to the first local maximum, the end date that corresponds to the second local maximum, the peak date that corresponds to the minimum precipitation between the onset and end dates, and the duration that corresponds to the period between the onset and end dates. The intensity of each MSD event is quantified using the formula: $$\:{I}_{msd}=\frac{mean\left(P1,\:P2\right)-{P}_{msd}}{mean\left(P1,P2\right)},$$ where \(\:P1\) and \(\:P2\) are the first and second local precipitation maximum corresponding to the onset and end of the MSD, respectively, and \(\:{P}_{msd}\) is the average precipitation during the MSD event (from onset to end). During the MSD detection, the ERA5 precipitation is regridded from its original 0.25 o grid cells to 0.5 o and 2939 out of 7381 grids (121*61) have been determined to exhibit climatological MSD characteristics, and 63240 MSD events are subsequently detected. Outputs from this method broadly captures the MSD regions identified in previous studies (Anderson et al., 2019 , Karnauskas et al., 2013 , Magaña et al., 1999, Maurer et al., 2022 ), including southern Mexico, the Pacific side of Central America, and the Caribbean. To investigate the diversity of MSD occurrences, we apply a K-means clustering algorithm (Hartigan and Wong, 1979 ) to the MSD-related precipitation time series. For each identified event, the corresponding seasonal precipitation record is extracted at the grid level. To maintain temporal consistency, precipitation on February 29 of non-leap years is linearly interpolated from February 28 and March 1. This procedure yields a matrix of 63,240 rows and 366 columns, where each row corresponds to a precipitation time series associated with an MSD year. The series are then smoothed using a 31-day Gaussian filter to emphasize seasonal variability. K-means clustering is performed on this dataset, with the number of clusters determined iteratively. Specifically, a new cluster is retained only if its centroid shows less than 0.9 correlation with existing centroids and if a two-sample t-test suggests a distinct distribution. Following this procedure, four clusters are identified as the optimal configuration, capturing the main patterns of MSD diversity while preserving inter-cluster dissimilarity. 2.3.MSD diversity in CMIP6 To evaluate the ability of climate models to reproduce the diversity of the midsummer drought (MSD), we analyze 33 general circulation models (GCMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6). Historical precipitation simulations from 1980 to 2014 (Table 1 ) are first interpolated to a common horizontal resolution of 1.0° × 1.0°. MSD events are then identified in each model using the method described in Section 2.2 , with detection applied to all grid cells within the study domain, irrespective of whether MSD occurs in the local climatology. This choice reflects the fact that simulated precipitation climatologies may place MSD-like features in different spatial locations across models, making it difficult to compare intermodel differences if detection were limited to a fixed set of MSD-prone grid cells. Table 1 List of CMIP6 model experiments used in this study. For all models here we use their r1i1p1f1 variant. If the variant is not available, r2i1p1f1 variant is considered. Institution Model Reference AS-RCEC TaiESM1 (Lee and Liang, 2020 ) AWI AWI-ESM-1-1-LR (Danek et al., 2020 ) BCC BCC-CSM2-MR (Xin et al., 2018 ) BCC BCC-ESM1 (Zhang et al., 2018) CAS FGOALS-f3-L (YU, 2019 ) CAS FGOALS-g3 (Li, 2019 ) CCCR-IITM IITM-ESM (Choudhury et al., 2019 ) CCCma CanESM5 (Swart et al., 2019 ) CMCC CMCC-CM2-HR4 (Scoccimarro et al., 2020 ) CMCC CMCC-CM2-SR5 (Lovato and Peano, 2020 ) CMCC CMCC-ESM2 (Lovato et al., 2021 ) EC-Earth-Consortium EC-Earth3-AerChem (Consortium, 2020a ) EC-Earth-Consortium EC-Earth3-CC (Consortium, 2021 ) EC-Earth-Consortium EC-Earth3-Veg-LR (Consortium, 2020b ) EC-Earth-Consortium EC-Earth3-Veg (Wyser et al., 2020 ) EC-Earth-Consortium EC-Earth3 (Consortium, 2019 ) HAMMOZ-Consortium MPI-ESM-1-2-HAM (Neubauer et al., 2019 ) IPSL IPSL-CM5A2-INCA (Boucher et al., 2020) IPSL IPSL-CM6A-LR-INCA (Boucher et al., 2021) IPSL IPSL-CM6A-LR (Boucher et al., 2018) MIROC MIROC-ES2L (Hajima et al., 2019) MIROC MIROC6 (Tatebe and Watanabe, 2018 ) MPI-M MPI-ESM1-2-HR (Jungclaus et al., 2019 ) MPI-M MPI-ESM1-2-LR (Wieners et al., 2019 ) MRI MRI-ESM2-0 (Yukimoto et al., 2019) NCAR CESM2-FV2 (Danabasoglu, 2019b ) NCAR CESM2-WACCM (Danabasoglu, 2019c ) NCAR CESM2 (Danabasoglu, 2019a ) NCC NorCPM1 (Bethke et al., 2019 ) NCC NorESM2-LM (Seland et al., 2019 ) NCC NorESM2-MM (Bentsen et al., 2019 ) NUIST NESM3 (Cao and Wang, 2019 ) SNU SAM0-UNICON (Park and Shin, 2019 ) The diversity of MSD events in each simulation is evaluated by classifying every detected event into one of the four reference clusters derived from ERA5 reanalysis data. The classification is based on the minimum Euclidean distance between the simulated event and the ERA5 cluster centroids. To ensure consistency, the ERA5 clusters are independently derived from ERA5 precipitation regridded to the same 1.0° × 1.0° resolution used for CMIP6 simulations. 3. Results The composite seasonal precipitation time series for each cluster and all MSD events are presented in Fig. 2 . Regions characterized by climatological MSD features display a canonical bimodal precipitation distribution, with a standard MSD event occurring from 23 June to 19 September and lasting 89 days (Fig. 2 a). During this period, a relatively shallow precipitation trough is observed (I msd = 0.31), corresponding to a 30% reduction in precipitation relative to the seasonal peaks. This reduction is weaker than the 40% decrease reported by Small et al. ( 2007 ) for the entire wet season, a discrepancy largely attributable to methodological differences, including the use of daily precipitation data in this study versus monthly data in theirs. Outputs from the cluster analysis reveal distinct variations in the bimodal seasonal precipitation cycle. Cluster 1 is characterized by a wetter seasonal cycle, with a relatively canonical bimodal structure but a weaker MSD intensity (I msd = 0.20) than in the climatological mean. Cluster 2 exhibits the longest MSD duration (126 days, from May to September) and a pronounced trough (I msd = 0.54), which is further distinguished by its asymmetric peaks, with the first precipitation maximum nearly 1.5 times larger than the second. In contrast, Cluster 3, which displays the strongest MSD intensity (I msd = 0.59), is characterized by the second peak clearly exceeding the first one. Cluster 4 resembles a compressed version of the climatological MSD, with moderate intensity (I msd = 0.39) and the shortest duration (70 days) (Fig. 2 a-b). Overall, the four clusters capture distinct modes of seasonal precipitation variability and MSD manifestation, demonstrating the effectiveness of cluster analysis for exploring the diversity of MSD events. To examine the spatiotemporal variability of MSD diversity, we mapped the number of calendar years exhibiting MSD events during 1979–2019 for each cluster (Fig. 3 ). Clear spatial heterogeneity is evident across Mesoamerica. MSD events in Cluster 1 predominantly occur south of 12°N, covering the Pacific coast of Central America as well as coastal and offshore regions of the eastern Pacific (Fig. 3 a). Cluster 2 shows a preference for Cuba, the coastal areas surrounding the Gulf of Mexico, and parts of the North Atlantic (Fig. 3 b). Cluster 3 is most pronounced over the Caribbean Sea, portions of the North Atlantic, the Yucatán Peninsula, and the Pacific coasts of Central America and Mexico (Fig. 3 c). Cluster 4 is mainly confined to regions north of 12°N, including southern Mexico and adjacent oceans. Clusters 1 and 4 together delineate the canonical MSD regions encompassing much of Central America and Mexico, while Clusters 2 and 3 highlight the occurrence of MSD events over the Caribbean and neighboring ocean basins. These results indicate that MSD diversity exhibits distinct spatial preferences, which is likely to reflect the differences in the underlying mechanisms which modulate regional precipitation variability. To investigate the mechanisms underlying MSD diversity, we examined atmospheric and oceanic conditions during MSD evolution, using the diagnostics described in Section 2.2 including SST, cloud fraction, and winds at 200 hPa and 850 hPa. SST and cloud fraction were selected as proxies for the SST–cloud feedback mechanism proposed to explain MSD onset and development (Herrera et al., 2015 , Magaña et al., 1999, Magaña and Caetano, 2005 ), with 200 hPa and 850 hPa winds selected as indicators for examing the upper- and lower-level atmospheric circulation, respectively. Anomalous composites were calculated by averaging these diagnostics across all MSD events in each cluster at key phases (onset, peak, and end), with low-level fields (SST, 850 hPa winds) shown in Fig. 4 and upper-level fields (cloud fraction, 200 hPa winds) in Fig. 5 . At both the onset and termination of the MSD, the diagnostics exhibit similar patterns: enhanced convection over canonical MSD regions, anomalous low-level westerlies from the Pacific contributing to convergence over the Yucatán Peninsula, and cooler SSTs in the adjacent eastern Pacific. This resemblance is consistent with the bimodal precipitation structure of the climatological MSD, which features two comparable rainfall peaks (Fig. 2 a). Such symmetric onset–end patterns are evident in Clusters 1 and 4, both of which closely resemble the canonical MSD (Fig. 2 a). In contrast, the diagnostic patterns diverge between the onset and end phases in Clusters 2 and 3, reflecting their non-canonical precipitation structures. In Cluster 2, low-level convergence over the Gulf of Mexico and the Caribbean Sea is evident at the onset but absent at the end (Fig. 4 ), consistent with stronger cloud anomalies and more vigorous convection at the onset than at the termination (Fig. 5 ). Cluster 3, however, maintains strong low-level convergence and active convection at both the onset and end. Notably, SSTs over the Caribbean Sea and Gulf of Mexico are warmer at the end (Fig. 5 ), suggesting an enhanced SST–cloud radiative feedback that may account for the stronger second precipitation peak characteristic of this cluster. At the MSD peak, distinct anomalies emerge, including anomalously warmer SSTs along the Pacific coasts, suppressed convection, and anomalous westerlies extending from the Caribbean Sea toward the Pacific (Fig. 4 ). These conditions contrast sharply with those observed at the onset and termination of the MSD. The opposing patterns between onset/end and peak suggest that MSD development may involve retreating processes, such as SST–cloud radiative feedback (Magaña et al., 1999) and variations associated with the solar declination angle (Karnauskas et al., 2013 ). Such contrasts are less evident in the MSD clusters (Fig. 4 ). For example, in Cluster 4, persistently cooler SSTs are observed along the Pacific coast of Central America throughout the MSD, while in Cluster 2, SST anomalies at the peak are only weakly positive relative to climatology. These findings point to potential variations in the mechanisms driving bimodal precipitation across different MSD clusters—variations that cannot be inferred solely from the climatological composite. The results presented above, together with previous studies, suggest that the SST and low-level winds act as key drivers of MSD. To further assess their roles, we examine the seasonal evolution of several important low-level flows—the CLLJ, the CHLLJ, and the EP wind flow—along with EP SST, and evaluate their coherence with MSD diversity. The CLLJ and EP wind are included because they have been suggested in previous literatures as either primary drivers or indicators of MSD events in canonical regions through their influence on moisture transport and regional convergence (Durán-Quesada et al., 2017 , García-Franco et al., 2023 , Magaña et al., 1999, Zhao et al., 2023 ). Although the CHLLJ has received comparatively less attention, earlier work has shown that its intensity is linked to seasonal moisture transport from the Pacific, implying a potential role in MSD development (Morales et al., 2017). EP SST is considered here as a proxy for SST–cloud radiative feedback. Indices for the CLLJ, CHLLJ, EP wind, and EP SST are constructed by averaging 925 hPa winds and SST within their respective domains (Fig. 1 a). To capture their seasonal variations associated with MSD diversity, the daily time series for each MSD year is averaged following the same procedure used to construct the MSD composites in Fig. 2 (see in Fig. 6 ). The CLLJ exhibits two distinct seasonal peaks, in March–May and September–November, respectively (Fig. 6 a), consistent with previous analyses (Cook and Vizy, 2010 , Wang, 2007 , Wang and Lee, 2007 ). There is a significant correlation between the CLLJ and MSD precipitation in clusters 2, 3, and 4, which broadly represent bimodal rainfall regimes over southern Mexico and the Caribbean (Fig. 6 e). This finding reinforces the critical role of the CLLJ in shaping MSD characteristics in these regions, while also underscoring its relatively weak influence on MSD precipitation in cluster 1, corresponding to Central America and the adjacent oceans. In MSD years, the CHLLJ displays a weak but discernible bimodal seasonal cycle, with a pronounced maximum from September to November and a secondary peak in May–July (Fig. 6 b). This pattern is consistent with the climatological seasonality of the CHLLJ reported in earlier studies (Rueda and Poveda, 2006 , Sierra et al., 2021 , Yepes et al., 2019 ). Notably, the CHLLJ shows significant correlations with MSD precipitation across all clusters of MSD diversity (Fig. 6 e). The seasonal cycle of the EP wind flow during MSD years exhibits a clear bimodal structure, with two westerly peaks in May–June and September–October (Fig. 6 c). This pattern closely resembles the bimodal rainfall structure of the MSD, as reflected in the significant correlations between EP wind anomalies and MSD precipitation across all diversity clusters (Fig. 6 e). In contrast, the EP SST seasonal cycle shows only a single distinct peak, with a modest warming in July–August followed by pronounced cooling at the canonical MSD termination in late August to September (Fig. 6 d). This result is consistent with García-Franco et al. ( 2023 ), who found similar features in pentad-mean SST, including weak bimodality and values below 29°C during the MSD. Correlations between EP SST and MSD precipitation are predominantly negative (clusters 3, 4, and all MSD events) but occasionally positive (clusters 1 and 2) (Fig. 6 e), underscoring that SST–cloud radiative feedback alone is insufficient to explain MSD dynamics. Overall, the diversity of MSD is not well represented in the seasonal cycles of the four selected indices, indicating that none of them individually resolves full mechanisms underlying MSD variability. Then, we would like to address the question, “How does the low-level wind flows influence the MSD diversity?”. Earlier research has contended that the impact of low-level wind patterns, particularly the CLLJ, on the MSD is effectively addressed by the convergence of moisture flux. To quantify the influence of the moisture flux convergence on the MSD, we use the formula of column-averaged moisture flux convergence (MFC) modifed from García‑Franco et al. (2023): \(\:MFC=--,\) where u and v indicate zonal and meridional wind flows, q indicates specific humidity, and indicates column average within troposphere. The MFC are calculated independently for each spatial grid and then averaged across corresponding MSD events to create composite representations of seasonal variation, either for all MSD events or MSD events assigned to a particular cluster. On the climatological scale, MFC shows a strong connection with MSD precipitation, as evidenced by their nearly identical bimodal seasonal cycles and significant correlations (Fig. 7 a). The bimodal structure of seasonal MFC during MSD years results from the combined influence of its zonal and meridional components: the zonal component contributes to moistening, whereas the meridional component exerts a drying effect. Even when MSD diversity is considered, the seasonal evolution of MFC continues to closely track MSD precipitation. In cluster 1, the annual MFC distribution is dominated by moistening from the zonal component, while the meridional component sustains mid-summer drying, producing a prolonged but weaker precipitation trough (Fig. 7 b). In clusters 2 and 3, where MSD events are concentrated over the Caribbean Sea, the bimodal MFC pattern is primarily driven by the meridional component, with the zonal component providing the background moistening (Figs. 7 c,d). In cluster 4, the seasonal MFC distribution resembles the canonical MSD structure, emerging from the combined effects of both zonal and meridional components. These results indicate that the bimodal structure of the MSD and its diversity are governed by the spatiotemporal variability of both the zonal and meridional components of MFC, rather than by individual zonal low-level jets such as the CLLJ or CHLLJ. Finally, we evaluate whether the CMIP6 models can reasonably reproduce the MSD and its diversity. We compare the simulated MSD frequencies during 1982–2014 with ERA5, focusing on both the overall occurrence and the diversity of MSDs, by mapping their spatial patterns and assessing the Pattern Correlation Coefficient (PCC) and Normalized Root Mean Square Error (NRMSE). From the perspective of all MSD events, the CMIP6 models capture the frequency patterns reasonably well, except for a notable underestimation over the Caribbean and adjacent regions. This is reflected by a mean PCC of 0.65 and a mean NRMSE of 0.88 estimated from multimodel ensemble mean (MMM), with 29 out of 33 models showing PCC ≥ 0.5 and NRMSE ≤ 1 (Fig. 8 c). In contrast, the models perform less consistently in simulating MSD diversity. For the two clusters characterized by symmetric bimodal distributions (Clusters 1 and 4), the simulated MSD frequencies are relatively well represented. Cluster 1 (4) shows mean PCC values of 0.58 (0.54) and mean NRMSE of 0.91 (0.92), with 20 (18) out of 33 models meeting the PCC ≥ 0.5 and NRMSE ≤ 1 criteria (Fig. 8 f, o). However, for the two asymmetric clusters (Clusters 2 and 3), model performance is substantially weaker. Cluster 2 (3) yields mean PCC values of 0.38 (0.40) and mean NRMSE of 1.21 (1.09), with only 0 (8) out of 33 models satisfying PCC ≥ 0.5 and NRMSE ≤ 1 (Fig. 8 i–l). Notably, the systematic underestimation of MSD frequency over the Caribbean in the all-model composite (Fig. 8 a–b) can be largely attributed to the poor simulation of Clusters 2 and 3 (Fig. 8 d, e, g, h). 4. Discussion and conclusions The Mesoamerican MSD has attracted considerable attention due to its profound influence on regional agriculture and socioeconomic activities. Numerous hypotheses have been proposed to explain the bimodal precipitation structure of the MSD, each emphasizing distinct local features. However, investigations on the diversity of MSD manifestations remains less explored, with most of the previous studies analyzed MSD mechanisms and characteristics using composite approaches. With the use of ERA5 reanalysis and CMIP6 model simulations, this study provides a comprehensive examination of the MSD diversity, including its characteristics, spatial distribution, and potential driving mechanisms. The four objectively identified clusters of MSD-related seasonal precipitation variations capture the breadth of MSD diversity. Each cluster exhibits distinct seasonal distributions and spatial preferences, encompassing a wide range of variations previously reported across southern Mexico, the Pacific and Caribbean slopes of Central America, the Caribbean basin, and adjacent oceanic regions. The clustering of MSD events based on daily data highlights the importance of fine-scale detection using sub-monthly observations (García-Franco et al., 2021; Zhao and Zhang, 2021 ), as it enables the identification of a sufficient number of events to train the clustering algorithm. Among the clusters, the pronounced asymmetry between the two precipitation peaks in the second and third clusters is particularly noteworthy, suggesting that MSD events in the Caribbean may arise from mechanisms distinct from those over continental Mesoamerica. Compared with the canonical MSD regions (e.g., García-Oliva et al., 2021; García-Franco et al., 2023 ), the Caribbean has received relatively little attention. Our results indicate that the MSD in this region cannot be neglected, as it exhibits unique features in both precipitation characteristics and underlying dynamics. Previous studies have highlighted the role of boreal summer sea surface temperatures (SSTs) and low-level wind patterns in shaping MSD evolution (Magaña et al., 1999, Magaña and Caetano, 2005 ). In contrast, our results show a weak connection between SSTs and MSD precipitation seasonality. Neither the climatic mean nor the seasonal evolution of eastern Pacific SSTs exhibits a bimodal signal consistent with MSD development, contrary to the hypothesis by Magaña et al. (1999), proposing a cooler SST (< 29°C) during the precipitation minimum period. Instead, anomalous warming or weak cooling is generally observed associated with precipitation reductions, suggesting that SST–cloud feedback is less influential than low-level wind dynamics, or that SST anomalies may reflect feedbacks from air–sea interactions driven by moisture convergence. This interpretation is consistent with recent findings using the same ERA5 dataset (García-Franco et al., 2023 ), which also point to a limited role of SSTs in MSD formation. The strong correspondence between seasonal bimodality in low-level wind indices and MSD precipitation emphasizes the central role of low-level jets. The CLLJ, a key feature of the Intra-Americas Seas circulation, has been widely recognized as a driver of MSD variability through its modulation of air–sea interaction and moisture convergence. Our results further suggest that other low-level flows, including the CHLLJ and easterly winds over the eastern Pacific, also play important roles. However, jet indices alone are insufficient to fully distinguish MSD diversity. By employing the MFC framework, we decomposed the contributions of zonal and meridional components of moisture transport. The results indicate that MFC reproduces the seasonal bimodal precipitation distribution and its variability across regions. MSD events over Central America, southern Mexico, and the eastern Pacific (clusters 1 and 4) are explained by combined zonal and meridional MFC components, consistent with wind patterns identified in earlier studies. In contrast, MSD events in the Caribbean are primarily shaped by meridional MFC, which induces the drying between precipitation peaks, while the zonal component provides a moist background. These findings highlight the need to pay greater attention to meridional moisture transport when analyzing MSD mechanisms in the Caribbean, where reliance on zonal jet indices such as the CLLJ alone may obscure important processes. Our results thus underscore that the CLLJ is one of several interacting drivers of MSD diversity, rather than its sole determinant. The CMIP6 evaluation further reveals that while models reproduce the overall MSD frequency reasonably well, they struggle to capture its diversity, particularly the asymmetric clusters prevalent over the Caribbean. This deficiency suggests that current models inadequately represent the regional moisture transport and associated air–sea interactions that underpin MSD variability. Improving the simulation of these processes is therefore critical for enhancing confidence in future projections of precipitation and its socioeconomic impacts across the Intra-Americas Seas. In conclusion, this study provides one of the first comprehensive analyses of MSD diversity, addressing its characteristics, underlying mechanisms, and representation in climate models, and thereby establishing a framework for future investigations. The findings highlight that assessments of climate model performance should explicitly consider MSD diversity, alongside mean states and long-term trends, since a model’s ability to reproduce distinct MSD types reflects its skill in capturing the spatiotemporal variability of precipitation systems in the Intra-Americas Seas. Advancing the understanding of MSD diversity, including its features and drivers, is essential for improving knowledge of regional precipitation variability and predictability, with direct implications for agriculture and socioeconomic planning. Future research should build on these results by integrating observational datasets and model experiments to develop a systematic framework for quantifying, evaluating, and exploring MSD diversity. Declarations 5. Data availability The ERA5 data used in this study are available at https://cds.climate.copernicus.eu/datasets , and the CMIP6 datasets can be accessed via https://esgf-node.ipsl.upmc.fr/projects/cmip6-ipsl/ . 6. Code availability Codes used to detect MSD events can be found via https://github.com/ZijieZhaoMMHW/MSD. 7. Funding information This work was supported by the ARC Centre of Excellence for Climate Extremes (CE17010023). 8. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References ALFARO EJ (2002) Some characteristics of the annual precipitation cycle in Central America and their relationships with its surrounding tropical oceans. Tópicos Meteorológicos y Oceanográficos 9:88–103 ALFARO EJ (2014) Caracterización del veranillo en dos cuencas de la vertiente del Pacífico de Costa Rica, América Central. Revista de Biología Trop 62:1–15 ALFARO EJ, HIDALGO H (2017) Propuesta Metodológica para la predicción climática estacional del veranillo en la cuenca del río Tempisque en el Pacífico Norte de Costa Rica. Tópicos Meteorológicos y Oceanográficos 16:62–74 AMADOR JA (2008) The intra-Americas sea low‐level jet: Overview and future research. Ann N Y Acad Sci 1146:153–188 ANDERSON TG, ANCHUKAITIS KJ, PONS D, TAYLOR M (2019) Multiscale trends and precipitation extremes in the Central American Midsummer Drought. Environ Res Lett 14:124016 BENTSEN M, TONIAZZO OLIVIÈDJLSELANDØ, GJERMUNDSEN T, GRAFF A, GUPTA LSDEBERNARDJB, HE AK, GRIESFELLER YKIRKEVÅGASCHWINGERJTJIPUTRAJAASKSBETHKEIFANY, GRINI J, GUO A, ILICAK C, LIAKKA MKARSETIHHLANDGRENOA, TANG JMOSEIDKONUMMELINASPENSBERGERC, ZHANG H (2019) In: SCHULZ M (ed) NCC NorESM2-MM model output prepared for CMIP6 CMIP historical. Z., HEINZE, C., IVERSEN. Earth System Grid Federation. BETHKE I, WANG Y, COUNILLON F, KIMMRITZ M, FRANSNER F, SAMUELSEN A, LANGEHAUG HR, BENTSEN CHIUP-G, GUO M, KIRKEVÅG CTJIPUTRAJ, SELAND AOLIVIÈDJL, ELDEVIK ØFANYLAWRENCEP, T., KEENLYSIDE N (2019) NCC NorCPM1 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation BOUCHER O, LEVAVASSEUR DENVILS, COZIC G, CAUBEL A, MEURDESOIF AFOUJOLSM-A, BALKANSKI Y, CHECA-GARCIA Y, R., HAUGLUSTAINE, D., BEKKI, S., MARCHAND M 2020. IPSL IPSL-CM5A2-INCA model output prepared for CMIP6 CMIP historical. Earth System Grid Federation BOUCHER O, LEVAVASSEUR DENVILS, COZIC G, CAUBEL A, MEURDESOIF AFOUJOLSM-A, BALKANSKI Y, CHECA-GARCIA Y, HAUGLUSTAINE R, D., BEKKI, S., MARCHAND M 2021. IPSL IPSL-CM6A-LR-INCA model output prepared for CMIP6 CMIP historical. Earth System Grid Federation BOUCHER O, LEVAVASSEUR DENVILS, COZIC G, CAUBEL A, FOUJOLS A, MEURDESOIF M-A, CADULE Y, DEVILLIERS P, LEBAS MGHATTASJ, LURTON N, MUSAT TMELLULL, I., MIGNOT, J., CHERUY F 2018. IPSL IPSL-CM6A-LR model output prepared for CMIP6 CMIP historical. Earth System Grid Federation CAO J, WANG B (2019) NUIST NESMv3 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation CENTELLA-ARTOLA A, BEZANILLA-MORLOT A, HERRERA TAYLORMA, MARTINEZ-CASTRO DA, SIERRA-LORENZO DGOUIRANDI, VICHOT-LLANO M, A., STEPHENSON, T., FONSECA C (2020) Evaluation of sixteen gridded precipitation datasets over the Caribbean region using gauge observations. Atmosphere , 11, 1334 CERÓN WL, ANDREOLI KAYANOMT, AVILA-DIAZ RV, DE A, SOUZA IP, SOUZA RA 2021a. Pacific and atlantic multidecadal variability relations with the choco and caribbean low-level jets during the 1900–2015 period. Atmosphere, 12, 1120 CERÓN WL, ANDREOLI KAYANOMT, CANCHALA RV, CARVAJAL-ESCOBAR T (2021b) Y. & ALFONSO-MORALES, W. Rainfall variability in southwestern Colombia: changes in ENSO-related features. Pure and Applied Geophysics , 178, 1087–1103 CHOUDHURY AD, RAGHAVAN K, NARAYANASETTI GOPINATHANPA, SINGH S (2019) In: MODI A (ed) CCCR-IITM IITM-ESM model output prepared for CMIP6 CMIP historical. M., PANICKAL CONSORTIUM EC-E (2019) EC-Earth-Consortium EC-Earth3 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation CONSORTIUM EC-E (2020a) EC-Earth-Consortium EC-Earth3-AerChem model output prepared for CMIP6 CMIP historical. Earth System Grid Federation CONSORTIUM EC-E (2020b) EC-Earth-Consortium EC-Earth3-Veg-LR model output prepared for CMIP6 CMIP historical. Earth System Grid Federation CONSORTIUM EC-E (2021) EC-Earth-Consortium EC-Earth-3-CC model output prepared for CMIP6 CMIP historical. Earth System Grid Federation COOK KH, VIZY EK (2010) Hydrodynamics of the Caribbean low-level jet and its relationship to precipitation. J Clim 23:1477–1494 CORRALES-SUASTEGUI A, FUENTES‐FRANCO R, PAVIA EG (2020) The mid‐summer drought over Mexico and Central America in the 21st century. Int J Climatol 40:1703–1715 DANABASOGLU G (2019a) NCAR CESM2 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation DANABASOGLU G (2019b) NCAR CESM2-FV2 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation DANABASOGLU G (2019c) NCAR CESM2-WACCM model output prepared for CMIP6 CMIP historical. Earth System Grid Federation DANEK C, SHI X, STEPANEK C, YANG H, BARBI D, HEGEWALD J, LOHMANN G (2020) IPCC DDC: AWI AWI-ESM1. 1LR model output prepared for CMIP6 CMIP historical DE SOUSA K, CASANOVES F, OSPINA SELLAREJ, AGUILAR ASUCHINIJG, A., MERCADO L (2018) How climate awareness influences farmers’ adaptation decisions in Central America? J Rural Stud 64:11–19 DILLEY M (1996) Synoptic controls on precipitation in the Valley of Oaxaca, Mexico. Int J Climatology: J Royal Meteorological Soc 16:1019–1031 DURÁN-QUESADA AM, GIMENO L, AMADOR J (2017) Role of moisture transport for Central American precipitation. Earth Sys Dyn 8:147–161 DURÁN-QUESADA AM, ORDOÑEZ SORÍR, P., GIMENO L (2020) Climate perspectives in the intra–Americas seas. Atmosphere 11:959 DURÁN-QUESADA AM, GIMENO L, AMADOR J, NIETO R (2010) Moisture sources for Central America: Identification of moisture sources using a Lagrangian analysis technique. J Geophys Research: Atmos, 115 EAKIN H, LERNER SWEENEYS, APPENDINI AM, PERALES K, DEWES HSTEIGERWALDDG, DAVENPORT CF, F., BAUSCH JC (2018) Agricultural change and resilience: Agricultural policy, climate trends and market integration in the Mexican maize system. Anthropocene 23:43–52 GALLEGO D, GARCÍA-HERRERA R, GÓMEZ-DELGADO FDP, ORDOÑEZ-PEREZ P, RIBERA P (2019) Tracking the moisture transport from the Pacific towards Central and northern South America since the late 19th century. Earth Sys Dyn 10:319–331 GAMBLE DW, PARNELL DB, CURTIS S (2008) Spatial variability of the Caribbean mid-summer drought and relation to north Atlantic high circulation. Int J Climatology: J Royal Meteorological Soc 28:343–350 GARCÍA-FRANCO JL, CHADWICK R, GRAY L, OSPREY S, ADAMS DK (2023) Revisiting mechanisms of the Mesoamerican Midsummer drought. Clim Dyn 60:549–569 HAJIMA T, ABE M, ARAKAWA O, SUZUKI T, KOMURO Y, OGURA T, OGOCHI K, WATANABE M, YAMAMOTO A, TATEBE H, OHGAITO NOGUCHIMA, ITO R, YAMAZAKI A, ITO D, TAKATA A, WATANABE K, S., KAWAMIYA, M., TACHIIRI K 2019. MIROC MIROC-ES2L model output prepared for CMIP6 CMIP historical. Earth System Grid Federation HARTIGAN JA, WONG MA (1979) Algorithm AS 136: A k-means clustering algorithm. J royal Stat Soc Ser c (applied statistics) 28:100–108 HASTENRATH S (1976) Variations in low-latitude circulation and extreme climatic events in the tropical Americas. J Atmos Sci 33:202–215 HELLIN J, COX R, LÓPEZ-RIDAURA S (2017) Maize diversity, market access, and poverty reduction in the Western Highlands of Guatemala. Mt Res Dev 37:188–197 HERRERA E, MAGAÑA V, CAETANO E (2015) Air-sea interactions and dynamical processes associated with the midsummer drought. Int J Climatol, 35 HERSBACH H, BELL B, BERRISFORD P, HIRAHARA S, MUÑOZ-SABATER HORÁNYIA, PEUBEY JNICOLASJ, RADU C, R., SCHEPERS D (2020) The ERA5 global reanalysis. Q J R Meteorol Soc 146:1999–2049 HIDALGO HG, DURÁN-QUESADA AM, AMADOR, J. A., ALFARO EJ (2015) The caribbean low‐level jet, the inter‐tropical convergence zone and precipitation patterns in the intra‐americas sea: a proposed dynamical mechanism. Geogr Annaler: Ser Phys Geogr 97:41–59 JUNGCLAUS J, SCHUPFNER BITTNERMWIENERSK-HWACHSMANNF, LEGUTKE M, GIORGETTA S, REICK M, GAYLER C, HAAK V, DE VRESE H, RADDATZ P, ESCH T, MAURITSEN M, CLAUSSEN TVONSTORCHJ-SBEHRENSJBROVKINV, CRUEGER M, FAST T, FIEDLER I, HAGEMANN S, HOHENEGGER S, JAHNS C, KLOSTER T, KINNE S, LASSLOP S, KORNBLUEH G, MATEI LMAROTZKEJ, MERANER D, K., ROECKNER E (2019), MIKOLAJEWICZ, U., MODALI, K., MÜLLER, W., NABEL, J., NOTZ, D., PETERS-VON GEHLEN, K., PINCUS MPI-M MPI-ESM1.2-HR model output prepared for CMIP6 CMIP historical. Earth System Grid Federation. KARNAUSKAS KB, GIANNINI SEAGERR, A., BUSALACCHI AJ (2013) A simple mechanism for the climatological midsummer drought along the Pacific coast of Central America. Atmósfera 26:261–281 LEE W, LIANG H (2020) AS-RCEC TaiESM1. 0 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation LI L (2019) CAS FGOALS-g3 model output prepared for CMIP6 CMIP historical. (No Title) LOAIZA CERON W, FERREIRA ANDREOLIRVKAYANOMT, DE SOUZA RA, JONES C, CARVALHO LM (2020) The influence of the atlantic multidecadal oscillation on the choco low-level jet and precipitation in colombia. Atmosphere 11:174 LOVATO T, PEANO D (2020) CMCC CMCC-CM2-SR5 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation LOVATO T, PEANO D, BUTENSCHÖN M (2021) CMCC CMCC-ESM2 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation MAGAÑA V, AMADOR, J. A., MEDINA S (1999) The midsummer drought over Mexico and Central America. J Clim 12:1577–1588 MAGAÑA V, CAETANO E (2005) Temporal evolution of summer convective activity over the Americas warm pools. Geophys Res Lett, 32 MALDONADO T, RUTGERSSON A, ALFARO E, AMADOR J, CLAREMAR B (2016) Interannual variability of the midsummer drought in Central America and the connection with sea surface temperatures. Adv Geosci 42:35–50 MAPES BE, LIU P, BUENNING N (2005) Indian monsoon onset and the Americas midsummer drought: Out-of-equilibrium responses to smooth seasonal forcing. J Clim 18:1109–1115 MARTINEZ C, GODDARD L, KUSHNIR Y, TING M (2019) Seasonal climatology and dynamical mechanisms of rainfall in the Caribbean. Clim Dyn 53:825–846 MAURER EP, STEWART IT, JOSEPH, K., HIDALGO HG (2022) The Mesoamerican mid-summer drought: the impact of its definition on occurrences and recent changes. Hydrol Earth Syst Sci 26:1425–1437 MEJÍA JF, POVEDA YEPESJHENAOJJ, ZULUAGA G, RAYMOND MD, D. J., FUCHS-STONE Ž (2021) Towards a mechanistic understanding of precipitation over the far eastern tropical Pacific and western Colombia, one of the rainiest spots on Earth. J Geophys Research: Atmos, 126, e2020JD033415. MORALES J, ARIAS, P. A., MARTÍNEZ JA (2017) Role of the Caribbean low-level jet and the Choco jet in the patterns of atmospheric moisture transport towards Central America. Sciforum Electronic Conference Series MORALES-VELÁZQUEZ MI, RAFIEEINASAB HERRERAGDSAPARICIOJ, A., LOBATO-SÁNCHEZ R (2021) Evaluating reanalysis and satellite-based precipitation at regional scale: A case study in southern Mexico. Atmósfera 34:189–206 MOSIÑO A, GARCÍA E (1966) Evaluación de la sequía intraestival en la República Mexicana. Proc. Conf. Reg. Latinoamericana Unión Geogr. Int, 500–516 NEUBAUER D, SIEGENTHALER-LE DRIAN FERRACHATS, STOLL C, FOLINI J, TEGEN DS, STEMMLER IWIENERSK-HMAURITSENT, BARTHEL I, BEY S, DASKALAKIS I, KOKKOLA NHEINOLDB, PARTRIDGE H, RAST D, SCHMIDT S, WATSON-PARRIS H, D., LOHMANN U (2019) HAMMOZ-Consortium MPI-ESM1.2-HAM model output prepared for CMIP6 CMIP historical. Earth System Grid Federation PARK S, SHIN J (2019) SNU SAM0-UNICON model output prepared for CMIP6 CMIP historical. Earth System Grid Federation PERDIGÓN-MORALES J, ROMERO-CENTENO R, ORDOÑEZ P, NIETO R, GIMENO L, BARRETT BS (2021) Influence of the Madden-Julian oscillation on moisture transport by the caribbean low level jet during the midsummer drought in Mexico. Atmos Res 248:105243 PERDIGÓN-MORALES J, ROMERO‐CENTENO R, PÉREZ PO, BARRETT BS (2018) The midsummer drought in Mexico: perspectives on duration and intensity from the CHIRPS precipitation database. Int J Climatol 38:2174–2186 PONS D, GRIFFIN TAYLORMJ, CASTELLANOS D, E. J., ANCHUKAITIS KJ (2017) On the production of climate information in the high mountain forests of Guatemala. Annals Am Association Geographers 107:323–335 POVEDA G, MESA OJ (2000) On the existence of Lloró (the rainiest locality on Earth): Enhanced ocean-land‐atmosphere interaction by a low‐level jet. Geophys Res Lett 27:1675–1678 RAUSCHER SA, GIORGI F, DIFFENBAUGH, N. S., SETH A (2008) Extension and intensification of the Meso-American mid-summer drought in the twenty-first century. Clim Dyn 31:551–571 RUEDA OA, POVEDA G (2006) Spatial and temporal variability of the Choco jet stream and its effect on the hydroclimatology of the Colombian pacific; variabilidad espacial y temporal del chorro del Choco y su efecto en la hidroclimatologia de la region del pacifico colombiano. Meteorologia Colombiana SCOCCIMARRO E, BELLUCCI A, PEANO D (2020) CMCC CMCC-CM2-HR4 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation SELAND Ø, BENTSEN M, TONIAZZO OLIVIÈDJL, GJERMUNDSEN T, GRAFF A, GUPTA LSDEBERNARDJB, HE AK, GRIESFELLER YKIRKEVÅGASCHWINGERJTJIPUTRAJAASKSBETHKEIFANY, GRINI J, GUO A, ILICAK C, LIAKKA MKARSETIHHLANDGRENOA, TANG JMOSEIDKONUMMELINASPENSBERGERC, ZHANG H (2019) In: SCHULZ M (ed) NCC NorESM2-LM model output prepared for CMIP6 CMIP historical. Z., HEINZE, C., IVERSEN. Earth System Grid Federation. SIERRA JP, ARIAS PA, DURÁN-QUESADA AM, TAPIAS, K. A., VIEIRA, S. C., MARTÍNEZ JA (2021) The Choco low-level jet: past, present and future. Clim Dyn 56:2667–2692 SMALL RJO, DE SZOEKE SP, XIE S-P (2007) The Central American midsummer drought: regional aspects and large-scale forcing. J Clim 20:4853–4873 SWART NC, KHARIN COLEJNS, LAZARE VV, SCINOCCA M, GILLETT JF, ARORA NPANSTEYJ, MAJAESS VCHRISTIANJRJIAOYLEEWG, SEILER FSAENKOOA, SEINEN C, SHAO C, SOLHEIM A, VON L, SALZEN K, YANG, D., WINTER, B., SIGMOND M (2019) CCCma CanESM5 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation TATEBE H, WATANABE M (2018) MIROC MIROC6 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation WANG C (2007) Variability of the Caribbean low-level jet and its relations to climate. Clim Dyn 29:411–422 WANG C, LEE SK (2007) Atlantic warm pool, Caribbean low-level jet, and their potential impact on Atlantic hurricanes. Geophys Res Lett, 34 WIENERS K-H, ESCH GIORGETTAMJUNGCLAUSJREICKC, BITTNER M, LEGUTKE M, SCHUPFNER S, WACHSMANN M, GAYLER F, HAAK V, DE VRESE H, RADDATZ P, MAURITSEN T, CLAUSSEN TVONSTORCHJ-SBEHRENSJBROVKINV, CRUEGER M, FAST T, FIEDLER I, HAGEMANN S, HOHENEGGER S, JAHNS C, KLOSTER T, KINNE S, LASSLOP S, KORNBLUEH G, MATEI LMAROTZKEJ, MERANER D, K., ROECKNER E (2019), MIKOLAJEWICZ, U., MODALI, K., MÜLLER, W., NABEL, J., NOTZ, D., PETERS-VON GEHLEN, K., PINCUS MPI-M MPI-ESM1.2-LR model output prepared for CMIP6 CMIP historical. Earth System Grid Federation. WYSER K, KJELLSTRÖM E, KOENIGK T, MARTINS H, DÖSCHER R (2020) Warmer climate projections in EC-Earth3-Veg: the role of changes in the greenhouse gas concentrations from CMIP5 to CMIP6. Environ Res Lett 15:054020 XIN X, ZHANG ZHANGJ, WU F, SHI T, LI X, LIU JCHUM, Q., YAN, J., MA Q (2018) BCC BCC-CSM2MR model output prepared for CMIP6 CMIP YEPES J, POVEDA G, MORENO MEJÍAJF, L., RUEDA C (2019) Choco-jex: A research experiment focused on the Chocó low-level jet over the far eastern Pacific and western Colombia. Bull Am Meteorol Soc 100:779–796 YU Y (2019) CAS FGOALS-f3-L model output prepared for CMIP6 CMIP historical YUKIMOTO S, KOSHIRO T, KAWAI H, URAKAWA OSHIMANYOSHIDAK, TSUJINO S, DEUSHI H, TANAKA M, HOSAKA T, YOSHIMURA M, SHINDO H, ISHII EMIZUTAR M., OBATA, A. & ADACHI Y 2019. MRI MRI-ESM2.0 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation ZHANG J, WU T, SHI X, ZHANG F, LI J, CHU M, MA LIUQYANJ, Q., WEI M 2018. BCC BCC-ESM1 model output prepared for CMIP6 CMIP. (No Title) ZHAO Z, HAN M, YANG K, HOLBROOK NJ (2023) Signatures of midsummer droughts over Central America and Mexico. Clim Dyn 60:3523–3542 ZHAO Z, HOLBROOK NJ, BALLESTERO OLIVEREC, D., VARGAS-HERNANDEZ JM (2020) Characteristic atmospheric states during mid-summer droughts over Central America and Mexico. Clim Dyn 55:681–701 ZHAO Z, ZHANG X (2021) Evaluation of methods to detect and quantify the bimodal precipitation over Central America and Mexico. Int J Climatol 41:E897–E911 Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Major Revision 15 Dec, 2025 Reviewers agreed at journal 23 Sep, 2025 Reviewers invited by journal 20 Sep, 2025 Editor assigned by journal 20 Sep, 2025 First submitted to journal 18 Sep, 2025 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-7652378","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":518047072,"identity":"6bc6b259-5af1-45f6-96eb-287dc6b0cb71","order_by":0,"name":"Zijie Zhao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/ElEQVRIiWNgGAWjYDACZoaEA0AqAcg6ABVKIFoLWwLDAaK0IEzmMSBOi3w7w8MDHxjq8vjbe75Jf6ipY+BnzzFg+NmGWwtjM0PCwRkMbMUSZ85ukzhw7DCDZM8bA8ZePFqYgX45zMPAk9hwIxeohe0Ag8ENoC28eLSwgbT8YZBInH8j55nEgX91DPZALYx/8WjhAWlhYDBI3HAjh03iYBszg4FEjgEzPlskgFoO9hgkJG48c8zY4mzfYR6JM88KDsucw61Fvv9M8ocfFXWJ8443P7xR8a1Ojr89eePDN2W4tQCdlgB0FxIXRBzAp4GBgZ2A/CgYBaNgFIwCACZyUgu4zZ5yAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-3403-878X","institution":"University of California Irvine","correspondingAuthor":true,"prefix":"","firstName":"Zijie","middleName":"","lastName":"Zhao","suffix":""},{"id":518047073,"identity":"a064fbfa-1eb2-403b-bd3b-e740ebc0b529","order_by":1,"name":"Yanxuan Du","email":"","orcid":"","institution":"The University of Melbourne","correspondingAuthor":false,"prefix":"","firstName":"Yanxuan","middleName":"","lastName":"Du","suffix":""}],"badges":[],"createdAt":"2025-09-18 18:59:51","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7652378/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7652378/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92580128,"identity":"41c1435e-1915-44e0-bb08-0e3e46898283","added_by":"auto","created_at":"2025-10-01 09:10:51","extension":"xml","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7802,"visible":true,"origin":"","legend":"","description":"","filename":"cldyCLDYD2501021.xml","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/425150d8a767d3fb9bae74bd.xml"},{"id":92578972,"identity":"bce867e5-6eb4-4cf7-b902-ef6354655abf","added_by":"auto","created_at":"2025-10-01 08:54:50","extension":"xml","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1211,"visible":true,"origin":"","legend":"","description":"","filename":"CLDYD250102121923.go.xml","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/4d78a926a6f7cdbd5a4ffa16.xml"},{"id":92578971,"identity":"fcf88ed8-0ad2-4d20-8e0a-3596274a352c","added_by":"auto","created_at":"2025-10-01 08:54:50","extension":"xml","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":733,"visible":true,"origin":"","legend":"","description":"","filename":"CLDYD2501021Import.xml","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/38ca0a537259b772aeaa4590.xml"},{"id":92578979,"identity":"ab9bdd5c-d94e-4bc6-a194-9f7f9fe8df69","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"xml","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":178358,"visible":true,"origin":"","legend":"","description":"","filename":"CLDYD25010210enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/fefee9404686b92fa6e10aa4.xml"},{"id":92578977,"identity":"25580631-e6a2-4bb2-a2db-e7e750567588","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"eps","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1662179,"visible":true,"origin":"","legend":"","description":"","filename":"Fig1.eps","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/f67d76c1bc2158782f7450e3.eps"},{"id":92578986,"identity":"c2a065d2-7f21-4ad7-935d-9eb2b3ee85c6","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"eps","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":403467,"visible":true,"origin":"","legend":"","description":"","filename":"Fig2.eps","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/5e1255322ab19aac4bd799c4.eps"},{"id":92578983,"identity":"0b7559af-33c8-4bf0-b047-8022464c0419","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"eps","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":898740,"visible":true,"origin":"","legend":"","description":"","filename":"Fig3.eps","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/2ef959df5a7f3c76db7641de.eps"},{"id":92578978,"identity":"1270fe83-2108-4063-b255-130532502979","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"eps","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1534919,"visible":true,"origin":"","legend":"","description":"","filename":"Fig4.eps","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/81744650c197a376280cb4c3.eps"},{"id":92578980,"identity":"e70e6cd2-870d-44b6-b55e-4334af913b9c","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"eps","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2009772,"visible":true,"origin":"","legend":"","description":"","filename":"Fig5.eps","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/50d7644977ef830694a1adcb.eps"},{"id":92578993,"identity":"1d530502-d85d-4abc-9fec-bd7fcd0fefb3","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"eps","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":539044,"visible":true,"origin":"","legend":"","description":"","filename":"Fig6.eps","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/332f22cdc1e922dd6d199d28.eps"},{"id":92578988,"identity":"0e8cb8b9-042f-45e8-ac37-f70842ef3c82","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"eps","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":136905,"visible":true,"origin":"","legend":"","description":"","filename":"Fig7.eps","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/79c1d58f077af2b54891bea2.eps"},{"id":92578987,"identity":"469dba2d-7608-4ff6-a44f-4fc378c59929","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"eps","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1915468,"visible":true,"origin":"","legend":"","description":"","filename":"Fig8.eps","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/159c7d6e37bd2920e961c2d8.eps"},{"id":92580129,"identity":"6a68ddd2-8f5c-48b4-962f-e39a913e3417","added_by":"auto","created_at":"2025-10-01 09:10:51","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1682343,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/7e9cc8220b58ef81e6b0cdb6.png"},{"id":92580418,"identity":"284ad996-8494-43bd-af09-058305f38aaa","added_by":"auto","created_at":"2025-10-01 09:18:51","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":314787,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/6b10658c3af8ed0ed2f40fd2.png"},{"id":92580132,"identity":"18ff929d-edfc-4b56-a69d-0b08a5af248f","added_by":"auto","created_at":"2025-10-01 09:10:51","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":711470,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/4a00b4fea96d163eb044d1c5.png"},{"id":92580130,"identity":"d64d0de5-d166-44c5-a2f0-d02d7d0b4f7a","added_by":"auto","created_at":"2025-10-01 09:10:51","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1549697,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/002e846cd505ebaab312ff4e.png"},{"id":92579501,"identity":"437603ee-4863-42be-9a33-be949be8c0f1","added_by":"auto","created_at":"2025-10-01 09:02:51","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2201663,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/6a5334640ee4c261dd8dda0c.png"},{"id":92579504,"identity":"7b101b2b-3055-4d47-8f66-599fcd0ed086","added_by":"auto","created_at":"2025-10-01 09:02:51","extension":"png","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":514660,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/c1707d33e914ba570eff9575.png"},{"id":92579004,"identity":"bad7167c-4641-465c-aed4-901448a1f1a0","added_by":"auto","created_at":"2025-10-01 08:54:52","extension":"png","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":445488,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/6c0beecb99089979e23dd18e.png"},{"id":92579507,"identity":"67b222ee-7d48-479c-b0a9-bcf4bbd75fff","added_by":"auto","created_at":"2025-10-01 09:02:51","extension":"png","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2002010,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/e8d201f4d0e5e2071b87df61.png"},{"id":92579506,"identity":"7406991a-42a8-499c-af67-87529c6d803d","added_by":"auto","created_at":"2025-10-01 09:02:51","extension":"png","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":368986,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/3d79bd62acab431a8e9b70c5.png"},{"id":92578989,"identity":"4c1756a6-8ed6-4905-973b-90649dbad59a","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"png","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":158630,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/a9aeb5c58213cf139bf1762f.png"},{"id":92578995,"identity":"20c7d9c8-7c12-4dd5-9777-3ea9abd3b6f8","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"png","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":392627,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/609bf575a5704e4aa30587cc.png"},{"id":92579000,"identity":"e228e6ef-49cc-462d-ad18-80b5c6d2edf6","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"png","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":811340,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/a2139076296f835ea4b5f77f.png"},{"id":92578997,"identity":"5aaf4503-905d-4dd1-a9a8-abd61b91dd8b","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"png","order_by":34,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1100143,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/95bf51207422edbe1e07d760.png"},{"id":92578990,"identity":"1d950589-2773-424a-87b5-152ab4f89de1","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"png","order_by":35,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":219878,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/0547b0f00c6e3cfb57fa5b88.png"},{"id":92579509,"identity":"3cc0dd93-bdb2-43bc-9d8a-0817524dfdaf","added_by":"auto","created_at":"2025-10-01 09:02:51","extension":"png","order_by":36,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":194961,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/b4a97a4cbc37a4ecf18998f6.png"},{"id":92578996,"identity":"7d095ae2-a659-4967-ba7c-5a896406352f","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"png","order_by":37,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":698995,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/5e3f8fa1e5089eb273dc606d.png"},{"id":92578998,"identity":"328b10f6-8bf5-4877-9568-dff023c673c6","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"xml","order_by":38,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":176637,"visible":true,"origin":"","legend":"","description":"","filename":"CLDYD25010210structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/944cb5ee2f2f8b601d4b460f.xml"},{"id":92578999,"identity":"102095af-2e04-42fd-9db2-365daebe6644","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"html","order_by":39,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":184956,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/90899f2076dfee6a4df516c9.html"},{"id":92579497,"identity":"c14047c4-fb63-4679-acb5-2e687072a90a","added_by":"auto","created_at":"2025-10-01 09:02:51","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1682343,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e Topography of the domain analyzed in this study includes southern Mexico, Central America, the Caribbean, northern South America, and adjacent seas such as the Caribbean Sea and the eastern Pacific. Black boxes indicate the domain used to calculate indices for CLLJ, EP SST and EP wind. \u003cstrong\u003eb\u003c/strong\u003e Schematic representation of a canonical precipitation seasonal cycle with the presence of the MSD. The periods of the MSD are shaded in orange, with its onset and end indicated by red and blue lines. Two shaded lines indicate the two linear regressions used to identify the existence of the MSD, as described in Section 2.2. \u003cstrong\u003ec\u003c/strong\u003eFrequency of MSD occurrences during 1979-2019, with colours indicating regions exhibiting climatological MSD features.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/28181a4d5f6f19ab4db4389d.png"},{"id":92580126,"identity":"0eb2a516-d6ee-4eb6-8d2d-6f87d77e5a88","added_by":"auto","created_at":"2025-10-01 09:10:50","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":314787,"visible":true,"origin":"","legend":"\u003cp\u003eComposites of seasonal precipitation cycle for each cluster and all MSD events using \u003cstrong\u003ea\u003c/strong\u003e raw and \u003cstrong\u003eb\u003c/strong\u003e normalized precipitation time series. These composites are calculated by averaging precipitation time series corresponding to MSD events assigned to a particular cluster. Dashed lines indicate the onset and end dates of a particular MSD diversity or climatological MSD, isolating the MSD period. \u003cstrong\u003ec\u003c/strong\u003e The inter-cluster correlations among the 4 clusters representing the MSD diversity derived from \u003cstrong\u003ea\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/1d42010fcd389bf90ed67a34.png"},{"id":92578966,"identity":"a0e93a01-f058-42ef-b41e-4f4abe579647","added_by":"auto","created_at":"2025-10-01 08:54:50","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":711470,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ea-d\u003c/strong\u003e Spatial distribution of the 4 clusters of MSD diversity (K1-4). Colours indicate counts of years with the presence of MSD during 1979-2019 (41 years). Percentages of MSD events assigned to each cluster is labelled in titles of each panel.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/aaede78299eedc7f5d302d65.png"},{"id":92579500,"identity":"16cdb5f2-b946-47b6-a021-b3ed93588f9c","added_by":"auto","created_at":"2025-10-01 09:02:51","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1549697,"visible":true,"origin":"","legend":"\u003cp\u003eComposites of low-level climate properties during the development of MSD diversity. Colours indicate SST anomalies while arrows indicate 850hPa wind anomalies. Each row represents composite mean from a cluster of MSD diversity (K1-4) or all MSD events, while each column indicates a particular date (onset, peak, or end date) during the development of MSD.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/fdadcb37d7db4f7dc99cff49.png"},{"id":92578975,"identity":"b0185dcb-0a95-4648-b807-5b50d86b2376","added_by":"auto","created_at":"2025-10-01 08:54:51","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2201663,"visible":true,"origin":"","legend":"\u003cp\u003eComposites of high-level climate properties during the development of MSD diversity. Colours indicate cloud fraction anomalies while arrows indicate 200hPa wind anomalies. Each row represents composite mean from a cluster of MSD diversity (K1-4) or all MSD events, while each column indicates a particular date (onset, peak, or end date) during the development of MSD.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/b5e7e163e7c9418a57d3a00b.png"},{"id":92578967,"identity":"fd9cb8b2-31a9-47f4-8a7a-ff17c05f39ab","added_by":"auto","created_at":"2025-10-01 08:54:50","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":514660,"visible":true,"origin":"","legend":"\u003cp\u003eComposites of daily mean seasonal cycle of \u003cstrong\u003ea\u003c/strong\u003ethe CLLJ, \u003cstrong\u003eb\u003c/strong\u003e the CHLLJ, \u003cstrong\u003ec\u003c/strong\u003e the EP wind flow, and \u003cstrong\u003ed\u003c/strong\u003e the EP SST. Lines in each panel a-d indicate result from MSD events assigned to a particular cluster (K1-4) or all MSD events. \u003cstrong\u003ee\u003c/strong\u003e The correlations between seasonal precipitation cycle (Fig 2a) and the 4 indices during the MSD period, which are indicated by dashed lines for each MSD cluster and all MSD events in Fig 2a. Staistical significance in the 95% confidence interval is marked by stars.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/31ddc1aa5051369a588a0403.png"},{"id":92579495,"identity":"ef36cc22-a0af-4732-96e7-586537a762fd","added_by":"auto","created_at":"2025-10-01 09:02:50","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":445488,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ea- e\u003c/strong\u003e Composites of daily mean seasonal cycle of the MFC. Each panel indicates results from \u003cstrong\u003ea\u003c/strong\u003e all MSD events or \u003cstrong\u003eb-e\u003c/strong\u003e a particular cluster of MSD diversity (K1-4). Percentages of MSD events assigned to each cluster is labelled in titles of each panel.\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/6505fb3dd4197f589586b81d.png"},{"id":92580417,"identity":"7b1cb0f8-aeb7-4dd7-adfc-e8beb1bf2f79","added_by":"auto","created_at":"2025-10-01 09:18:50","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":2002010,"visible":true,"origin":"","legend":"\u003cp\u003eEvaluation of MSD simulation skill in 33 CMIP6 models for 1979–2014. The first two columns show \u003cstrong\u003ea, d, g, j, m\u003c/strong\u003e observed MSD frequency distributions and \u003cstrong\u003eb, e, h, k, n\u003c/strong\u003ethe corresponding multi-model composite for all events and for each MSD cluster. The third column presents the PCC and NRMSE of each individual model relative to observations, with the solid line denoting the linear fit and the Pearson correlation coefficient (r) given alongside. Note that the observed MSD frequencies in panels (a, d, g, j, m) differ from those in Fig 2, as here the events are independently identified from the regridded 1° ERA5 data without constraining grid cells by climatological MSD features (see Section. 2.3).\u003c/p\u003e","description":"","filename":"image8.png","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/0656fe725acf9b3c209585c1.png"},{"id":92581369,"identity":"b059a9d7-5107-4ac1-a21c-7ac206253253","added_by":"auto","created_at":"2025-10-01 09:27:00","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":8182041,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7652378/v1/c56538c8-330b-4600-bcaf-9179b5dd04ff.pdf"}],"financialInterests":"","formattedTitle":"Diversity of Mesoamerican Midsummer Drought","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eA large portion of the Intra-Americas Seas region\u0026mdash;including Central America, southern Mexico, the Caribbean, and adjacent oceanic areas\u0026mdash;exhibits a bimodal annual precipitation cycle, typically marked by a relative \u0026ldquo;trough\u0026rdquo; within the canonical May\u0026ndash;October rainy season (Dur\u0026aacute;n-Quesada et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Garc\u0026iacute;a-Franco et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, Maga\u0026ntilde;a et al., 1999, Mosi\u0026ntilde;o and Garc\u0026iacute;a, \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e1966\u003c/span\u003e). This feature, commonly referred to as the Midsummer Drought (MSD) (Maga\u0026ntilde;a et al., 1999) or veranillo in local terminology (Alfaro, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2002\u003c/span\u003e, Alfaro, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2014\u003c/span\u003e, Dilley, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e1996\u003c/span\u003e), exerts substantial socioeconomic impacts. In particular, it has been linked to agricultural losses (de Sousa et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, Eakin et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, Hellin et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, Pons et al., \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) and to notable shifts in its timing and spatial extent over recent decades (Anderson et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, Corrales-Suastegui et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Rauscher et al., \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Given its regional importance, it is critical to advance our understanding of MSD characteristics and underlying mechanisms for improved seasonal precipitation prediction.\u003c/p\u003e\u003cp\u003eA canonical seasonal precipitation cycle associated with the MSD is characterized by two distinct rainfall peaks: the first between May and July and the second between August and October, separated by a relative reduction in precipitation (Alfaro, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2002\u003c/span\u003e, Alfaro and Hidalgo, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, Garc\u0026iacute;a-Franco et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, Maga\u0026ntilde;a et al., 1999, Maurer et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). However, both the bimodal structure and the timing and intensity of the MSD exhibit pronounced spatiotemporal variability. Strong MSD signatures have been identified along the Pacific side of Central America, the Yucat\u0026aacute;n Peninsula, parts of the Pacific coast of southern Mexico, and across several Caribbean islands, including Cuba and Haiti (Alfaro, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2002\u003c/span\u003e, Alfaro, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2014\u003c/span\u003e, Gamble et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2008\u003c/span\u003e, Maga\u0026ntilde;a et al., 1999, Perdig\u0026oacute;n-Morales et al., \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). However, the occurrence of evident MSD signals along the Caribbean side of Central America remains less certain While some studies report only weak evidence of MSD in this region, others suggest that portions of the Caribbean coast may display a bimodal annual rainfall distribution (Karnauskas et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2013\u003c/span\u003e, Maldonado et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). In the latter cases, however, the second peak often occurs in November, extending beyond the canonical May\u0026ndash;October rainy season (Zhao and Zhang, \u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Spatial variability is further evident in the increasing duration and intensity of the MSD toward the southeast along the Pacific side of southern Mexico and Central America (Anderson et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, Zhao et al., \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Projections for the 21st century suggest substantial changes in MSD characteristics\u0026mdash;including their timing, intensity, and spatial extent\u0026mdash;indicating a potential emergence of enhanced spatial variability in MSD under future climate conditions. (Corrales‐Suastegui et al., 2020, Rauscher et al., \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). This pronounced spatiotemporal variability underscores the complexity of characterizing the MSD, suggesting that a pure climatological definition based on canonical bimodal precipitation may be insufficient to capture its full range of expressions.\u003c/p\u003e\u003cp\u003eThe complexity of the MSD is further reflected in extensive efforts to elucidate its underlying mechanisms in literatures. Since the initial hypothesis that the two rainfall peaks of the MSD correspond to the double passage of the Intertropical Convergence Zone (ITCZ) (Hastenrath, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e1976\u003c/span\u003e), numerous mechanisms have been proposed to explain the origin of its bimodal precipitation structure. Among them, the SST\u0026ndash;cloud radiative feedback mechanism proposed by Maga\u0026ntilde;a et al. (1999) emphasizes the role of Eastern Pacific (EP) sea surface temperature (SST) variability and associated cloud feedbacks. In this framework, early-summer SST warming enhances cloudiness through strong radiative effects, producing the first precipitation peak. The resulting increase in cloud cover reduces shortwave radiation while strengthened low-level winds enhance latent heat flux, together leading to SST cooling and suppressed convection, consistent with the midsummer reduction in rainfall. Subsequently, weakened convection allows greater solar insolation, that warms the SST again and supports a second rainfall peak. This mechanism was later refined by Maga\u0026ntilde;a and Caetano (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), who documented a bimodal SST structure resembling the MSD during the boreal summer of 2001. The SST\u0026ndash;cloud radiative feedback implies that seasonal forcing alone can sustain recurrent MSD events, whereas their interannual diversity arises primarily from remote forcings. However, the correspondence between SST and MSD remains ambiguous: it is less evident in reanalysis products and climate model simulations, and SST variability does not fully resolve the observed precipitation seasonality associated with the MSD (Garc\u0026iacute;a-Franco et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eBeyond eastern Pacific SST, the MSD has also been linked to low-level pressure gyres and their associated wind circulations (Amador, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Seasonal variations of the North Atlantic Subtropical High (NASH) modulate the strength of the Caribbean Low-Level Jet (CLLJ) (Mapes et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), giving rise to its bimodal annual cycle (Wang, \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2007\u003c/span\u003e, Wang and Lee, \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). The CLLJ is considered a key driver of moisture transport from the Caribbean into the canonical MSD region and plays a central role in regulating convective activity during MSD evolution through moisture accumulation and dissipation (Corrales-Suastegui et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Gamble et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2008\u003c/span\u003e, Hidalgo et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2015\u003c/span\u003e, Martinez et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, Small et al., \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). In addition to its role in shaping the year-to-year occurrence of the MSD, the CLLJ has been shown to influence precipitation variability across multiple time scales, including decadal (Cer\u0026oacute;n et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2021b\u003c/span\u003e) and intraseasonal (Perdig\u0026oacute;n-Morales et al., \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) variations over Central America.\u003c/p\u003e\u003cp\u003eWhile the CLLJ plays a central role in modulating regional circulation, other low-level wind systems also contribute to precipitation variability across Mesoamerica and the Caribbean, most notably the eastern Pacific trade winds and the Choc\u0026oacute; Low-Level Jet (CHLLJ). Variations in the eastern Pacific trade winds have been recognized as indicators of MSD timing and duration, as they modulate surface latent heat fluxes and may influence the development of low-level convergence associated with the first precipitation peak. The CHLLJ, in turn, serves as an important conveyor of moisture from the Pacific into Central America and Colombia, intensifying eastward during the rainy season (Cer\u0026oacute;n et al., 2021a, Gallego et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, Sierra et al., \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Its strong moisture transport capacity favors extreme rainfall in regions such as Llor\u0026oacute;, one of the rainiest localities worldwide (Poveda and Mesa, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). The CHLLJ further interacts across multiple scales with both the ITCZ and the CLLJ, jointly modulating the development of mesoscale convective systems during the rainy season (Dur\u0026aacute;n-Quesada et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Loaiza Ceron et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Mej\u0026iacute;a et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2021\u003c/span\u003e, Small et al., \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). Although its contribution to the MSD has received relatively little attention, seasonal variations in the CHLLJ are likely linked to moisture availability during the MSD over Central America (Dur\u0026aacute;n-Quesada et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Dur\u0026aacute;n-Quesada et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, Gallego et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Collectively, these low-level wind systems enhance the spatiotemporal variability of the MSD and represent a key source of its regional diversity.\u003c/p\u003e\u003cp\u003eAlthough multiple theories have been proposed to explain the mechanisms underlying the MSD, most have emphasized its canonical bimodal structure in seasonal precipitation distribution. Yet, previous studies have revealed substantial spatiotemporal variability, suggesting diverse characteristics, drivers, and physical backgrounds of the MSD. A better understanding of this diversity is essential to extend the concept of the MSD beyond its canonical framework and to improve fine-scale predictions that require explicit characterization of MSD features at the grid-cell level. This study aims to investigate the general aspects of MSD diversity, including its characterization, variability, and potential mechanisms. Section \u003cspan refid=\"Sec2\" class=\"InternalRef\"\u003e2\u003c/span\u003e describes the data and methods, with particular attention to the identification of MSD events and their diversity. Section \u003cspan refid=\"Sec6\" class=\"InternalRef\"\u003e3\u003c/span\u003e presents the main results, followed by a discussion in Section \u003cspan refid=\"Sec7\" class=\"InternalRef\"\u003e4\u003c/span\u003e.\u003c/p\u003e"},{"header":"2. Data and Method","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1.Data\u003c/h2\u003e\u003cp\u003eIn this study, we analyze precipitation from the ERA5 reanalysis (Hersbach et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) for the period 1979\u0026ndash;2019 over the domain [0\u0026ndash;30\u0026deg;N, 60\u0026ndash;120\u0026deg;W] (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). ERA5 is chosen because it provides a long-term (41 years), fine-scale dataset that reliably represents precipitation over both land and ocean, allowing for the construction of a robust climatology. Previous evaluations show that ERA5 reproduces the seasonal cycle and interannual variability of precipitation over Central America and Mexico reasonably well (Centella-Artola et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Morales-Vel\u0026aacute;zquez et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and it captures the mean state and timing of the MSD in close agreement with satellite-based and gauge-based products such as TRMM and CHIRPS (Garc\u0026iacute;a-Franco et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Additional diagnostics\u0026mdash;including cloud fraction, SST, horizontal winds at different pressure levels, and specific humidity\u0026mdash;are also obtained from ERA5. These fields have been extensively validated and show improved consistency with observations compared to earlier reanalyses. Taken together, these assessments indicate that ERA5 provides a suitable reference for characterizing the fine-scale features and dynamics of the MSD, which are often not fully captured by observations alone. Although small-scale biases remain, their influence is limited in this study, which focuses on larger-scale spatiotemporal climatology and intraseasonal variability.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2.MSD detection\u003c/h2\u003e\u003cp\u003eIn this study, we use an event-detection method consisting of two steps to identify the MSD and its associated metrics (Zhao et al., \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb). We first calculate the seasonally varying precipitation climatology in each grid and locate two local maxima: one between May 15th and July 15th, and the other between August 15th and October 15th. Next, we perform two independent linear regressions for each grid: one from the first day of the year to the first maximum, and the other from the second maximum to the end of the year. If the fitted linear trend from the first regression is significantly positive and the trend from the second regression is significantly negative, the climatological existence of the MSD in that particular grid is confirmed (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec). Subsequently, we employ the aforementioned criteria to evaluate the precipitation time series of each year in that grid and determine the yearly occurrence of the MSD. For every MSD event, some metrics can be determined, such as the onset date that corresponds to the first local maximum, the end date that corresponds to the second local maximum, the peak date that corresponds to the minimum precipitation between the onset and end dates, and the duration that corresponds to the period between the onset and end dates. The intensity of each MSD event is quantified using the formula:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:{I}_{msd}=\\frac{mean\\left(P1,\\:P2\\right)-{P}_{msd}}{mean\\left(P1,P2\\right)},$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:P1\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:P2\\)\u003c/span\u003e\u003c/span\u003e are the first and second local precipitation maximum corresponding to the onset and end of the MSD, respectively, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{P}_{msd}\\)\u003c/span\u003e\u003c/span\u003e is the average precipitation during the MSD event (from onset to end). During the MSD detection, the ERA5 precipitation is regridded from its original 0.25\u003csup\u003eo\u003c/sup\u003e grid cells to 0.5\u003csup\u003eo\u003c/sup\u003e and 2939 out of 7381 grids (121*61) have been determined to exhibit climatological MSD characteristics, and 63240 MSD events are subsequently detected. Outputs from this method broadly captures the MSD regions identified in previous studies (Anderson et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, Karnauskas et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2013\u003c/span\u003e, Maga\u0026ntilde;a et al., 1999, Maurer et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), including southern Mexico, the Pacific side of Central America, and the Caribbean.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo investigate the diversity of MSD occurrences, we apply a K-means clustering algorithm (Hartigan and Wong, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e1979\u003c/span\u003e) to the MSD-related precipitation time series. For each identified event, the corresponding seasonal precipitation record is extracted at the grid level. To maintain temporal consistency, precipitation on February 29 of non-leap years is linearly interpolated from February 28 and March 1. This procedure yields a matrix of 63,240 rows and 366 columns, where each row corresponds to a precipitation time series associated with an MSD year. The series are then smoothed using a 31-day Gaussian filter to emphasize seasonal variability. K-means clustering is performed on this dataset, with the number of clusters determined iteratively. Specifically, a new cluster is retained only if its centroid shows less than 0.9 correlation with existing centroids and if a two-sample t-test suggests a distinct distribution. Following this procedure, four clusters are identified as the optimal configuration, capturing the main patterns of MSD diversity while preserving inter-cluster dissimilarity.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3.MSD diversity in CMIP6\u003c/h2\u003e\u003cp\u003eTo evaluate the ability of climate models to reproduce the diversity of the midsummer drought (MSD), we analyze 33 general circulation models (GCMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6). Historical precipitation simulations from 1980 to 2014 (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) are first interpolated to a common horizontal resolution of 1.0\u0026deg; \u0026times; 1.0\u0026deg;. MSD events are then identified in each model using the method described in Section \u003cspan refid=\"Sec4\" class=\"InternalRef\"\u003e2.2\u003c/span\u003e, with detection applied to all grid cells within the study domain, irrespective of whether MSD occurs in the local climatology. This choice reflects the fact that simulated precipitation climatologies may place MSD-like features in different spatial locations across models, making it difficult to compare intermodel differences if detection were limited to a fixed set of MSD-prone grid cells.\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\u003eList of CMIP6 model experiments used in this study. For all models here we use their r1i1p1f1 variant. If the variant is not available, r2i1p1f1 variant is considered.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInstitution\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eReference\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAS-RCEC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTaiESM1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Lee and Liang, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2020\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAWI\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAWI-ESM-1-1-LR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Danek et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2020\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBCC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBCC-CSM2-MR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Xin et al., \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e2018\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBCC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBCC-ESM1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Zhang et al., 2018)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCAS\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFGOALS-f3-L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(YU, \u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCAS\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFGOALS-g3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Li, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCCCR-IITM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eIITM-ESM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Choudhury et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCCCma\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCanESM5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Swart et al., \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCMCC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCMCC-CM2-HR4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Scoccimarro et al., \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2020\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCMCC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCMCC-CM2-SR5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Lovato and Peano, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2020\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCMCC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCMCC-ESM2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Lovato et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEC-Earth-Consortium\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEC-Earth3-AerChem\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Consortium, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2020a\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEC-Earth-Consortium\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEC-Earth3-CC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Consortium, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2021\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEC-Earth-Consortium\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEC-Earth3-Veg-LR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Consortium, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2020b\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEC-Earth-Consortium\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEC-Earth3-Veg\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Wyser et al., \u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e2020\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEC-Earth-Consortium\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEC-Earth3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Consortium, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHAMMOZ-Consortium\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMPI-ESM-1-2-HAM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Neubauer et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIPSL\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eIPSL-CM5A2-INCA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Boucher et al., 2020)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIPSL\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eIPSL-CM6A-LR-INCA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Boucher et al., 2021)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIPSL\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eIPSL-CM6A-LR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Boucher et al., 2018)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMIROC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMIROC-ES2L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Hajima et al., 2019)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMIROC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMIROC6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Tatebe and Watanabe, \u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e2018\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMPI-M\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMPI-ESM1-2-HR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Jungclaus et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMPI-M\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMPI-ESM1-2-LR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Wieners et al., \u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMRI\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMRI-ESM2-0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Yukimoto et al., 2019)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNCAR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCESM2-FV2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Danabasoglu, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019b\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNCAR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCESM2-WACCM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Danabasoglu, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2019c\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNCAR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCESM2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Danabasoglu, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2019a\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNCC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNorCPM1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Bethke et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNCC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNorESM2-LM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Seland et al., \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNCC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNorESM2-MM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Bentsen et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNUIST\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNESM3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Cao and Wang, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSNU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSAM0-UNICON\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(Park and Shin, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)\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\u003eThe diversity of MSD events in each simulation is evaluated by classifying every detected event into one of the four reference clusters derived from ERA5 reanalysis data. The classification is based on the minimum Euclidean distance between the simulated event and the ERA5 cluster centroids. To ensure consistency, the ERA5 clusters are independently derived from ERA5 precipitation regridded to the same 1.0\u0026deg; \u0026times; 1.0\u0026deg; resolution used for CMIP6 simulations.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cp\u003eThe composite seasonal precipitation time series for each cluster and all MSD events are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Regions characterized by climatological MSD features display a canonical bimodal precipitation distribution, with a standard MSD event occurring from 23 June to 19 September and lasting 89 days (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). During this period, a relatively shallow precipitation trough is observed (I\u003csub\u003emsd\u003c/sub\u003e = 0.31), corresponding to a 30% reduction in precipitation relative to the seasonal peaks. This reduction is weaker than the 40% decrease reported by Small et al. (\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2007\u003c/span\u003e) for the entire wet season, a discrepancy largely attributable to methodological differences, including the use of daily precipitation data in this study versus monthly data in theirs. Outputs from the cluster analysis reveal distinct variations in the bimodal seasonal precipitation cycle. Cluster 1 is characterized by a wetter seasonal cycle, with a relatively canonical bimodal structure but a weaker MSD intensity (I\u003csub\u003emsd\u003c/sub\u003e = 0.20) than in the climatological mean. Cluster 2 exhibits the longest MSD duration (126 days, from May to September) and a pronounced trough (I\u003csub\u003emsd\u003c/sub\u003e = 0.54), which is further distinguished by its asymmetric peaks, with the first precipitation maximum nearly 1.5 times larger than the second. In contrast, Cluster 3, which displays the strongest MSD intensity (I\u003csub\u003emsd\u003c/sub\u003e = 0.59), is characterized by the second peak clearly exceeding the first one. Cluster 4 resembles a compressed version of the climatological MSD, with moderate intensity (I\u003csub\u003emsd\u003c/sub\u003e = 0.39) and the shortest duration (70 days) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea-b). Overall, the four clusters capture distinct modes of seasonal precipitation variability and MSD manifestation, demonstrating the effectiveness of cluster analysis for exploring the diversity of MSD events.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo examine the spatiotemporal variability of MSD diversity, we mapped the number of calendar years exhibiting MSD events during 1979\u0026ndash;2019 for each cluster (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Clear spatial heterogeneity is evident across Mesoamerica. MSD events in Cluster 1 predominantly occur south of 12\u0026deg;N, covering the Pacific coast of Central America as well as coastal and offshore regions of the eastern Pacific (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). Cluster 2 shows a preference for Cuba, the coastal areas surrounding the Gulf of Mexico, and parts of the North Atlantic (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb). Cluster 3 is most pronounced over the Caribbean Sea, portions of the North Atlantic, the Yucat\u0026aacute;n Peninsula, and the Pacific coasts of Central America and Mexico (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec). Cluster 4 is mainly confined to regions north of 12\u0026deg;N, including southern Mexico and adjacent oceans. Clusters 1 and 4 together delineate the canonical MSD regions encompassing much of Central America and Mexico, while Clusters 2 and 3 highlight the occurrence of MSD events over the Caribbean and neighboring ocean basins. These results indicate that MSD diversity exhibits distinct spatial preferences, which is likely to reflect the differences in the underlying mechanisms which modulate regional precipitation variability.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo investigate the mechanisms underlying MSD diversity, we examined atmospheric and oceanic conditions during MSD evolution, using the diagnostics described in Section \u003cspan refid=\"Sec4\" class=\"InternalRef\"\u003e2.2\u003c/span\u003e including SST, cloud fraction, and winds at 200 hPa and 850 hPa. SST and cloud fraction were selected as proxies for the SST\u0026ndash;cloud feedback mechanism proposed to explain MSD onset and development (Herrera et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2015\u003c/span\u003e, Maga\u0026ntilde;a et al., 1999, Maga\u0026ntilde;a and Caetano, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), with 200 hPa and 850 hPa winds selected as indicators for examing the upper- and lower-level atmospheric circulation, respectively. Anomalous composites were calculated by averaging these diagnostics across all MSD events in each cluster at key phases (onset, peak, and end), with low-level fields (SST, 850 hPa winds) shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e and upper-level fields (cloud fraction, 200 hPa winds) in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. At both the onset and termination of the MSD, the diagnostics exhibit similar patterns: enhanced convection over canonical MSD regions, anomalous low-level westerlies from the Pacific contributing to convergence over the Yucat\u0026aacute;n Peninsula, and cooler SSTs in the adjacent eastern Pacific. This resemblance is consistent with the bimodal precipitation structure of the climatological MSD, which features two comparable rainfall peaks (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). Such symmetric onset\u0026ndash;end patterns are evident in Clusters 1 and 4, both of which closely resemble the canonical MSD (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). In contrast, the diagnostic patterns diverge between the onset and end phases in Clusters 2 and 3, reflecting their non-canonical precipitation structures. In Cluster 2, low-level convergence over the Gulf of Mexico and the Caribbean Sea is evident at the onset but absent at the end (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), consistent with stronger cloud anomalies and more vigorous convection at the onset than at the termination (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). Cluster 3, however, maintains strong low-level convergence and active convection at both the onset and end. Notably, SSTs over the Caribbean Sea and Gulf of Mexico are warmer at the end (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), suggesting an enhanced SST\u0026ndash;cloud radiative feedback that may account for the stronger second precipitation peak characteristic of this cluster.\u003c/p\u003e\u003cp\u003eAt the MSD peak, distinct anomalies emerge, including anomalously warmer SSTs along the Pacific coasts, suppressed convection, and anomalous westerlies extending from the Caribbean Sea toward the Pacific (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). These conditions contrast sharply with those observed at the onset and termination of the MSD. The opposing patterns between onset/end and peak suggest that MSD development may involve retreating processes, such as SST\u0026ndash;cloud radiative feedback (Maga\u0026ntilde;a et al., 1999) and variations associated with the solar declination angle (Karnauskas et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Such contrasts are less evident in the MSD clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). For example, in Cluster 4, persistently cooler SSTs are observed along the Pacific coast of Central America throughout the MSD, while in Cluster 2, SST anomalies at the peak are only weakly positive relative to climatology. These findings point to potential variations in the mechanisms driving bimodal precipitation across different MSD clusters\u0026mdash;variations that cannot be inferred solely from the climatological composite.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe results presented above, together with previous studies, suggest that the SST and low-level winds act as key drivers of MSD. To further assess their roles, we examine the seasonal evolution of several important low-level flows\u0026mdash;the CLLJ, the CHLLJ, and the EP wind flow\u0026mdash;along with EP SST, and evaluate their coherence with MSD diversity. The CLLJ and EP wind are included because they have been suggested in previous literatures as either primary drivers or indicators of MSD events in canonical regions through their influence on moisture transport and regional convergence (Dur\u0026aacute;n-Quesada et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2017\u003c/span\u003e, Garc\u0026iacute;a-Franco et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, Maga\u0026ntilde;a et al., 1999, Zhao et al., \u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Although the CHLLJ has received comparatively less attention, earlier work has shown that its intensity is linked to seasonal moisture transport from the Pacific, implying a potential role in MSD development (Morales et al., 2017). EP SST is considered here as a proxy for SST\u0026ndash;cloud radiative feedback. Indices for the CLLJ, CHLLJ, EP wind, and EP SST are constructed by averaging 925 hPa winds and SST within their respective domains (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). To capture their seasonal variations associated with MSD diversity, the daily time series for each MSD year is averaged following the same procedure used to construct the MSD composites in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e (see in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe CLLJ exhibits two distinct seasonal peaks, in March\u0026ndash;May and September\u0026ndash;November, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea), consistent with previous analyses (Cook and Vizy, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, Wang, \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2007\u003c/span\u003e, Wang and Lee, \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). There is a significant correlation between the CLLJ and MSD precipitation in clusters 2, 3, and 4, which broadly represent bimodal rainfall regimes over southern Mexico and the Caribbean (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ee). This finding reinforces the critical role of the CLLJ in shaping MSD characteristics in these regions, while also underscoring its relatively weak influence on MSD precipitation in cluster 1, corresponding to Central America and the adjacent oceans. In MSD years, the CHLLJ displays a weak but discernible bimodal seasonal cycle, with a pronounced maximum from September to November and a secondary peak in May\u0026ndash;July (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb). This pattern is consistent with the climatological seasonality of the CHLLJ reported in earlier studies (Rueda and Poveda, \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2006\u003c/span\u003e, Sierra et al., \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e2021\u003c/span\u003e, Yepes et al., \u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Notably, the CHLLJ shows significant correlations with MSD precipitation across all clusters of MSD diversity (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ee). The seasonal cycle of the EP wind flow during MSD years exhibits a clear bimodal structure, with two westerly peaks in May\u0026ndash;June and September\u0026ndash;October (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec). This pattern closely resembles the bimodal rainfall structure of the MSD, as reflected in the significant correlations between EP wind anomalies and MSD precipitation across all diversity clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ee). In contrast, the EP SST seasonal cycle shows only a single distinct peak, with a modest warming in July\u0026ndash;August followed by pronounced cooling at the canonical MSD termination in late August to September (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ed). This result is consistent with Garc\u0026iacute;a-Franco et al. (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), who found similar features in pentad-mean SST, including weak bimodality and values below 29\u0026deg;C during the MSD. Correlations between EP SST and MSD precipitation are predominantly negative (clusters 3, 4, and all MSD events) but occasionally positive (clusters 1 and 2) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ee), underscoring that SST\u0026ndash;cloud radiative feedback alone is insufficient to explain MSD dynamics. Overall, the diversity of MSD is not well represented in the seasonal cycles of the four selected indices, indicating that none of them individually resolves full mechanisms underlying MSD variability.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThen, we would like to address the question, \u0026ldquo;How does the low-level wind flows influence the MSD diversity?\u0026rdquo;. Earlier research has contended that the impact of low-level wind patterns, particularly the CLLJ, on the MSD is effectively addressed by the convergence of moisture flux. To quantify the influence of the moisture flux convergence on the MSD, we use the formula of column-averaged moisture flux convergence (MFC) modifed from Garc\u0026iacute;a‑Franco et al. (2023):\u003c/p\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:MFC=-\u0026lt;\\frac{duq}{dx}\u0026gt;-\u0026lt;\\frac{dvq}{dy}\u0026gt;,\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e\u003cp\u003ewhere u and v indicate zonal and meridional wind flows, q indicates specific humidity, and \u0026lt;\u0026thinsp;\u0026gt;\u0026thinsp;indicates column average within troposphere. The MFC are calculated independently for each spatial grid and then averaged across corresponding MSD events to create composite representations of seasonal variation, either for all MSD events or MSD events assigned to a particular cluster. On the climatological scale, MFC shows a strong connection with MSD precipitation, as evidenced by their nearly identical bimodal seasonal cycles and significant correlations (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ea). The bimodal structure of seasonal MFC during MSD years results from the combined influence of its zonal and meridional components: the zonal component contributes to moistening, whereas the meridional component exerts a drying effect. Even when MSD diversity is considered, the seasonal evolution of MFC continues to closely track MSD precipitation. In cluster 1, the annual MFC distribution is dominated by moistening from the zonal component, while the meridional component sustains mid-summer drying, producing a prolonged but weaker precipitation trough (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eb). In clusters 2 and 3, where MSD events are concentrated over the Caribbean Sea, the bimodal MFC pattern is primarily driven by the meridional component, with the zonal component providing the background moistening (Figs.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ec,d). In cluster 4, the seasonal MFC distribution resembles the canonical MSD structure, emerging from the combined effects of both zonal and meridional components. These results indicate that the bimodal structure of the MSD and its diversity are governed by the spatiotemporal variability of both the zonal and meridional components of MFC, rather than by individual zonal low-level jets such as the CLLJ or CHLLJ.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFinally, we evaluate whether the CMIP6 models can reasonably reproduce the MSD and its diversity. We compare the simulated MSD frequencies during 1982\u0026ndash;2014 with ERA5, focusing on both the overall occurrence and the diversity of MSDs, by mapping their spatial patterns and assessing the Pattern Correlation Coefficient (PCC) and Normalized Root Mean Square Error (NRMSE). From the perspective of all MSD events, the CMIP6 models capture the frequency patterns reasonably well, except for a notable underestimation over the Caribbean and adjacent regions. This is reflected by a mean PCC of 0.65 and a mean NRMSE of 0.88 estimated from multimodel ensemble mean (MMM), with 29 out of 33 models showing PCC\u0026thinsp;\u0026ge;\u0026thinsp;0.5 and NRMSE\u0026thinsp;\u0026le;\u0026thinsp;1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ec). In contrast, the models perform less consistently in simulating MSD diversity. For the two clusters characterized by symmetric bimodal distributions (Clusters 1 and 4), the simulated MSD frequencies are relatively well represented. Cluster 1 (4) shows mean PCC values of 0.58 (0.54) and mean NRMSE of 0.91 (0.92), with 20 (18) out of 33 models meeting the PCC\u0026thinsp;\u0026ge;\u0026thinsp;0.5 and NRMSE\u0026thinsp;\u0026le;\u0026thinsp;1 criteria (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ef, o). However, for the two asymmetric clusters (Clusters 2 and 3), model performance is substantially weaker. Cluster 2 (3) yields mean PCC values of 0.38 (0.40) and mean NRMSE of 1.21 (1.09), with only 0 (8) out of 33 models satisfying PCC\u0026thinsp;\u0026ge;\u0026thinsp;0.5 and NRMSE\u0026thinsp;\u0026le;\u0026thinsp;1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ei\u0026ndash;l). Notably, the systematic underestimation of MSD frequency over the Caribbean in the all-model composite (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ea\u0026ndash;b) can be largely attributed to the poor simulation of Clusters 2 and 3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ed, e, g, h).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"4. Discussion and conclusions","content":"\u003cp\u003eThe Mesoamerican MSD has attracted considerable attention due to its profound influence on regional agriculture and socioeconomic activities. Numerous hypotheses have been proposed to explain the bimodal precipitation structure of the MSD, each emphasizing distinct local features. However, investigations on the diversity of MSD manifestations remains less explored, with most of the previous studies analyzed MSD mechanisms and characteristics using composite approaches. With the use of ERA5 reanalysis and CMIP6 model simulations, this study provides a comprehensive examination of the MSD diversity, including its characteristics, spatial distribution, and potential driving mechanisms.\u003c/p\u003e\u003cp\u003eThe four objectively identified clusters of MSD-related seasonal precipitation variations capture the breadth of MSD diversity. Each cluster exhibits distinct seasonal distributions and spatial preferences, encompassing a wide range of variations previously reported across southern Mexico, the Pacific and Caribbean slopes of Central America, the Caribbean basin, and adjacent oceanic regions. The clustering of MSD events based on daily data highlights the importance of fine-scale detection using sub-monthly observations (Garc\u0026iacute;a-Franco et al., 2021; Zhao and Zhang, \u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), as it enables the identification of a sufficient number of events to train the clustering algorithm. Among the clusters, the pronounced asymmetry between the two precipitation peaks in the second and third clusters is particularly noteworthy, suggesting that MSD events in the Caribbean may arise from mechanisms distinct from those over continental Mesoamerica. Compared with the canonical MSD regions (e.g., Garc\u0026iacute;a-Oliva et al., 2021; Garc\u0026iacute;a-Franco et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), the Caribbean has received relatively little attention. Our results indicate that the MSD in this region cannot be neglected, as it exhibits unique features in both precipitation characteristics and underlying dynamics.\u003c/p\u003e\u003cp\u003ePrevious studies have highlighted the role of boreal summer sea surface temperatures (SSTs) and low-level wind patterns in shaping MSD evolution (Maga\u0026ntilde;a et al., 1999, Maga\u0026ntilde;a and Caetano, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). In contrast, our results show a weak connection between SSTs and MSD precipitation seasonality. Neither the climatic mean nor the seasonal evolution of eastern Pacific SSTs exhibits a bimodal signal consistent with MSD development, contrary to the hypothesis by Maga\u0026ntilde;a et al. (1999), proposing a cooler SST (\u0026lt;\u0026thinsp;29\u0026deg;C) during the precipitation minimum period. Instead, anomalous warming or weak cooling is generally observed associated with precipitation reductions, suggesting that SST\u0026ndash;cloud feedback is less influential than low-level wind dynamics, or that SST anomalies may reflect feedbacks from air\u0026ndash;sea interactions driven by moisture convergence. This interpretation is consistent with recent findings using the same ERA5 dataset (Garc\u0026iacute;a-Franco et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), which also point to a limited role of SSTs in MSD formation.\u003c/p\u003e\u003cp\u003eThe strong correspondence between seasonal bimodality in low-level wind indices and MSD precipitation emphasizes the central role of low-level jets. The CLLJ, a key feature of the Intra-Americas Seas circulation, has been widely recognized as a driver of MSD variability through its modulation of air\u0026ndash;sea interaction and moisture convergence. Our results further suggest that other low-level flows, including the CHLLJ and easterly winds over the eastern Pacific, also play important roles. However, jet indices alone are insufficient to fully distinguish MSD diversity. By employing the MFC framework, we decomposed the contributions of zonal and meridional components of moisture transport. The results indicate that MFC reproduces the seasonal bimodal precipitation distribution and its variability across regions. MSD events over Central America, southern Mexico, and the eastern Pacific (clusters 1 and 4) are explained by combined zonal and meridional MFC components, consistent with wind patterns identified in earlier studies. In contrast, MSD events in the Caribbean are primarily shaped by meridional MFC, which induces the drying between precipitation peaks, while the zonal component provides a moist background. These findings highlight the need to pay greater attention to meridional moisture transport when analyzing MSD mechanisms in the Caribbean, where reliance on zonal jet indices such as the CLLJ alone may obscure important processes. Our results thus underscore that the CLLJ is one of several interacting drivers of MSD diversity, rather than its sole determinant.\u003c/p\u003e\u003cp\u003eThe CMIP6 evaluation further reveals that while models reproduce the overall MSD frequency reasonably well, they struggle to capture its diversity, particularly the asymmetric clusters prevalent over the Caribbean. This deficiency suggests that current models inadequately represent the regional moisture transport and associated air\u0026ndash;sea interactions that underpin MSD variability. Improving the simulation of these processes is therefore critical for enhancing confidence in future projections of precipitation and its socioeconomic impacts across the Intra-Americas Seas.\u003c/p\u003e\u003cp\u003eIn conclusion, this study provides one of the first comprehensive analyses of MSD diversity, addressing its characteristics, underlying mechanisms, and representation in climate models, and thereby establishing a framework for future investigations. The findings highlight that assessments of climate model performance should explicitly consider MSD diversity, alongside mean states and long-term trends, since a model\u0026rsquo;s ability to reproduce distinct MSD types reflects its skill in capturing the spatiotemporal variability of precipitation systems in the Intra-Americas Seas. Advancing the understanding of MSD diversity, including its features and drivers, is essential for improving knowledge of regional precipitation variability and predictability, with direct implications for agriculture and socioeconomic planning. Future research should build on these results by integrating observational datasets and model experiments to develop a systematic framework for quantifying, evaluating, and exploring MSD diversity.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003e5. Data availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe ERA5 data used in this study are available at https://cds.climate.copernicus.eu/datasets\u003c/p\u003e\n\u003cp\u003e, and the CMIP6 datasets can be accessed via https://esgf-node.ipsl.upmc.fr/projects/cmip6-ipsl/\u003c/p\u003e\n\u003cp\u003e.\u003cstrong\u003e6. Code availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCodes used to detect MSD events can be found via https://github.com/ZijieZhaoMMHW/MSD.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e7. Funding information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work\u0026nbsp;was supported by the ARC Centre of Excellence for Climate Extremes (CE17010023).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e8. Declaration of competing interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eALFARO EJ (2002) Some characteristics of the annual precipitation cycle in Central America and their relationships with its surrounding tropical oceans. T\u0026oacute;picos Meteorol\u0026oacute;gicos y Oceanogr\u0026aacute;ficos 9:88\u0026ndash;103\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eALFARO EJ (2014) Caracterizaci\u0026oacute;n del veranillo en dos cuencas de la vertiente del Pac\u0026iacute;fico de Costa Rica, Am\u0026eacute;rica Central. Revista de Biolog\u0026iacute;a Trop 62:1\u0026ndash;15\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eALFARO EJ, HIDALGO H (2017) Propuesta Metodol\u0026oacute;gica para la predicci\u0026oacute;n clim\u0026aacute;tica estacional del veranillo en la cuenca del r\u0026iacute;o Tempisque en el Pac\u0026iacute;fico Norte de Costa Rica. T\u0026oacute;picos Meteorol\u0026oacute;gicos y Oceanogr\u0026aacute;ficos 16:62\u0026ndash;74\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAMADOR JA (2008) The intra-Americas sea low‐level jet: Overview and future research. Ann N Y Acad Sci 1146:153\u0026ndash;188\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eANDERSON TG, ANCHUKAITIS KJ, PONS D, TAYLOR M (2019) Multiscale trends and precipitation extremes in the Central American Midsummer Drought. Environ Res Lett 14:124016\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBENTSEN M, TONIAZZO OLIVI\u0026Egrave;DJLSELAND\u0026Oslash;, GJERMUNDSEN T, GRAFF A, GUPTA LSDEBERNARDJB, HE AK, GRIESFELLER YKIRKEV\u0026Aring;GASCHWINGERJTJIPUTRAJAASKSBETHKEIFANY, GRINI J, GUO A, ILICAK C, LIAKKA MKARSETIHHLANDGRENOA, TANG JMOSEIDKONUMMELINASPENSBERGERC, ZHANG H (2019) In: SCHULZ M (ed) NCC NorESM2-MM model output prepared for CMIP6 CMIP historical. Z., HEINZE, C., IVERSEN. Earth System Grid Federation.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBETHKE I, WANG Y, COUNILLON F, KIMMRITZ M, FRANSNER F, SAMUELSEN A, LANGEHAUG HR, BENTSEN CHIUP-G, GUO M, KIRKEV\u0026Aring;G CTJIPUTRAJ, SELAND AOLIVI\u0026Egrave;DJL, ELDEVIK \u0026Oslash;FANYLAWRENCEP, T., KEENLYSIDE N (2019) NCC NorCPM1 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBOUCHER O, LEVAVASSEUR DENVILS, COZIC G, CAUBEL A, MEURDESOIF AFOUJOLSM-A, BALKANSKI Y, CHECA-GARCIA Y, R., HAUGLUSTAINE, D., BEKKI, S., MARCHAND M 2020. IPSL IPSL-CM5A2-INCA model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBOUCHER O, LEVAVASSEUR DENVILS, COZIC G, CAUBEL A, MEURDESOIF AFOUJOLSM-A, BALKANSKI Y, CHECA-GARCIA Y, HAUGLUSTAINE R, D., BEKKI, S., MARCHAND M 2021. IPSL IPSL-CM6A-LR-INCA model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBOUCHER O, LEVAVASSEUR DENVILS, COZIC G, CAUBEL A, FOUJOLS A, MEURDESOIF M-A, CADULE Y, DEVILLIERS P, LEBAS MGHATTASJ, LURTON N, MUSAT TMELLULL, I., MIGNOT, J., CHERUY F 2018. IPSL IPSL-CM6A-LR model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCAO J, WANG B (2019) NUIST NESMv3 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCENTELLA-ARTOLA A, BEZANILLA-MORLOT A, HERRERA TAYLORMA, MARTINEZ-CASTRO DA, SIERRA-LORENZO DGOUIRANDI, VICHOT-LLANO M, A., STEPHENSON, T., FONSECA C (2020) Evaluation of sixteen gridded precipitation datasets over the Caribbean region using gauge observations. \u003cem\u003eAtmosphere\u003c/em\u003e, 11, 1334\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCER\u0026Oacute;N WL, ANDREOLI KAYANOMT, AVILA-DIAZ RV, DE A, SOUZA IP, SOUZA RA 2021a. Pacific and atlantic multidecadal variability relations with the choco and caribbean low-level jets during the 1900\u0026ndash;2015 period. Atmosphere, 12, 1120\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCER\u0026Oacute;N WL, ANDREOLI KAYANOMT, CANCHALA RV, CARVAJAL-ESCOBAR T (2021b) Y. \u0026amp; ALFONSO-MORALES, W. Rainfall variability in southwestern Colombia: changes in ENSO-related features. \u003cem\u003ePure and Applied Geophysics\u003c/em\u003e, 178, 1087\u0026ndash;1103\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCHOUDHURY AD, RAGHAVAN K, NARAYANASETTI GOPINATHANPA, SINGH S (2019) In: MODI A (ed) CCCR-IITM IITM-ESM model output prepared for CMIP6 CMIP historical. M., PANICKAL\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCONSORTIUM EC-E (2019) EC-Earth-Consortium EC-Earth3 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCONSORTIUM EC-E (2020a) EC-Earth-Consortium EC-Earth3-AerChem model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCONSORTIUM EC-E (2020b) EC-Earth-Consortium EC-Earth3-Veg-LR model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCONSORTIUM EC-E (2021) EC-Earth-Consortium EC-Earth-3-CC model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCOOK KH, VIZY EK (2010) Hydrodynamics of the Caribbean low-level jet and its relationship to precipitation. J Clim 23:1477\u0026ndash;1494\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCORRALES-SUASTEGUI A, FUENTES‐FRANCO R, PAVIA EG (2020) The mid‐summer drought over Mexico and Central America in the 21st century. Int J Climatol 40:1703\u0026ndash;1715\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDANABASOGLU G (2019a) NCAR CESM2 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDANABASOGLU G (2019b) NCAR CESM2-FV2 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDANABASOGLU G (2019c) NCAR CESM2-WACCM model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDANEK C, SHI X, STEPANEK C, YANG H, BARBI D, HEGEWALD J, LOHMANN G (2020) IPCC DDC: AWI AWI-ESM1. 1LR model output prepared for CMIP6 CMIP historical\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDE SOUSA K, CASANOVES F, OSPINA SELLAREJ, AGUILAR ASUCHINIJG, A., MERCADO L (2018) How climate awareness influences farmers\u0026rsquo; adaptation decisions in Central America? J Rural Stud 64:11\u0026ndash;19\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDILLEY M (1996) Synoptic controls on precipitation in the Valley of Oaxaca, Mexico. Int J Climatology: J Royal Meteorological Soc 16:1019\u0026ndash;1031\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDUR\u0026Aacute;N-QUESADA AM, GIMENO L, AMADOR J (2017) Role of moisture transport for Central American precipitation. Earth Sys Dyn 8:147\u0026ndash;161\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDUR\u0026Aacute;N-QUESADA AM, ORDO\u0026Ntilde;EZ SOR\u0026Iacute;R, P., GIMENO L (2020) Climate perspectives in the intra\u0026ndash;Americas seas. Atmosphere 11:959\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDUR\u0026Aacute;N-QUESADA AM, GIMENO L, AMADOR J, NIETO R (2010) Moisture sources for Central America: Identification of moisture sources using a Lagrangian analysis technique. J Geophys Research: Atmos, 115\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEAKIN H, LERNER SWEENEYS, APPENDINI AM, PERALES K, DEWES HSTEIGERWALDDG, DAVENPORT CF, F., BAUSCH JC (2018) Agricultural change and resilience: Agricultural policy, climate trends and market integration in the Mexican maize system. Anthropocene 23:43\u0026ndash;52\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGALLEGO D, GARC\u0026Iacute;A-HERRERA R, G\u0026Oacute;MEZ-DELGADO FDP, ORDO\u0026Ntilde;EZ-PEREZ P, RIBERA P (2019) Tracking the moisture transport from the Pacific towards Central and northern South America since the late 19th century. Earth Sys Dyn 10:319\u0026ndash;331\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGAMBLE DW, PARNELL DB, CURTIS S (2008) Spatial variability of the Caribbean mid-summer drought and relation to north Atlantic high circulation. Int J Climatology: J Royal Meteorological Soc 28:343\u0026ndash;350\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGARC\u0026Iacute;A-FRANCO JL, CHADWICK R, GRAY L, OSPREY S, ADAMS DK (2023) Revisiting mechanisms of the Mesoamerican Midsummer drought. Clim Dyn 60:549\u0026ndash;569\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHAJIMA T, ABE M, ARAKAWA O, SUZUKI T, KOMURO Y, OGURA T, OGOCHI K, WATANABE M, YAMAMOTO A, TATEBE H, OHGAITO NOGUCHIMA, ITO R, YAMAZAKI A, ITO D, TAKATA A, WATANABE K, S., KAWAMIYA, M., TACHIIRI K 2019. MIROC MIROC-ES2L model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHARTIGAN JA, WONG MA (1979) Algorithm AS 136: A k-means clustering algorithm. J royal Stat Soc Ser c (applied statistics) 28:100\u0026ndash;108\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHASTENRATH S (1976) Variations in low-latitude circulation and extreme climatic events in the tropical Americas. J Atmos Sci 33:202\u0026ndash;215\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHELLIN J, COX R, L\u0026Oacute;PEZ-RIDAURA S (2017) Maize diversity, market access, and poverty reduction in the Western Highlands of Guatemala. Mt Res Dev 37:188\u0026ndash;197\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHERRERA E, MAGA\u0026Ntilde;A V, CAETANO E (2015) Air-sea interactions and dynamical processes associated with the midsummer drought. Int J Climatol, 35\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHERSBACH H, BELL B, BERRISFORD P, HIRAHARA S, MU\u0026Ntilde;OZ-SABATER HOR\u0026Aacute;NYIA, PEUBEY JNICOLASJ, RADU C, R., SCHEPERS D (2020) The ERA5 global reanalysis. Q J R Meteorol Soc 146:1999\u0026ndash;2049\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHIDALGO HG, DUR\u0026Aacute;N-QUESADA AM, AMADOR, J. A., ALFARO EJ (2015) The caribbean low‐level jet, the inter‐tropical convergence zone and precipitation patterns in the intra‐americas sea: a proposed dynamical mechanism. Geogr Annaler: Ser Phys Geogr 97:41\u0026ndash;59\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJUNGCLAUS J, SCHUPFNER BITTNERMWIENERSK-HWACHSMANNF, LEGUTKE M, GIORGETTA S, REICK M, GAYLER C, HAAK V, DE VRESE H, RADDATZ P, ESCH T, MAURITSEN M, CLAUSSEN TVONSTORCHJ-SBEHRENSJBROVKINV, CRUEGER M, FAST T, FIEDLER I, HAGEMANN S, HOHENEGGER S, JAHNS C, KLOSTER T, KINNE S, LASSLOP S, KORNBLUEH G, MATEI LMAROTZKEJ, MERANER D, K., ROECKNER E (2019), MIKOLAJEWICZ, U., MODALI, K., M\u0026Uuml;LLER, W., NABEL, J., NOTZ, D., PETERS-VON GEHLEN, K., PINCUS MPI-M MPI-ESM1.2-HR model output prepared for CMIP6 CMIP historical. Earth System Grid Federation.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKARNAUSKAS KB, GIANNINI SEAGERR, A., BUSALACCHI AJ (2013) A simple mechanism for the climatological midsummer drought along the Pacific coast of Central America. Atm\u0026oacute;sfera 26:261\u0026ndash;281\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLEE W, LIANG H (2020) AS-RCEC TaiESM1. 0 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLI L (2019) CAS FGOALS-g3 model output prepared for CMIP6 CMIP historical. \u003cem\u003e(No Title)\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLOAIZA CERON W, FERREIRA ANDREOLIRVKAYANOMT, DE SOUZA RA, JONES C, CARVALHO LM (2020) The influence of the atlantic multidecadal oscillation on the choco low-level jet and precipitation in colombia. Atmosphere 11:174\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLOVATO T, PEANO D (2020) CMCC CMCC-CM2-SR5 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLOVATO T, PEANO D, BUTENSCH\u0026Ouml;N M (2021) CMCC CMCC-ESM2 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMAGA\u0026Ntilde;A V, AMADOR, J. A., MEDINA S (1999) The midsummer drought over Mexico and Central America. J Clim 12:1577\u0026ndash;1588\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMAGA\u0026Ntilde;A V, CAETANO E (2005) Temporal evolution of summer convective activity over the Americas warm pools. Geophys Res Lett, 32\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMALDONADO T, RUTGERSSON A, ALFARO E, AMADOR J, CLAREMAR B (2016) Interannual variability of the midsummer drought in Central America and the connection with sea surface temperatures. Adv Geosci 42:35\u0026ndash;50\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMAPES BE, LIU P, BUENNING N (2005) Indian monsoon onset and the Americas midsummer drought: Out-of-equilibrium responses to smooth seasonal forcing. J Clim 18:1109\u0026ndash;1115\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMARTINEZ C, GODDARD L, KUSHNIR Y, TING M (2019) Seasonal climatology and dynamical mechanisms of rainfall in the Caribbean. Clim Dyn 53:825\u0026ndash;846\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMAURER EP, STEWART IT, JOSEPH, K., HIDALGO HG (2022) The Mesoamerican mid-summer drought: the impact of its definition on occurrences and recent changes. Hydrol Earth Syst Sci 26:1425\u0026ndash;1437\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMEJ\u0026Iacute;A JF, POVEDA YEPESJHENAOJJ, ZULUAGA G, RAYMOND MD, D. J., FUCHS-STONE Ž (2021) Towards a mechanistic understanding of precipitation over the far eastern tropical Pacific and western Colombia, one of the rainiest spots on Earth. J Geophys Research: Atmos, 126, e2020JD033415.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMORALES J, ARIAS, P. A., MART\u0026Iacute;NEZ JA (2017) Role of the Caribbean low-level jet and the Choco jet in the patterns of atmospheric moisture transport towards Central America. Sciforum Electronic Conference Series\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMORALES-VEL\u0026Aacute;ZQUEZ MI, RAFIEEINASAB HERRERAGDSAPARICIOJ, A., LOBATO-S\u0026Aacute;NCHEZ R (2021) Evaluating reanalysis and satellite-based precipitation at regional scale: A case study in southern Mexico. Atm\u0026oacute;sfera 34:189\u0026ndash;206\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMOSI\u0026Ntilde;O A, GARC\u0026Iacute;A E (1966) Evaluaci\u0026oacute;n de la sequ\u0026iacute;a intraestival en la Rep\u0026uacute;blica Mexicana. Proc. Conf. Reg. Latinoamericana Uni\u0026oacute;n Geogr. Int, 500\u0026ndash;516\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNEUBAUER D, SIEGENTHALER-LE DRIAN FERRACHATS, STOLL C, FOLINI J, TEGEN DS, STEMMLER IWIENERSK-HMAURITSENT, BARTHEL I, BEY S, DASKALAKIS I, KOKKOLA NHEINOLDB, PARTRIDGE H, RAST D, SCHMIDT S, WATSON-PARRIS H, D., LOHMANN U (2019) HAMMOZ-Consortium MPI-ESM1.2-HAM model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePARK S, SHIN J (2019) SNU SAM0-UNICON model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePERDIG\u0026Oacute;N-MORALES J, ROMERO-CENTENO R, ORDO\u0026Ntilde;EZ P, NIETO R, GIMENO L, BARRETT BS (2021) Influence of the Madden-Julian oscillation on moisture transport by the caribbean low level jet during the midsummer drought in Mexico. Atmos Res 248:105243\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePERDIG\u0026Oacute;N-MORALES J, ROMERO‐CENTENO R, P\u0026Eacute;REZ PO, BARRETT BS (2018) The midsummer drought in Mexico: perspectives on duration and intensity from the CHIRPS precipitation database. Int J Climatol 38:2174\u0026ndash;2186\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePONS D, GRIFFIN TAYLORMJ, CASTELLANOS D, E. J., ANCHUKAITIS KJ (2017) On the production of climate information in the high mountain forests of Guatemala. Annals Am Association Geographers 107:323\u0026ndash;335\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePOVEDA G, MESA OJ (2000) On the existence of Llor\u0026oacute; (the rainiest locality on Earth): Enhanced ocean-land‐atmosphere interaction by a low‐level jet. Geophys Res Lett 27:1675\u0026ndash;1678\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRAUSCHER SA, GIORGI F, DIFFENBAUGH, N. S., SETH A (2008) Extension and intensification of the Meso-American mid-summer drought in the twenty-first century. Clim Dyn 31:551\u0026ndash;571\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRUEDA OA, POVEDA G (2006) Spatial and temporal variability of the Choco jet stream and its effect on the hydroclimatology of the Colombian pacific; variabilidad espacial y temporal del chorro del Choco y su efecto en la hidroclimatologia de la region del pacifico colombiano. \u003cem\u003eMeteorologia Colombiana\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSCOCCIMARRO E, BELLUCCI A, PEANO D (2020) CMCC CMCC-CM2-HR4 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSELAND \u0026Oslash;, BENTSEN M, TONIAZZO OLIVI\u0026Egrave;DJL, GJERMUNDSEN T, GRAFF A, GUPTA LSDEBERNARDJB, HE AK, GRIESFELLER YKIRKEV\u0026Aring;GASCHWINGERJTJIPUTRAJAASKSBETHKEIFANY, GRINI J, GUO A, ILICAK C, LIAKKA MKARSETIHHLANDGRENOA, TANG JMOSEIDKONUMMELINASPENSBERGERC, ZHANG H (2019) In: SCHULZ M (ed) NCC NorESM2-LM model output prepared for CMIP6 CMIP historical. Z., HEINZE, C., IVERSEN. Earth System Grid Federation.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSIERRA JP, ARIAS PA, DUR\u0026Aacute;N-QUESADA AM, TAPIAS, K. A., VIEIRA, S. C., MART\u0026Iacute;NEZ JA (2021) The Choco low-level jet: past, present and future. Clim Dyn 56:2667\u0026ndash;2692\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSMALL RJO, DE SZOEKE SP, XIE S-P (2007) The Central American midsummer drought: regional aspects and large-scale forcing. J Clim 20:4853\u0026ndash;4873\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSWART NC, KHARIN COLEJNS, LAZARE VV, SCINOCCA M, GILLETT JF, ARORA NPANSTEYJ, MAJAESS VCHRISTIANJRJIAOYLEEWG, SEILER FSAENKOOA, SEINEN C, SHAO C, SOLHEIM A, VON L, SALZEN K, YANG, D., WINTER, B., SIGMOND M (2019) CCCma CanESM5 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTATEBE H, WATANABE M (2018) MIROC MIROC6 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWANG C (2007) Variability of the Caribbean low-level jet and its relations to climate. Clim Dyn 29:411\u0026ndash;422\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWANG C, LEE SK (2007) Atlantic warm pool, Caribbean low-level jet, and their potential impact on Atlantic hurricanes. Geophys Res Lett, 34\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWIENERS K-H, ESCH GIORGETTAMJUNGCLAUSJREICKC, BITTNER M, LEGUTKE M, SCHUPFNER S, WACHSMANN M, GAYLER F, HAAK V, DE VRESE H, RADDATZ P, MAURITSEN T, CLAUSSEN TVONSTORCHJ-SBEHRENSJBROVKINV, CRUEGER M, FAST T, FIEDLER I, HAGEMANN S, HOHENEGGER S, JAHNS C, KLOSTER T, KINNE S, LASSLOP S, KORNBLUEH G, MATEI LMAROTZKEJ, MERANER D, K., ROECKNER E (2019), MIKOLAJEWICZ, U., MODALI, K., M\u0026Uuml;LLER, W., NABEL, J., NOTZ, D., PETERS-VON GEHLEN, K., PINCUS MPI-M MPI-ESM1.2-LR model output prepared for CMIP6 CMIP historical. Earth System Grid Federation.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWYSER K, KJELLSTR\u0026Ouml;M E, KOENIGK T, MARTINS H, D\u0026Ouml;SCHER R (2020) Warmer climate projections in EC-Earth3-Veg: the role of changes in the greenhouse gas concentrations from CMIP5 to CMIP6. Environ Res Lett 15:054020\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXIN X, ZHANG ZHANGJ, WU F, SHI T, LI X, LIU JCHUM, Q., YAN, J., MA Q (2018) BCC BCC-CSM2MR model output prepared for CMIP6 CMIP\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYEPES J, POVEDA G, MORENO MEJ\u0026Iacute;AJF, L., RUEDA C (2019) Choco-jex: A research experiment focused on the Choc\u0026oacute; low-level jet over the far eastern Pacific and western Colombia. Bull Am Meteorol Soc 100:779\u0026ndash;796\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYU Y (2019) CAS FGOALS-f3-L model output prepared for CMIP6 CMIP historical\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYUKIMOTO S, KOSHIRO T, KAWAI H, URAKAWA OSHIMANYOSHIDAK, TSUJINO S, DEUSHI H, TANAKA M, HOSAKA T, YOSHIMURA M, SHINDO H, ISHII EMIZUTAR M., OBATA, A. \u0026amp; ADACHI Y 2019. MRI MRI-ESM2.0 model output prepared for CMIP6 CMIP historical. Earth System Grid Federation\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZHANG J, WU T, SHI X, ZHANG F, LI J, CHU M, MA LIUQYANJ, Q., WEI M 2018. BCC BCC-ESM1 model output prepared for CMIP6 CMIP. \u003cem\u003e(No Title)\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZHAO Z, HAN M, YANG K, HOLBROOK NJ (2023) Signatures of midsummer droughts over Central America and Mexico. Clim Dyn 60:3523\u0026ndash;3542\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZHAO Z, HOLBROOK NJ, BALLESTERO OLIVEREC, D., VARGAS-HERNANDEZ JM (2020) Characteristic atmospheric states during mid-summer droughts over Central America and Mexico. Clim Dyn 55:681\u0026ndash;701\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZHAO Z, ZHANG X (2021) Evaluation of methods to detect and quantify the bimodal precipitation over Central America and Mexico. Int J Climatol 41:E897\u0026ndash;E911\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"climate-dynamics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"cldy","sideBox":"Learn more about [Climate Dynamics](https://www.springer.com/journal/382)","snPcode":"382","submissionUrl":"https://submission.nature.com/new-submission/382/3","title":"Climate Dynamics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Midsummer Drought, Precipitation, Drought, CMIP6, Cluster Analysis","lastPublishedDoi":"10.21203/rs.3.rs-7652378/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7652378/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe Mesoamerican midsummer drought (MSD) is a distinctive precipitation feature characterized by a mid-season rainfall reduction within the boreal summer wet season. Despite its socioeconomic significance, most previous studies have emphasized its canonical bimodal structure, leaving the diversity of MSD expressions less explored. Using ERA5 reanalysis (1979\u0026ndash;2019), we objectively identify MSD events for each grid cells over the domain and classify them into four distinct clusters via K-means analysis. These clusters reveal diverse temporal structures, intensities, and spatial preferences, spanning southern Mexico, Central America, the Caribbean basin, and adjacent oceans. Composite analyses of sea surface temperature (SST), cloud fraction, winds, and moisture flux convergence (MFC) indicate that low-level circulations\u0026mdash;particularly the Caribbean and Choc\u0026oacute; low-level jets, along with eastern Pacific winds\u0026mdash;play a central role in shaping MSD diversity. In contrast, eastern Pacific SST anomalies exhibit only weak and inconsistent associations, suggesting a secondary role of SST\u0026ndash;cloud feedbacks. Decomposition of MFC further highlights the combined zonal and meridional moisture transport as the primary driver of bimodality, with meridional fluxes being especially important for Caribbean MSD events. An evaluation of 33 CMIP6 models shows that while most capture the overall MSD frequency, they underperform in reproducing asymmetric precipitation structures, particularly those over the Caribbean. These results emphasize the need to incorporate MSD diversity into model evaluation frameworks to improve regional precipitation projections. Our findings provide a new perspective on the mechanisms underlying MSD variability and establish a foundation for more reliable seasonal prediction and climate change assessments in the Intra-Americas Seas region.\u003c/p\u003e","manuscriptTitle":"Diversity of Mesoamerican Midsummer Drought","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-01 08:54:46","doi":"10.21203/rs.3.rs-7652378/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major Revision","date":"2025-12-16T03:40:31+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"","date":"2025-09-23T18:46:38+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-20T13:50:17+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-20T13:22:08+00:00","index":"","fulltext":""},{"type":"submitted","content":"Climate Dynamics","date":"2025-09-18T14:58:50+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"climate-dynamics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"cldy","sideBox":"Learn more about [Climate Dynamics](https://www.springer.com/journal/382)","snPcode":"382","submissionUrl":"https://submission.nature.com/new-submission/382/3","title":"Climate Dynamics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"e1f98255-3974-43cf-a661-efeb45aaabc7","owner":[],"postedDate":"October 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-02-14T10:15:11+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-01 08:54:46","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7652378","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7652378","identity":"rs-7652378","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}