Lightning Behavior and Its Relationship with Topography, Precipitation, and Land Use in the São Francisco River Basin

Lightning Behavior and Its Relationship with Topography, Precipitation, and Land Use in the São Francisco River Basin | 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 Lightning Behavior and Its Relationship with Topography, Precipitation, and Land Use in the São Francisco River Basin Samuel Amorim Silva, Weber Andrade Gonçalves, Lizando Pereira Abreu, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7566962/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 11 You are reading this latest preprint version Abstract With one of the highest rates of lightning activity globally, Brazil faces a significant natural hazard. The São Francisco River Basin (BHSF) represents a key area for studying this phenomenon. This study analyzes the spatiotemporal distribution of lightning activity within the basin, identifies areas of maximum concentration (hotspots), and investigates its relationship with surface and atmospheric drivers. We used lightning data from the Lightning Imaging Sensor (LIS; 1998–2013), precipitation (CHIRPS), topography (ASTER), and land use (MapBiomas) data, applying the K-means clustering technique for pattern segmentation. Results indicate that hotspots, with flash rates up to 39.9 flashes km⁻² yr⁻¹, are concentrated in the western portion of the basin, predominantly over plateau areas characterized by agricultural use. The temporal analysis revealed a distinct seasonal cycle, with maximum activity in summer coupled with the rainfall regime, and a diurnal peak in the late afternoon. We conclude that the interaction between topography-induced air uplift and surface alterations from agricultural land use are the primary modulators of the storm regime in the BHSF, offering valuable insights for risk mitigation strategies. Cluster Hotspots Slope Remote Sensing Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. INTRODUCTION Lightning is the result of electrification processes within deep convective clouds (Williams, 1989 ; Mattos et al., 2024 ). In these systems, strong updrafts promote collisions between ice particles and water droplets, leading to charge separation (Reynolds et al., 1957 ; Saunders, 2008 ). The frequency and intensity of these discharges are not random; they are directly modulated by atmospheric conditions and, crucially, by terrestrial surface factors (Fernandes, 2006 ). When these discharges reach the ground, they become a significant natural hazard, capable of causing serious incidents and fatalities (Pinto Junior & Pinto, 2008 ). Among the most influential surface drivers are topography and land use. Topography often acts as a triggering mechanism for convection by forcing moist air to rise, thereby intensifying storm formation (Bourscheidt et al., 2009 ; Schneider et al., 2019 ). Studies across various Brazilian biomes, from the Amazon to the southern regions, have shown that areas with steeper slopes, such as mountain ridges and hillsides, tend to exhibit higher lightning densities (Bourscheidt et al., 2009 ; Santos et al., 2024 ). Analogously, changes in land cover alter the surface energy and moisture balance, which can also influence lightning patterns (Potdar et al., 2025 ; Santos et al., 2024 ). This dynamic intrinsically links lightning to precipitation events, a relationship documented throughout Brazil (Abreu et al., 2020 ; Mattos & Machado, 2011; Zoboli & Silva, 2023 ). With an average of 96.4 million flashes annually, Brazil is one of the countries most affected by this phenomenon, exhibiting a fatality rate approximately four times higher than that of developed nations (Oda et al., 2022 ; Cardoso et al., 2014 ). This national context contrasts with extreme global hotspots like Lake Maracaibo in Venezuela, which records the world's highest flash density at 232.52 flashes km⁻² yr⁻¹ (Albrecht et al., 2016 ). In Brazil, lightning causes an average of 132 fatalities per year and generates economic losses that can reach R $ 1 billion annually, primarily affecting the electricity and telecommunications sectors (Cardoso et al., 2014 ; G1, 2013). Within this context, the São Francisco River Basin (BHSF) in Brazil emerges as an area of particular interest and concern due to its significant social and economic relevance. The São Francisco River is a critical water source for the country, essential for irrigated agriculture, livestock farming, and hydroelectric power generation (Mutti et al., 2022 ). Spanning multiple regions, the BHSF is situated in territories that collectively account for about 76% of lightning-related fatalities in Brazil, highlighting a pronounced vulnerability, especially for rural activities, which represent 19% of these deaths (Cardoso et al., 2014 ). However, investigating this phenomenon in the BHSF faces a methodological challenge: the ground-based lightning monitoring network has limited spatial coverage in the region (Abreu, 2023 ). This limitation makes the use of remote sensing data, such as from the Lightning Imaging Sensor (LIS), a critical alternative, as it offers comprehensive and consistent spatiotemporal coverage over the basin, including in remote and hard-to-access areas (Abreu, 2023 ). Considering the high frequency of lightning and fatalities within the regions of the BHSF (Cardoso et al., 2014 ; Oda et al., 2022 ), the complex interaction between its surface drivers, and the existing gap in event monitoring, this study aims to characterize and analyze the spatiotemporal patterns of lightning in the basin. We will identify the areas of highest density (hotspots) and correlate them with topographic, precipitation, and land use variables, seeking to fill a knowledge gap and provide insights for risk mitigation strategies in the region. 2. MATERIALS AND METHODS 2.1. Study Area The São Francisco River Basin (BHSF) is one of the most extensive and strategic drainage areas in Brazil, covering approximately 8% of the national territory (CBHSF, 2016). The river flows for 2,863 km through the Northeast, Southeast, and Center-West regions of the country. The basin spans 505 municipalities across six states: Minas Gerais (MG), Goiás (GO), Bahia (BA), Pernambuco (PE), Alagoas (AL), and Sergipe (SE), in addition to the Federal District (DF) (CBHSF, 2016; Fig. 1 ). For planning and management purposes, the BHSF is divided into four sub-regions: Upper (ASF), Middle (MSF), Lower-Middle (SMSF), and Lower São Francisco (BSF). This division is based on the distinct climatic and ecological characteristics along the river's course (Bezerra et al., 2019 ). 2.2. Data This study utilized four primary datasets: lightning data from the Lightning Imaging Sensor (LIS), topographic data from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), precipitation data from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS), and land use and land cover data from the MapBiomas project. 2.2.1. Lightning Data Lightning data were obtained from the Lightning Imaging Sensor (LIS), which operated aboard the Tropical Rainfall Measurement Mission (TRMM) satellite between 1997 and 2015. The satellite maintained an orbit inclined at 35° and an altitude of 350 km, with the mission to detect the optical signatures of lightning (Albrecht et al., 2016 ; Cecil et al., 2014 ). The sensor used a 128x128 pixel array and a bandpass filter centered at 777.4 nm to capture near-infrared oxygen emissions, a strong and consistent spectral signature of lightning that allows it to be distinguished from other light sources (Albrecht et al., 2016 ). The high-resolution LIS climatology, with a 0.1° spatial resolution, is made available by the Global Hydrometeorology Resource Center (GHRC). The sensor's flash detection efficiency has been estimated at 93% at night and 73% during the day (Qie et al., 2003 ). Although the orbital nature of LIS can result in temporal subsampling—meaning it does not observe a specific location continuously—its extensive data collection period establishes it as one of the most important lightning databases for studies involving long time series (Albrecht et al., 2016 ). This is particularly true when compared to more recent products like the Geostationary Lightning Mapper (GLM), which has only provided data for the last seven years (2018–2025) but offers far superior temporal coverage. For this study, the analysis period was set from 1998 to 2013. The exclusion of data from 2014 and 2015 is due to the start of the TRMM satellite's decommissioning process in 2014, which introduced uncertainties and interruptions in the observations. This is a frequent practice adopted in other studies using the same database (Abreu et al., 2020 ; Boccippio et al., 2002 ; Dewan et al., 2018 ; Zhang et al., 2019 ). 2.2.2. Land Use and Land Cover Data MapBiomas is an initiative formed by a collaborative network of Brazilian ONGs, universities, and technology companies that uses time series of satellite imagery to map and monitor land cover and use across the entire Brazilian territory (MapBiomas, 2024 ). Through machine learning algorithms and cloud processing, the platform categorizes the land into classes such as agriculture, pasture, natural vegetation, and urban areas, with high spatial and temporal resolution. This methodology allows for a detailed analysis of changes in land use over the study period (1998 to 2013), providing a robust foundation for studies relating terrestrial dynamics to atmospheric phenomena like lightning. The accuracy and comprehensiveness of MapBiomas make it an indispensable tool for land cover information in regional studies. 2.2.3. Topography and Elevation Data Topographic information was derived from data captured by the ASTER sensor. With its ability to capture stereoscopic images in the near-infrared band at a 15-meter spatial resolution, ASTER data allows for the generation of a Digital Elevation Model (DEM) for the study area. DEMs are fundamental inputs for analyzing orographic effects on atmospheric convection. The ASTER DEM has a vertical accuracy of 20 meters with 95% confidence, without the need for ground control points (Fujisada et al., 2005 ). 2.2.4. Precipitation Data For the precipitation analysis, the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) dataset was used. This product, developed by the U.S. Geological Survey (USGS) and the University of California, Santa Barbara (UCSB), combines satellite estimates with rain gauge station data to create a detailed and reliable historical time series extending from 1981 to the present (Funk et al., 2014 ). This hybrid approach is particularly powerful as it merges the broad spatial coverage of satellites with the ground-truthed accuracy of in-situ measurements. For this study, CHIRPS data with a spatial resolution of 0.05° (~ 5 km) and at a monthly aggregation scale were used, making them suitable for the proposed climatological analysis. 2.3. METHODOLOGY The methodology was structured into four main steps: (1) processing and standardization of the datasets; (2) cluster analysis to identify spatial lightning patterns within the basin; (3) identification and characterization of areas with maximum activity (hotspots); and (4) analysis of the relationship between hotspots and surface drivers (topography and land use). This multi-step approach ensures comprehensive analysis, progressing from initial data preparation to the final physical interpretation of the results. 2.3.1. Data Processing and Standardization Due to the differing spatial resolutions of the source data, it was necessary to standardize all information onto a common grid. This is a critical step in geospatial analysis, as it allows for direct, pixel-by-pixel comparison between different environmental variables. A sample grid was created based on the native 0.1° resolution of the LIS sensor. This grid was then used to extract corresponding values for land use and precipitation. For each LIS grid cell, the predominant land use class from the high-resolution MapBiomas data (30 m) and the corresponding precipitation value from the CHIRPS dataset were assigned, ensuring spatial compatibility across all datasets. 2.3.2. K-means Clustering of Total Lightning To understand the spatial structure of the lightning data, the k-means clustering technique was employed. This algorithm, first proposed by MacQueen ( 1967 ), is an iterative clustering method that partitions a dataset into k predefined clusters, such that the distance from each point to its cluster's centroid is minimized (Peña et al., 1999 ). The goal of k-means is to find the optimal partition that minimizes the objective function, defined as the sum of squared distances between the points and their respective centroids (Wu, 2012 ). The algorithm groups data points with similar values around a central point, with each resulting cluster representing a distinct pattern of lightning activity in the data. The choice of k-means for this analysis was driven by its computational efficiency and simplicity, which are particularly advantageous for large-scale datasets (Arthur & Vassilvitskii, 2007 ). However, it is acknowledged that the method has limitations: it requires the number of clusters (k) to be defined beforehand and is sensitive to the initial placement of centroids, which can lead to suboptimal results (Celebi et al., 2013 ). In this study, the total lightning flash rate density was used as the input variable for the algorithm. Groupings ranging from two to ten clusters were evaluated. The final configuration of four clusters was chosen with the aid of the elbow method and the silhouette index, which are statistical techniques used to estimate the optimal number of clusters in a dataset (Kodinariya & Makwana, 2013). For the dissimilarity measure, both Euclidean and squared Euclidean distances were compared, with the former being selected for its simplicity and effectiveness with the data in question (Peña et al., 1999 ). This approach differs from that of Abreu et al. ( 2020 ); while the latter used seasonal data for clustering, this study used the total lightning density. This approach allows for an analysis focused on the intrinsic, year-round spatial patterns of lightning within the basin. 2.3.3. Hotspot Identification and Characterization Areas with the highest lightning density (hotspots) were identified by sorting the total LIS lightning climatology data in descending order, which allowed for the precise localization of coordinates with the highest flash rates in the basin (Albrecht et al., 2016 ). For a detailed individual analysis of each hotspot, multiple temporal scales of LIS data were used: the single band of total lightning density, the 12 bands corresponding to the monthly average, and the 24 bands for the hourly average, all for the 1998–2013 period. This multi-temporal analysis enables a deeper understanding of each hotspot's behavior, revealing not only where it is most active, but also when throughout the day and year. Numerical values were extracted from sample points matching the original sensor's pixel size, enabling the calculation of descriptive statistics such as mean, standard deviation, and maximum/minimum values. 2.3.4. Analysis of Surface Factors and 3D Modeling Following a methodology based on the work of Bourscheidt et al. ( 2009 ) and Abreu et al. ( 2020 ), ASTER data were used to generate a Digital Elevation Model (DEM). From this DEM, slope maps were calculated in QGIS software, and three-dimensional (3D) models were constructed using the open-source software Blender. This 3D modeling approach, differing from that of Abreu et al. ( 2020 ), allowed for a detailed graphical analysis of the influence of topographic features on the spatial distribution of lightning. This visual approach offers a more intuitive understanding of the complex spatial relationships between terrain and lightning hotspots than traditional 2D maps can provide. 3. RESULTS AND DISCUSSIONS 3.1. Clustering and Spatial Distribution of Lightning The application of the K-means clustering algorithm proved to be an effective tool for segmenting the electrical activity in the basin, resulting in four clusters with distinct lightning regimes (Fig. 2 a). Clusters 1, 2, and 3 exhibit a relatively homogeneous spatial distribution that aligns with the BHSF sub-regions: cluster 1 is concentrated in the Lower and Lower-Middle São Francisco sections, cluster 2 predominantly covers the Middle São Francisco, and cluster 3 is located in the Upper São Francisco. This correspondence suggests that macroclimatic and geographic factors, which define the sub-basins themselves, modulate large-scale electrical activity (Marengo et al., 2012 ). In contrast, cluster 4, which aggregates the pixels with the highest density, exhibits a localized distribution, primarily superimposed over areas belonging to clusters 2 and 3. This configuration indicates the strong influence of local-scale forcings, such as variations in topography and land use, in the genesis of severe storms, highlighting a hierarchy of atmospheric controls, from large-scale climate to local terrain effects. This observation is consistent with studies that identify lightning hotspots anchored by specific geographic features (Albrecht et al., 2016 ; Diaz et al., 2022 ). The analysis of Fig. 2 b demonstrates that lightning hotspots are concentrated in the most elevated areas of the basin, notably in its western portion, where the flash density exceeds 25 flashes km⁻² yr⁻¹. This region is characterized by the predominance of agriculture (Fig. 2 d) and by a plateau topography. The association between electrical activity and elevated areas is a well-documented phenomenon, as orography intensifies atmospheric convection by providing the necessary mechanical lift for the formation of thunderstorms (Kotroni & Lagouvardos, 2008 ; Mondal et al., 2022 ). The eastern portion of the BHSF, despite also having elevated topography, records a significantly lower lightning density (less than 15 flashes km⁻² yr⁻¹). This pattern suggests that the eastern edge of the basin acts as an orographic barrier, placing the study area on the leeward side of prevailing moisture systems. This "rain shadow" effect results in warmer and drier atmospheric conditions that are less conducive to electrification (Roe, 2005 ). The terrain slope also proves to be a crucial modulating factor (Fig. 2 e). Precipitation-inducing systems, upon encountering the rugged topography in the central portion of the basin, undergo orographic uplift, which enhances instability and favors storm formation (Roe, 2005 ; Whiteman, 2000 ). Notably, the highest lightning density values predominantly occur after the systems pass over this zone of steeper slopes. A detailed analysis of the hotspot locations (Fig. 2 f) corroborates this observation, showing that the majority are not located directly on the steepest slopes but rather on adjacent plateau areas. This pattern suggests that the convection, once initiated by the orography, subsequently propagates and reaches maturity over the flatter, elevated areas. Hotspot 5, however, is an exception, as it is located directly over a steep region, indicating that at this point, the intensification of convection may be more immediately anchored by the steep slope itself (Markowski & Richardson, 2010 ). 3.2. Spatiotemporal Characteristics of Lightning Statistical analysis of the clusters quantifies the differences between the lightning regimes (Fig. 3 a). Cluster 1, located in the north, exhibits the lowest mean density (2.86 flashes km⁻² yr⁻¹). Clusters 2 (5.86 flashes km⁻² yr⁻¹) and 3 (8.37 flashes km⁻² yr⁻¹) represent intermediate and transitional conditions. Cluster 4 stands out with the highest mean density (16.68 flashes km⁻² yr⁻¹) and the largest standard deviation (σ = 4.10), concentrating the extreme pixels (outliers). This high variability is characteristic of regions where convection is not only intense but also intermittent, driven by localized, powerful storm cells (Albrecht et al., 2016 ). The hourly distribution (Fig. 3 b) reveals a well-defined diurnal cycle, with a peak in activity concentrated in the late afternoon, between 14:00 and 20:00 UTC. This pattern is typical of continental tropical and subtropical regions, where maximum surface heating from the sun intensifies atmospheric instability and convection, culminating in thunderstorm formation (Burgesser et al., 2014 ; Mondal et al., 2022 ). This is the classic signature of "air-mass thunderstorms," which build up during the day and are fueled by solar radiation. During the early morning hours, residual activity is observed in clusters 2 and 3, possibly associated with the dissipation of Mesoscale Convective Systems (MCSs) that typically form during the preceding evening (Nesbitt et al., 2000 ). The seasonality of electrical activity (Fig. 3 d) directly follows the region's rainfall cycle (Fig. 3 c), with a maximum in summer (December-January-February, DJF) and a minimum in winter (June-July-August, JJA). This strong correlation highlights that the availability of moisture and the dynamics of seasonal weather systems, such as the South Atlantic Convergence Zone (SACZ), are the primary drivers of storm activity in the BHSF (Jayaratne & Kuleshov, 2006 ; Vogt & Hodanish, 2016 ). This strong seasonal agreement is clearly visualized in Fig. 4. During the summer (DJF), there is significant spatial variation, with minimum lightning records in regions closer to the coast (eastern BHSF) and maximums in the interior, notably in the state of Bahia (Figs. 4a and 4b). In stark contrast, winter (JJA) is the season of lowest electrical activity, with a near-total absence of lightning across much of the BHSF, except for its southern portion (Figs. 4e and 4f). The spring (SON) and autumn (MAM) months clearly represent transitional seasons (Figs. 4c, 4d, 4g, 4h). During the winter dry period, the near-zero lightning and precipitation values in the central basin are attributable to the large-scale reduction in convective activity and atmospheric moisture (Nesbitt et al., 2000 ; Zipser et al., 2006 ), reinforcing the influence of climatic and topographic factors on the seasonal distribution of lightning, as noted by Bourscheidt et al. ( 2009 ). Rasmussen and Houze ( 2011 ) identified and mapped deep and extensive convective cores in South America, with a particular focus on the summer period. A comparison between their results and those of the present study reveals overlaps between some of these convective cores and the lightning hotspots identified in the BHSF. This spatial agreement suggests a strong link between intense convective activity and the occurrence of electrical discharges during the summer in the basin. This correlation is further corroborated by studies from Zipser et al. ( 2006 ) and Romatschke and Houze ( 2010 ), which emphasize the role of deep convection in lightning generation, especially in tropical and subtropical regions. This connection reinforces the physical basis of our statistically-identified hotspots, grounding them in the known meteorology of the region. 3.3 Analysis of Lightning Hotspots Ten hotspots of maximum lightning concentration were identified in the BHSF, all belonging to cluster 4 (Table 1 ). The most intense hotspot, located in São Desidério, Bahia, recorded a rate of 39.9 flashes km⁻² yr⁻¹. Although this value is lower than those of major global hotspots, such as Lake Maracaibo in Venezuela (233 flashes km⁻² yr⁻¹) and the Congo Basin (205 flashes km⁻² yr⁻¹) (Albrecht et al., 2016 ), it is comparable to other highly active regions in South America, such as areas in Colombia that record average rates of up to 70 flashes km⁻² yr⁻¹ (Diaz et al., 2022 ). This places the BHSF's most active areas as significant regional centers of intense convective activity. Table 1 Lightning Hotspots in the BHSF. Ranking City State Point Coordinate \(\:Land\:use\) Cluster Topography (m) \(\:flash\:{km}^{-2}{ano}^{-1}\) 1 São Desidério BA 12°31'48.0"S 46°07'48.0"W Agriculture 4 515,0 39,9 2 Luís Eduardo Magalhães BA 11°55'48.0"S 45°55'48.0"W Agriculture 4 542,5 38,8 3 Correntina BA 13°19'48.0"S 45°37'48.0"W Agriculture 4 419,3 35,4 4 Luís Eduardo Magalhães BA 12°25'48.0"S 46°07'48.0"W Agriculture 4 517,2 34,9 5 Formosa do Rio Preto BA 11°19'48.0"S 46°31'48.0"W Pasture 4 677,8 32,0 6 Luís Eduardo Magalhães BA 12°01'48.0"S 45°55'48.0"W Forest Formation 4 507,1 30,5 7 Formoso MG 14°49'48.0"S 46°25'48.0"W Forest Formation 4 1182,4 29,7 8 Santa Fé de Minas MG 16°37'12.0"S 45°13'48.0"W Forest Formation 4 479,9 29,6 9 Buritis MG 15°07'48.0"S 46°49'48.0"W Grassland 4 454,8 29,2 10 Três Marias MG 18°01'12.0"S 45°01'48.0"W Forest Formation 4 603,0 29,2 The São Desidério hotspot (Fig. 5) reveals that maximum activity does not occur on the topographic peaks themselves, but on plateau areas near depressions. This pattern suggests that terrain-induced moisture convergence, combined with mesoscale circulations generated by the differential heating between the plateau and the valleys, may be a key mechanism for the initiation and anchoring of convection (Bourscheidt et al., 2009 ; Whiteman, 2000 ). Hotspots 7 and 10, in turn, are situated in plains surrounded by steep terrain, indicating that the channeling of moisture flows and the convergence of valley-mountain breezes may be the dominant triggers in these locations (Albrecht et al., 2016 ). Land use also stands out as a modulating factor. The four most active hotspots are located in areas of intensive agriculture (Table 1 ). The conversion of native vegetation to agriculture alters the surface energy and moisture balance. Although the relationship is complex, studies indicate that landscape heterogeneity, such as the boundaries between agricultural fields and native vegetation, can create temperature and moisture gradients that induce local circulations and trigger convection (Pielke, 2001 ; Taylor et al., 2011 ). Furthermore, agricultural activities, such as biomass burning, can inject aerosols into the atmosphere, which, under certain moisture conditions, can invigorate cloud updrafts and enhance lightning production (Rosenfeld et al., 2008 ; Wang et al., 2018 ). Therefore, the high incidence of lightning in western Bahia appears to be the result of a synergistic interaction between the favorable topography of the Western Bahia Plateau, seasonal moisture availability, and the surface alterations imposed by agricultural land use. 4. CONCLUSION This study concludes that lightning activity in the BHSF is spatially heterogeneous and can be segmented into four distinct regimes. The large-scale distribution is modulated by macroclimatic factors, while intense convective activity is concentrated in 10 well-defined hotspots controlled by local-scale forcings. The temporal behavior of lightning follows a clear pattern, with a diurnal peak in the late afternoon driven by surface heating and a seasonal cycle with a maximum in the summer, strongly coupled with the region's rainy season. The lightning hotspots are primarily associated with specific topographic features. The majority are located on the Western Bahia Plateau, where convection is intensified by the interaction of the plateau topography with surrounding valleys and depressions. Orographic uplift and terrain-induced mesoscale circulations are identified as the primary triggering mechanisms. Furthermore, agricultural land use emerges as a key modulating factor, with the most intense hotspots situated directly over cultivated areas. It is concluded that while topography acts as the primary trigger for storm initiation, the surface conditions modified by agriculture likely contribute to the intensification of these storms. The primary contribution of this work is the demonstration of the synergistic interaction among topography, land use, and atmospheric dynamics in determining the storm regime of the BHSF. These findings are relevant for improving severe weather warning systems, for electrical and agricultural infrastructure planning, and for a better understanding of hydrometeorological processes in one of Brazil's most important river basins. Looking ahead, future research could employ high-resolution numerical modeling to quantify the relative contribution of each of these factors to the formation of lightning hotspots, an essential step for developing more precise predictive models in a context of ongoing climate and land-use change. Declarations Funding This work was partially supported by the National Council for Scientific and Technological Development (CNPq) [Grant number 312707/2021-5]. Competing Interests The authors have no relevant financial or non-financial interests to disclose. Author Contributions All authors contributed to the study conception and design. The methodology, software development, data curation and formal analysis were performed by Samuel Amorim Silva, with contributions from Weber Andrade Gonçalves, Lizando Pereira de Abreu and Douglas Leonardo Sales Pedrosa. The first draft of the manuscript was written by Samuel Amorim Silva and all authors commented on previous versions of the manuscript. Supervision was provided by Weber Andrade Gonçalves. All authors read and approved the final manuscript. Data Availability Lightning data are freely available from NASA EARTHDATA (DOI: 10.5067/LIS/LIS/DATA304). CHIRPS precipitation data are freely available from the USGS Climate Hazards Center (CHC) at UC Santa Barbara (https://data.chc.ucsb.edu/products/CHIRPS-2.0/). Key processed datasets generated during this study are available on Zenodo (DOI: 10.5281/zenodo.15588150). MapBiomas data can be downloaded from the MapBiomas Brazil project website (https://brasil.mapbiomas.org/downloads/). References Abreu, L. P. (2023). Caracterização dos relâmpagos ocorridos na região Nordeste do Brasil, por meio de sensoriamento remoto [Characterization of lightning in the Northeast region of Brazil using remote sensing] [Doctoral dissertation, Universidade Federal do Rio Grande do Norte]. Abreu, L. P., Gonçalves, W. A., Mattos, E. V., & Albrecht, R. I. (2020). Assessment of the total lightning flash rate density (FRD) in northeast Brazil (NEB) based on TRMM orbital data from 1998 to 2013. International Journal of Applied Earth Observation and Geoinformation , 93 , 102195. https://doi.org/10.1016/j.jag.2020.102195 Albrecht, R. I., Goodman, S. J., Buechler, D. E., Blakeslee, R. J., & Christian, H. J. (2016). Where are the lightning hotspots on Earth? Bulletin of the American Meteorological Society , 97 (11), 2051–2068. https://doi.org/10.1175/BAMS-D-14-00193.1 Albrecht, R. I., Morales, C. A., & Silva Dias, M. A. F. (2011). Electrification of precipitating systems over the Amazon: Physical processes of thunderstorm development. Journal of Geophysical Research: Atmospheres , 116 (D8), D08208. https://doi.org/10.1029/2010JD015401 Alvares, C. A., Stape, J. L., Sentelhas, P. C., de Moraes Gonçalves, J. L., & Sparovek, G. (2013). Köppen’s climate classification map for Brazil. Meteorologische Zeitschrift , 22 (6), 711–728. https://doi.org/10.1127/0941-2948/2013/0507 Arthur, D., & Vassilvitskii, S. (2007). K-Means++: The advantages of careful seeding. In Proceedings of the eighteenth annual ACM-SIAM symposium on Discrete algorithms (pp. 1027–1035). Society for Industrial and Applied Mathematics. Bezerra, B. G., da Silva, L. L., dos Santos e Silva, C. M., & de Carvalho, G. G. (2019). Changes of precipitation extremes indices in São Francisco River Basin, Brazil from 1947 to 2012. Theoretical and Applied Climatology , 135 (1–2), 565–576. https://doi.org/10.1007/s00704-018-2396-6 Boccippio, D. J., Koshak, W. J., & Blakeslee, R. J. (2002). Performance Assessment of the Optical Transient Detector and Lightning Imaging Sensor. Part I: Predicted Diurnal Variability. Journal of Atmospheric and Oceanic Technology , 19 (9), 1318–1332. https://doi.org/10.1175/1520-0426(2002)0192.0.CO;2 Bourscheidt, V., Pinto, O., Jr., Naccarato, K. P., & Pinto, I. R. C. A. (2009). The influence of topography on the cloud-to-ground lightning density in South Brazil. Atmospheric Research , 91 (2–4), 508–513. https://doi.org/10.1016/j.atmosres.2008.06.010 Burgesser, R. E., Chameides, W. L., & De-Gaetano, A. T. (2014). Projected increase in lightning strikes in the United States due to global warming. Science , 346 (6211), 851-854. https://doi.org/10.1126/science.1259100 Cardoso, I., Naccarato, K. P., Pinto, O., Jr., & Pinto, I. R. C. A. (2014). Lightning casualty demographics in Brazil and their implications for safety rules. Atmospheric Research , 135–136 , 374–379. https://doi.org/10.1016/j.atmosres.2012.12.006 Cecil, D. J., Buechler, D. E., & Blakeslee, R. J. (2014). Gridded lightning climatology from TRMM-LIS and OTD: Dataset description. Atmospheric Research , 135–136 , 404–414. https://doi.org/10.1016/j.atmosres.2012.06.028 Celebi, M. E., Kingravi, H. A., & Vela, P. A. (2013). A comparative study of efficient initialization methods for the k-means clustering algorithm. Expert Systems with Applications , 40 (1), 200–210. https://doi.org/10.1016/j.eswa.2012.07.021 Christian, H. J., Blakeslee, R. J., & Goodman, S. J. (1992). Lightning Imaging Sensor (LIS) for the Earth Observing System (NASA Technical Memorandum 4350). NASA. Comitê da Bacia Hidrográfica do Rio São Francisco. (2016). Plano de Recursos Hídricos da Bacia Hidrográfica do Rio São Francisco 2016-2025 [Water Resources Plan for the São Francisco River Basin 2016-2025]. http://cbhsaofrancisco.org.br Costa, M. O., de Matos Quintela, M., de Jesus, R. M., Almeida, A., de Andrade, H., & Fia, F. R. L. (2017). Evaluation of violations in water quality standards in the monitoring network of São Francisco River basin, the third largest in Brazil. Environmental Monitoring and Assessment , 189 (11), 590. https://doi.org/10.1007/s10661-017-6266-y Dewan, A., Ongee, E. T., Rafiuddin, M., Rahman, M. M., & Mahmood, R. (2018). Spatial and temporal analysis of a 17-year lightning climatology over Bangladesh with LIS data. Theoretical and Applied Climatology , 134 (1–2), 347–362. https://doi.org/10.1007/s00704-017-2278-3 Diaz, F., Ortiz, D., & Roman, F. (2022). A climatological overview of lightning activity in Colombia using lightning imaging sensor (LIS) data. Theoretical and Applied Climatology , 149 (1), 69–82. https://doi.org/10.1007/s00704-022-04029-w Fernandes, W. A. (2006). Características dos relâmpagos gerados por nuvens de tempestades em ambientes sob a influência das queimadas em Rondônia [Characteristics of lightning generated by storm clouds in environments under the influence of fires in Rondônia] [Doctoral dissertation, Instituto Nacional de Pesquisas Espaciais]. Fujisada, H., Bailey, G. B., Kelly, G. G., Hara, S., & Abrams, M. J. (2005). ASTER DEM performance. IEEE Transactions on Geoscience and Remote Sensing , 43 (12), 2707–2714. https://doi.org/10.1109/TGRS.2005.847924 Funk, C., Peterson, P., Landsfeld, M., Pedreros, D., Verdin, J., Shukla, S., Husak, G., Rowland, J., Harrison, L., Hoell, A., & Michaelsen, J. (2014). A quasi-global precipitation time series for drought monitoring (U.S. Geological Survey Data Series 832). https://doi.org/10.3133/ds832 G1. (2013, March 4). Raios causam prejuízo de R$ 1 bilhão à economia do país, diz Inpe [Lightning causes R$ 1 billion loss to the country's economy, says Inpe]. G1. http://g1.globo.com/brasil/noticia/2013/03/raios-causam-prejuizo-de-r-1-bilhao-economia-do-pais-diz-inpe.html Jayaratne, E. R., & Kuleshov, Y. (2006). Geographical and seasonal characteristics of the relationship between lightning ground flash density and rainfall within the continent of Australia. Atmospheric Research , 79 (1), 1–14. https://doi.org/10.1016/j.atmosres.2005.03.004 Kodinaríya, T. M., & Makwana, P. R. (2013). Review on determining number of Cluster in K-Means Clustering. International Journal of Advance Research in Computer Science and Management Studies , 1 (6), 90–95. Kotroni, V., & Lagouvardos, K. (2008). Lightning occurrence in relation with elevation, terrain slope, and vegetation cover in the Mediterranean. Journal of Geophysical Research: Atmospheres , 113 (D21), D21113. https://doi.org/10.1029/2008JD010605 MacQueen, J. (1967). Some methods for classification and analysis of multivariate observations. In L. M. Le Cam & J. Neyman (Eds.), Proceedings of the fifth Berkeley symposium on mathematical statistics and probability, Volume 1: Theory of statistics (pp. 281–297). University of California Press. MapBiomas. (2024). Algorithm Theoretical Basis Document (ATBD) - Collection 9 (Version 2). Marengo, J. A., Tomasella, J., Alves, L. M., Soares, W. R., & Rodriguez, D. A. (2012). Extreme climatic events in the Amazon basin. Theoretical and Applied Climatology , 107 (1–2), 73–85. https://doi.org/10.1007/s00704-011-0465-1 Markowski, P., & Richardson, Y. (2010). Mesoscale meteorology in midlatitudes . John Wiley & Sons. Mattos, E. V., Machado, L. A. T., & Gonçalves, W. A. (2011). Cloud-to-ground lightning and Mesoscale Convective Systems. Atmospheric Research , 99 (3–4), 377–390. https://doi.org/10.1016/j.atmosres.2010.11.007 Mattos, E. V., da Silva, J. B., Machado, L. A. T., & Albrecht, R. I. (2024). Relâmpagos: gênese, tipos e impactos [Lightning: Genesis, types and impacts]. Revista Brasileira de Geografia Física , 17 (5), 3876–3905. https://doi.org/10.26848/rbgf.v17.5.p3876-3905 Mondal, S., Das, S., & Maitra, A. (2022). Spatio-temporal variability of lightning climatology and its association with thunderstorm indices over India. Theoretical and Applied Climatology , 149 (3-4), 1195–1212. https://doi.org/10.1007/s00704-022-04104-2 Mutti, P. R., de Medeiros, G. A., & de Assis de Souza Filho, F. (2022). Long-term meteorological drought characterization in the São Francisco watershed, Brazil: A climatic water balance approach. International Journal of Climatology , 42 (13), 6757–6775. https://doi.org/10.1002/joc.7701 Nesbitt, S. W., Zipser, E. J., & Cecil, D. J. (2000). A Census of Precipitation Features in the Tropics Using TRMM: Radar, Ice Scattering, and Lightning Observations. Journal of Climate , 13 (23), 4087–4106. https://doi.org/10.1175/1520-0442(2000)0132.0.CO;2 Oda, F. S. S., Naccarato, K. P., Pinto, O., Jr., & Holle, R. L. (2022). An initial assessment of the distribution of total Flash Rate Density (FRD) in Brazil from GOES-16 Geostationary Lightning Mapper (GLM) observations. Atmospheric Research , 270 , 106081. https://doi.org/10.1016/j.atmosres.2022.106081 Peña, J. M., Lozano, J. A., & Larrañaga, P. (1999). An empirical comparison of four initialization methods for the K-Means algorithm. Pattern Recognition Letters , 20 (10), 1027–1040. https://doi.org/10.1016/S0167-8655(99)00069-0 Pereira, R. G., Naccarato, K. P., & Pinto, O., Jr. (2025). Temporal Evolution of Lightning Properties in the Metropolitan Area of São Paulo (MASP) During the CHUVA-Vale Campaign. Atmosphere , 16 (4), 426. https://doi.org/10.3390/atmos16040426 Pielke, R. A., Sr. (2001). Influence of the spatial distribution of vegetation and soils on the prediction of cumulus convective rainfall. Reviews of Geophysics , 39 (2), 151–177. https://doi.org/10.1029/1999RG000072 Pinto Junior, O., & Pinto, I. R. C. A. (2008). Mortes por raios no Brasil: uma nova estimativa e considerações sobre a sua distribuição, sazonalidade e letalidade [Deaths by lightning in Brazil: A new estimate and considerations about its distribution, seasonality and lethality]. Revista Brasileira de Geofísica , 26 (1), 77–87. Potdar, S. S., Singh, D., & Singh, R. (2025). Influence of land use land cover and topography on lightning distribution over north and north-east Indian region. Journal of Atmospheric and Solar-Terrestrial Physics , 268 , 106460. https://doi.org/10.1016/j.jastp.2025.106460 Qie, X., Toumi, R., & Yuan, T. (2003). Lightning activities on the Tibetan Plateau as observed by the lightning imaging sensor. Journal of Geophysical Research: Atmospheres , 108 (D17), 4553. https://doi.org/10.1029/2002JD003304 Rasmussen, K. L., & Houze, R. A., Jr. (2011). Orogenic convection in subtropical South America as seen by the TRMM satellite. Monthly Weather Review , 139 (8), 2399–2420. https://doi.org/10.1175/MWR-D-10-05006.1 Reynolds, S. E., Brook, M., & Gourley, M. F. (1957). Thunderstorm Charge Separation. Journal of Meteorology , 14 (5), 426–436. https://doi.org/10.1175/1520-0469(1957)0142.0.CO;2 Roe, G. H. (2005). Orographic Precipitation. Annual Review of Earth and Planetary Sciences , 33 , 645–671. https://doi.org/10.1146/annurev.earth.33.092203.122541 Romatschke, U., & Houze, R. A., Jr. (2010). Extreme Summer Convection in South America. Journal of Climate , 23 (14), 3761–3791. https://doi.org/10.1175/2010JCLI3465.1 Rosenfeld, D., Lohmann, U., Rissman, G. B., Flagan, R. C., Seinfeld, J. H., Andreae, M. O., Asmi, E., & Jurkat, T. (2008). Flood or drought: how do aerosols affect precipitation? Science , 321 (5894), 1309–1313. https://doi.org/10.1126/science.1160606 dos Santos, A. P. P., da S. S. de Souza, E. B., & dos Santos, H. L. C. (2024). Lightning under Different Land Use and Cover, and the Influence of Topography in the Carajás Mineral Province, Eastern Amazon. Atmosphere , 15 (3), 375. https://doi.org/10.3390/atmos15030375 Saunders, C. (2008). Charge Separation Mechanisms in Clouds. Space Science Reviews , 137 (1–4), 335–353. https://doi.org/10.1007/s11214-008-9345-0 Schneider, L., Eremenko, M., Cuesta, J., Dufour, G., & Foret, G. (2019). Relative impact of aerosol, soil moisture, and orography perturbations on deep convection. Atmospheric Chemistry and Physics , 19 (19), 12343–12359. https://doi.org/10.5194/acp-19-12343-2019 Simões, F. T. A. (2023). Sensibilidade da vazão média anual ao clima na bacia hidrográfica do Rio São Francisco [Sensitivity of the average annual flow to the climate in the São Francisco river basin] [Undergraduate thesis, Universidade de Brasília]. Taylor, C. M., Gounou, A., Guichard, F., Harris, P. P., Ellis, R. J., Couvreux, F., & De Kauwe, M. (2011). Frequency of Sahelian storm initiation enhanced over mesoscale soil-moisture patterns. Nature Geoscience , 4 (7), 430–433. https://doi.org/10.1038/ngeo1173 Unfer, G. R., Dias, N. L., Chamecki, M., Martins, L. D., & Gonçalves, W. A. (2025). Decoding the Relationship Between Cloud Electrification, Downdrafts, and Surface Ozone in the Amazon Basin. Journal of Geophysical Research: Atmospheres , 130 (4), e2024JD042158. https://doi.org/10.1029/2024JD042158 Vogt, B. J., & Hodanish, S. J. (2016). A geographical analysis of warm season lightning/landscape interactions across Colorado, USA. Applied Geography , 75 , 93–103. https://doi.org/10.1016/j.apgeog.2016.08.006 Wang, Y., Fan, J., Zhang, R., Zhang, D., & Li, P. (2018). Aerosol microphysical effects on lightning activity under different cloud-dominated regimes. Atmospheric Chemistry and Physics , 18 (17), 12797–12809. https://doi.org/10.5194/acp-18-12797-2018 Whiteman, C. D. (2000). Mountain meteorology: Fundamentals and applications . Oxford University Press. Williams, E. R. (1989). The tripole structure of thunderstorms. Journal of Geophysical Research: Atmospheres , 94 (D11), 13151–13167. https://doi.org/10.1029/JD094iD11p13151 Wu, J. (2012). Cluster analysis and K-means clustering: An introduction. In J. He, X. Li, & X. Li (Eds.), Advances in K-means clustering (pp. 1–16). Springer. Zhang, D., Koshak, W., & Boccippio, D. (2019). Evaluation of the Performance Characteristics of the Lightning Imaging Sensor. Journal of Atmospheric and Oceanic Technology , 36 (6), 1015–1031. https://doi.org/10.1175/JTECH-D-18-0173.1 Zipser, E. J., Cecil, D. J., Liu, C., Nesbitt, S. W., & Yorty, D. P. (2006). Where are the most intense thunderstorms on earth? Bulletin of the American Meteorological Society , 87 (8), 1057–1071. https://doi.org/10.1175/BAMS-87-8-1057 Zoboli, V. G., & da Silva, T. A. (2023). Estudo da incidência de nuvens CBS no centro-oeste brasileiro: Uma avaliação decenal (2013-2023) a partir das cartas SIGWX PROG [Study of the incidence of CB clouds in the Brazilian midwest: A decennial evaluation (2013-2023) from SIGWX PROG charts]. Revista Brasileira de Aviação Civil & Ciências Aeronáuticas , 3 (5), 178–201. https://rbac.cia.emnuvens.com.br/revista/article/view/197 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 10 Nov, 2025 Reviews received at journal 10 Nov, 2025 Reviews received at journal 28 Oct, 2025 Reviews received at journal 28 Oct, 2025 Reviewers agreed at journal 12 Oct, 2025 Reviewers agreed at journal 10 Oct, 2025 Reviewers agreed at journal 25 Sep, 2025 Reviewers invited by journal 22 Sep, 2025 Editor assigned by journal 11 Sep, 2025 Submission checks completed at journal 10 Sep, 2025 First submitted to journal 08 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-7566962","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":512096344,"identity":"290158e5-e1cb-4bc5-ab32-ead15c47c853","order_by":0,"name":"Samuel Amorim Silva","email":"data:image/png;base64,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","orcid":"","institution":"Universidade Federal do Rio Grande do Norte (UFRN)","correspondingAuthor":true,"prefix":"","firstName":"Samuel","middleName":"Amorim","lastName":"Silva","suffix":""},{"id":512096345,"identity":"459edad7-aa43-4ca6-a150-fc92b7b39f0a","order_by":1,"name":"Weber Andrade Gonçalves","email":"","orcid":"","institution":"Universidade Federal do Rio Grande do Norte (UFRN)","correspondingAuthor":false,"prefix":"","firstName":"Weber","middleName":"Andrade","lastName":"Gonçalves","suffix":""},{"id":512096346,"identity":"c23156e0-0654-429f-b9d3-2f0bc70410bf","order_by":2,"name":"Lizando Pereira Abreu","email":"","orcid":"","institution":"Instituto Federal de Educação, Ciência e Tecnologia do Piauí (IFPI)","correspondingAuthor":false,"prefix":"","firstName":"Lizando","middleName":"Pereira","lastName":"Abreu","suffix":""},{"id":512096347,"identity":"2c336d1f-da77-4e4b-b7b3-c62e301b3600","order_by":3,"name":"Douglas Leonardo Sales Pedrosa","email":"","orcid":"","institution":"Universidade Federal do Rio Grande do Norte (UFRN)","correspondingAuthor":false,"prefix":"","firstName":"Douglas","middleName":"Leonardo Sales","lastName":"Pedrosa","suffix":""},{"id":512096348,"identity":"7b883084-09d7-4da8-8aa6-d2477047ff88","order_by":4,"name":"Evandro Moimaz Anselmo","email":"","orcid":"","institution":"Fundação Cearense de Meteorologia e Recursos Hídricos (FUNCEME)","correspondingAuthor":false,"prefix":"","firstName":"Evandro","middleName":"Moimaz","lastName":"Anselmo","suffix":""},{"id":512096349,"identity":"d7e6d7bb-c318-4833-904a-951763df5e80","order_by":5,"name":"Enrique Vieira Mattos","email":"","orcid":"","institution":"Universidade Federal de Itajubá (UNIFEI)","correspondingAuthor":false,"prefix":"","firstName":"Enrique","middleName":"Vieira","lastName":"Mattos","suffix":""},{"id":512096350,"identity":"c5afe418-c80a-4716-8cab-9a4bd926612c","order_by":6,"name":"Moisés Elias Nascimento Rufino Costa","email":"","orcid":"","institution":"Universidade Federal do Rio Grande do Norte (UFRN)","correspondingAuthor":false,"prefix":"","firstName":"Moisés","middleName":"Elias Nascimento Rufino","lastName":"Costa","suffix":""},{"id":512096351,"identity":"915d7c30-80d1-4892-ac8c-cc96d2af5dd1","order_by":7,"name":"Glenda Yasmin Pereira Carvalho","email":"","orcid":"","institution":"Universidade Federal do Rio Grande do Norte (UFRN)","correspondingAuthor":false,"prefix":"","firstName":"Glenda","middleName":"Yasmin Pereira","lastName":"Carvalho","suffix":""},{"id":512096353,"identity":"8896d0a8-0180-4544-9167-f5c13dfb3c04","order_by":8,"name":"Bruno Felipe Moreira Lima","email":"","orcid":"","institution":"Instituto Metrópole Digital (IMD)","correspondingAuthor":false,"prefix":"","firstName":"Bruno","middleName":"Felipe Moreira","lastName":"Lima","suffix":""}],"badges":[],"createdAt":"2025-09-08 18:38:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7566962/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7566962/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":91079996,"identity":"4009077a-8ee1-43f0-b720-fd46b0ccf516","added_by":"auto","created_at":"2025-09-11 11:28:20","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":13757339,"visible":true,"origin":"","legend":"\u003cp\u003eGeographic location of the São Francisco River Basin (BHSF), shown in red. The hydrographic sub-regions are highlighted: Upper (ASF) in black, Middle (MSF) in dark gray, Lower-Middle (SMSF) in gray, and Lower (BSF) in light gray.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7566962/v1/8e35ca1f39429962a6807245.png"},{"id":91078554,"identity":"5340d846-b39b-4bed-8a79-e0d0b445c44d","added_by":"auto","created_at":"2025-09-11 11:20:20","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":3836996,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of (a) lightning clustering based on flash rate density (FRD) in the BHSF, with hotspot locations shown (cluster 1 is blue, cluster 2 is green, cluster 3 is yellow, and cluster 4 is red); (b) topography (limited to 0–1700 m), with hotspot locations; (c) total flash rate density distribution (limited to 30 flashes km⁻² yr⁻¹), with hotspot locations; (d) predominant land use and land cover in the BHSF, with hotspot locations; (e) topographic slope, classified as flat (0–3%), gently rolling (3–8%), rolling (8–20%), strongly rolling (20–45%), mountainous (45–75%), and steep (\u0026gt;75%); and (f) a close-up view of the hotspot locations overlaid on the slope map.\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7566962/v1/17b0a8b85190064a91168074.jpg"},{"id":91079987,"identity":"d2f0d0a2-cc71-4462-8769-aea7cf4d0003","added_by":"auto","created_at":"2025-09-11 11:28:20","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1470495,"visible":true,"origin":"","legend":"\u003cp\u003eTemporal behavior of lightning in the BHSF based on the k-means clustering. (a) Boxplot of total lightning for each cluster, (b) hourly distribution of FRD, (c) monthly distribution of precipitation, and (d) monthly distribution of FRD. Cluster 1 is represented by blue, cluster 2 by green, cluster 3 by yellow, and cluster 4 by red.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7566962/v1/d65c9629f13bcace4da54f42.png"},{"id":91079997,"identity":"ef04a3d3-a8a0-451d-9e7d-fa817e26b5a9","added_by":"auto","created_at":"2025-09-11 11:28:20","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1745170,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of seasonal precipitation and FRD in the BHSF from 1998 to 2013. For December-January-February (DJF): (a) Precipitation and (b) FRD. For March-April-May (MAM): (c) Precipitation and (d) FRD. For June-July-August (JJA): (e) Precipitation and (f) FRD. For September-October-November (SON): (g) Precipitation and (h) FRD.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-7566962/v1/b9cbf9e6720acd7dc466815d.png"},{"id":91079992,"identity":"332dea56-2f69-4e8b-ad51-b28338022dae","added_by":"auto","created_at":"2025-09-11 11:28:20","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":7007397,"visible":true,"origin":"","legend":"\u003cp\u003eThree-dimensional (3D) model of the Digital Elevation Model (DEM) for the BHSF. The pixels corresponding to the São Desidério, Bahia hotspot (ranked 1st), which has the highest FRD in the BHSF (39.9 flashes km⁻² yr⁻¹), are highlighted by red circles. Panel (a) shows a horizontal 3D view of hotspot 1; (b) shows a 3D profile near this same point; and (c) shows the corresponding 2D topography. The red arrows indicate the flat plateau regions where the hotspot pixel is located, while the black arrow indicates the movement of weather systems that cause orographic precipitation in the region.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-7566962/v1/5e00e4d4dee0b4c0828649b6.png"},{"id":91081656,"identity":"4db53d84-2c53-4b86-b9d7-ac78749aeedd","added_by":"auto","created_at":"2025-09-11 11:44:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":33572529,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7566962/v1/c72c7a6b-e3d6-47bb-898e-01afa6ca1c6d.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Lightning Behavior and Its Relationship with Topography, Precipitation, and Land Use in the São Francisco River Basin","fulltext":[{"header":"1. INTRODUCTION","content":"\u003cp\u003eLightning is the result of electrification processes within deep convective clouds (Williams, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e1989\u003c/span\u003e; Mattos et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In these systems, strong updrafts promote collisions between ice particles and water droplets, leading to charge separation (Reynolds et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1957\u003c/span\u003e; Saunders, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). The frequency and intensity of these discharges are not random; they are directly modulated by atmospheric conditions and, crucially, by terrestrial surface factors (Fernandes, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). When these discharges reach the ground, they become a significant natural hazard, capable of causing serious incidents and fatalities (Pinto Junior \u0026amp; Pinto, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2008\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAmong the most influential surface drivers are topography and land use. Topography often acts as a triggering mechanism for convection by forcing moist air to rise, thereby intensifying storm formation (Bourscheidt et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Schneider et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Studies across various Brazilian biomes, from the Amazon to the southern regions, have shown that areas with steeper slopes, such as mountain ridges and hillsides, tend to exhibit higher lightning densities (Bourscheidt et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Santos et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Analogously, changes in land cover alter the surface energy and moisture balance, which can also influence lightning patterns (Potdar et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Santos et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This dynamic intrinsically links lightning to precipitation events, a relationship documented throughout Brazil (Abreu et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Mattos \u0026amp; Machado, 2011; Zoboli \u0026amp; Silva, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eWith an average of 96.4\u0026nbsp;million flashes annually, Brazil is one of the countries most affected by this phenomenon, exhibiting a fatality rate approximately four times higher than that of developed nations (Oda et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Cardoso et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). This national context contrasts with extreme global hotspots like Lake Maracaibo in Venezuela, which records the world's highest flash density at 232.52 flashes km⁻\u0026sup2; yr⁻\u0026sup1; (Albrecht et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). In Brazil, lightning causes an average of 132 fatalities per year and generates economic losses that can reach R\u003cspan\u003e$\u003c/span\u003e1\u0026nbsp;billion annually, primarily affecting the electricity and telecommunications sectors (Cardoso et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; G1, 2013).\u003c/p\u003e\u003cp\u003eWithin this context, the S\u0026atilde;o Francisco River Basin (BHSF) in Brazil emerges as an area of particular interest and concern due to its significant social and economic relevance. The S\u0026atilde;o Francisco River is a critical water source for the country, essential for irrigated agriculture, livestock farming, and hydroelectric power generation (Mutti et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Spanning multiple regions, the BHSF is situated in territories that collectively account for about 76% of lightning-related fatalities in Brazil, highlighting a pronounced vulnerability, especially for rural activities, which represent 19% of these deaths (Cardoso et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eHowever, investigating this phenomenon in the BHSF faces a methodological challenge: the ground-based lightning monitoring network has limited spatial coverage in the region (Abreu, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This limitation makes the use of remote sensing data, such as from the Lightning Imaging Sensor (LIS), a critical alternative, as it offers comprehensive and consistent spatiotemporal coverage over the basin, including in remote and hard-to-access areas (Abreu, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eConsidering the high frequency of lightning and fatalities within the regions of the BHSF (Cardoso et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Oda et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), the complex interaction between its surface drivers, and the existing gap in event monitoring, this study aims to characterize and analyze the spatiotemporal patterns of lightning in the basin. We will identify the areas of highest density (hotspots) and correlate them with topographic, precipitation, and land use variables, seeking to fill a knowledge gap and provide insights for risk mitigation strategies in the region.\u003c/p\u003e"},{"header":"2. MATERIALS AND METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1. Study Area\u003c/h2\u003e\u003cp\u003eThe S\u0026atilde;o Francisco River Basin (BHSF) is one of the most extensive and strategic drainage areas in Brazil, covering approximately 8% of the national territory (CBHSF, 2016). The river flows for 2,863 km through the Northeast, Southeast, and Center-West regions of the country. The basin spans 505 municipalities across six states: Minas Gerais (MG), Goi\u0026aacute;s (GO), Bahia (BA), Pernambuco (PE), Alagoas (AL), and Sergipe (SE), in addition to the Federal District (DF) (CBHSF, 2016; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). For planning and management purposes, the BHSF is divided into four sub-regions: Upper (ASF), Middle (MSF), Lower-Middle (SMSF), and Lower S\u0026atilde;o Francisco (BSF). This division is based on the distinct climatic and ecological characteristics along the river's course (Bezerra et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2. Data\u003c/h2\u003e\u003cp\u003eThis study utilized four primary datasets: lightning data from the Lightning Imaging Sensor (LIS), topographic data from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), precipitation data from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS), and land use and land cover data from the MapBiomas project.\u003c/p\u003e\u003cdiv id=\"Sec5\" class=\"Section3\"\u003e\u003ch2\u003e2.2.1. Lightning Data\u003c/h2\u003e\u003cp\u003eLightning data were obtained from the Lightning Imaging Sensor (LIS), which operated aboard the Tropical Rainfall Measurement Mission (TRMM) satellite between 1997 and 2015. The satellite maintained an orbit inclined at 35\u0026deg; and an altitude of 350 km, with the mission to detect the optical signatures of lightning (Albrecht et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Cecil et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). The sensor used a 128x128 pixel array and a bandpass filter centered at 777.4 nm to capture near-infrared oxygen emissions, a strong and consistent spectral signature of lightning that allows it to be distinguished from other light sources (Albrecht et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe high-resolution LIS climatology, with a 0.1\u0026deg; spatial resolution, is made available by the Global Hydrometeorology Resource Center (GHRC). The sensor's flash detection efficiency has been estimated at 93% at night and 73% during the day (Qie et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). Although the orbital nature of LIS can result in temporal subsampling\u0026mdash;meaning it does not observe a specific location continuously\u0026mdash;its extensive data collection period establishes it as one of the most important lightning databases for studies involving long time series (Albrecht et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). This is particularly true when compared to more recent products like the Geostationary Lightning Mapper (GLM), which has only provided data for the last seven years (2018\u0026ndash;2025) but offers far superior temporal coverage.\u003c/p\u003e\u003cp\u003eFor this study, the analysis period was set from 1998 to 2013. The exclusion of data from 2014 and 2015 is due to the start of the TRMM satellite's decommissioning process in 2014, which introduced uncertainties and interruptions in the observations. This is a frequent practice adopted in other studies using the same database (Abreu et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Boccippio et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Dewan et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Zhang et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section3\"\u003e\u003ch2\u003e2.2.2. Land Use and Land Cover Data\u003c/h2\u003e\u003cp\u003eMapBiomas is an initiative formed by a collaborative network of Brazilian ONGs, universities, and technology companies that uses time series of satellite imagery to map and monitor land cover and use across the entire Brazilian territory (MapBiomas, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Through machine learning algorithms and cloud processing, the platform categorizes the land into classes such as agriculture, pasture, natural vegetation, and urban areas, with high spatial and temporal resolution. This methodology allows for a detailed analysis of changes in land use over the study period (1998 to 2013), providing a robust foundation for studies relating terrestrial dynamics to atmospheric phenomena like lightning. The accuracy and comprehensiveness of MapBiomas make it an indispensable tool for land cover information in regional studies.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section3\"\u003e\u003ch2\u003e2.2.3. Topography and Elevation Data\u003c/h2\u003e\u003cp\u003eTopographic information was derived from data captured by the ASTER sensor. With its ability to capture stereoscopic images in the near-infrared band at a 15-meter spatial resolution, ASTER data allows for the generation of a Digital Elevation Model (DEM) for the study area. DEMs are fundamental inputs for analyzing orographic effects on atmospheric convection. The ASTER DEM has a vertical accuracy of 20 meters with 95% confidence, without the need for ground control points (Fujisada et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2005\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section3\"\u003e\u003ch2\u003e2.2.4. Precipitation Data\u003c/h2\u003e\u003cp\u003eFor the precipitation analysis, the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) dataset was used. This product, developed by the U.S. Geological Survey (USGS) and the University of California, Santa Barbara (UCSB), combines satellite estimates with rain gauge station data to create a detailed and reliable historical time series extending from 1981 to the present (Funk et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). This hybrid approach is particularly powerful as it merges the broad spatial coverage of satellites with the ground-truthed accuracy of in-situ measurements. For this study, CHIRPS data with a spatial resolution of 0.05\u0026deg; (~\u0026thinsp;5 km) and at a monthly aggregation scale were used, making them suitable for the proposed climatological analysis.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e2.3. METHODOLOGY\u003c/h2\u003e\u003cp\u003eThe methodology was structured into four main steps: (1) processing and standardization of the datasets; (2) cluster analysis to identify spatial lightning patterns within the basin; (3) identification and characterization of areas with maximum activity (hotspots); and (4) analysis of the relationship between hotspots and surface drivers (topography and land use). This multi-step approach ensures comprehensive analysis, progressing from initial data preparation to the final physical interpretation of the results.\u003c/p\u003e\u003cdiv id=\"Sec10\" class=\"Section3\"\u003e\u003ch2\u003e2.3.1. Data Processing and Standardization\u003c/h2\u003e\u003cp\u003eDue to the differing spatial resolutions of the source data, it was necessary to standardize all information onto a common grid. This is a critical step in geospatial analysis, as it allows for direct, pixel-by-pixel comparison between different environmental variables. A sample grid was created based on the native 0.1\u0026deg; resolution of the LIS sensor. This grid was then used to extract corresponding values for land use and precipitation. For each LIS grid cell, the predominant land use class from the high-resolution MapBiomas data (30 m) and the corresponding precipitation value from the CHIRPS dataset were assigned, ensuring spatial compatibility across all datasets.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section3\"\u003e\u003ch2\u003e2.3.2. K-means Clustering of Total Lightning\u003c/h2\u003e\u003cp\u003eTo understand the spatial structure of the lightning data, the k-means clustering technique was employed. This algorithm, first proposed by MacQueen (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e1967\u003c/span\u003e), is an iterative clustering method that partitions a dataset into k predefined clusters, such that the distance from each point to its cluster's centroid is minimized (Pe\u0026ntilde;a et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e1999\u003c/span\u003e). The goal of k-means is to find the optimal partition that minimizes the objective function, defined as the sum of squared distances between the points and their respective centroids (Wu, \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). The algorithm groups data points with similar values around a central point, with each resulting cluster representing a distinct pattern of lightning activity in the data.\u003c/p\u003e\u003cp\u003eThe choice of k-means for this analysis was driven by its computational efficiency and simplicity, which are particularly advantageous for large-scale datasets (Arthur \u0026amp; Vassilvitskii, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). However, it is acknowledged that the method has limitations: it requires the number of clusters (k) to be defined beforehand and is sensitive to the initial placement of centroids, which can lead to suboptimal results (Celebi et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn this study, the total lightning flash rate density was used as the input variable for the algorithm. Groupings ranging from two to ten clusters were evaluated. The final configuration of four clusters was chosen with the aid of the elbow method and the silhouette index, which are statistical techniques used to estimate the optimal number of clusters in a dataset (Kodinariya \u0026amp; Makwana, 2013). For the dissimilarity measure, both Euclidean and squared Euclidean distances were compared, with the former being selected for its simplicity and effectiveness with the data in question (Pe\u0026ntilde;a et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e1999\u003c/span\u003e). This approach differs from that of Abreu et al. (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e); while the latter used seasonal data for clustering, this study used the total lightning density. This approach allows for an analysis focused on the intrinsic, year-round spatial patterns of lightning within the basin.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section3\"\u003e\u003ch2\u003e2.3.3. Hotspot Identification and Characterization\u003c/h2\u003e\u003cp\u003eAreas with the highest lightning density (hotspots) were identified by sorting the total LIS lightning climatology data in descending order, which allowed for the precise localization of coordinates with the highest flash rates in the basin (Albrecht et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). For a detailed individual analysis of each hotspot, multiple temporal scales of LIS data were used: the single band of total lightning density, the 12 bands corresponding to the monthly average, and the 24 bands for the hourly average, all for the 1998\u0026ndash;2013 period. This multi-temporal analysis enables a deeper understanding of each hotspot's behavior, revealing not only where it is most active, but also when throughout the day and year. Numerical values were extracted from sample points matching the original sensor's pixel size, enabling the calculation of descriptive statistics such as mean, standard deviation, and maximum/minimum values.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section3\"\u003e\u003ch2\u003e2.3.4. Analysis of Surface Factors and 3D Modeling\u003c/h2\u003e\u003cp\u003eFollowing a methodology based on the work of Bourscheidt et al. (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) and Abreu et al. (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), ASTER data were used to generate a Digital Elevation Model (DEM). From this DEM, slope maps were calculated in QGIS software, and three-dimensional (3D) models were constructed using the open-source software Blender. This 3D modeling approach, differing from that of Abreu et al. (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), allowed for a detailed graphical analysis of the influence of topographic features on the spatial distribution of lightning. This visual approach offers a more intuitive understanding of the complex spatial relationships between terrain and lightning hotspots than traditional 2D maps can provide.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"3. RESULTS AND DISCUSSIONS","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n \u003ch2\u003e3.1. Clustering and Spatial Distribution of Lightning\u003c/h2\u003e\n \u003cp\u003eThe application of the K-means clustering algorithm proved to be an effective tool for segmenting the electrical activity in the basin, resulting in four clusters with distinct lightning regimes (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ea). Clusters 1, 2, and 3 exhibit a relatively homogeneous spatial distribution that aligns with the BHSF sub-regions: cluster 1 is concentrated in the Lower and Lower-Middle S\u0026atilde;o Francisco sections, cluster 2 predominantly covers the Middle S\u0026atilde;o Francisco, and cluster 3 is located in the Upper S\u0026atilde;o Francisco. This correspondence suggests that macroclimatic and geographic factors, which define the sub-basins themselves, modulate large-scale electrical activity (Marengo et al., \u003cspan class=\"CitationRef\"\u003e2012\u003c/span\u003e). In contrast, cluster 4, which aggregates the pixels with the highest density, exhibits a localized distribution, primarily superimposed over areas belonging to clusters 2 and 3. This configuration indicates the strong influence of local-scale forcings, such as variations in topography and land use, in the genesis of severe storms, highlighting a hierarchy of atmospheric controls, from large-scale climate to local terrain effects. This observation is consistent with studies that identify lightning hotspots anchored by specific geographic features (Albrecht et al., \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e; Diaz et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThe analysis of Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eb demonstrates that lightning hotspots are concentrated in the most elevated areas of the basin, notably in its western portion, where the flash density exceeds 25 flashes km⁻\u0026sup2; yr⁻\u0026sup1;. This region is characterized by the predominance of agriculture (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ed) and by a plateau topography. The association between electrical activity and elevated areas is a well-documented phenomenon, as orography intensifies atmospheric convection by providing the necessary mechanical lift for the formation of thunderstorms (Kotroni \u0026amp; Lagouvardos, \u003cspan class=\"CitationRef\"\u003e2008\u003c/span\u003e; Mondal et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). The eastern portion of the BHSF, despite also having elevated topography, records a significantly lower lightning density (less than 15 flashes km⁻\u0026sup2; yr⁻\u0026sup1;). This pattern suggests that the eastern edge of the basin acts as an orographic barrier, placing the study area on the leeward side of prevailing moisture systems. This \u0026quot;rain shadow\u0026quot; effect results in warmer and drier atmospheric conditions that are less conducive to electrification (Roe, \u003cspan class=\"CitationRef\"\u003e2005\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThe terrain slope also proves to be a crucial modulating factor (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ee). Precipitation-inducing systems, upon encountering the rugged topography in the central portion of the basin, undergo orographic uplift, which enhances instability and favors storm formation (Roe, \u003cspan class=\"CitationRef\"\u003e2005\u003c/span\u003e; Whiteman, \u003cspan class=\"CitationRef\"\u003e2000\u003c/span\u003e). Notably, the highest lightning density values predominantly occur after the systems pass over this zone of steeper slopes. A detailed analysis of the hotspot locations (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ef) corroborates this observation, showing that the majority are not located directly on the steepest slopes but rather on adjacent plateau areas. This pattern suggests that the convection, once initiated by the orography, subsequently propagates and reaches maturity over the flatter, elevated areas. Hotspot 5, however, is an exception, as it is located directly over a steep region, indicating that at this point, the intensification of convection may be more immediately anchored by the steep slope itself (Markowski \u0026amp; Richardson, \u003cspan class=\"CitationRef\"\u003e2010\u003c/span\u003e).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\n \u003ch2\u003e3.2. Spatiotemporal Characteristics of Lightning\u003c/h2\u003e\n \u003cp\u003eStatistical analysis of the clusters quantifies the differences between the lightning regimes (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ea). Cluster 1, located in the north, exhibits the lowest mean density (2.86 flashes km⁻\u0026sup2; yr⁻\u0026sup1;). Clusters 2 (5.86 flashes km⁻\u0026sup2; yr⁻\u0026sup1;) and 3 (8.37 flashes km⁻\u0026sup2; yr⁻\u0026sup1;) represent intermediate and transitional conditions. Cluster 4 stands out with the highest mean density (16.68 flashes km⁻\u0026sup2; yr⁻\u0026sup1;) and the largest standard deviation (\u0026sigma;\u0026thinsp;=\u0026thinsp;4.10), concentrating the extreme pixels (outliers). This high variability is characteristic of regions where convection is not only intense but also intermittent, driven by localized, powerful storm cells (Albrecht et al., \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThe hourly distribution (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eb) reveals a well-defined diurnal cycle, with a peak in activity concentrated in the late afternoon, between 14:00 and 20:00 UTC. This pattern is typical of continental tropical and subtropical regions, where maximum surface heating from the sun intensifies atmospheric instability and convection, culminating in thunderstorm formation (Burgesser et al., \u003cspan class=\"CitationRef\"\u003e2014\u003c/span\u003e; Mondal et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). This is the classic signature of \u0026quot;air-mass thunderstorms,\u0026quot; which build up during the day and are fueled by solar radiation. During the early morning hours, residual activity is observed in clusters 2 and 3, possibly associated with the dissipation of Mesoscale Convective Systems (MCSs) that typically form during the preceding evening (Nesbitt et al., \u003cspan class=\"CitationRef\"\u003e2000\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThe seasonality of electrical activity (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ed) directly follows the region\u0026apos;s rainfall cycle (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ec), with a maximum in summer (December-January-February, DJF) and a minimum in winter (June-July-August, JJA). This strong correlation highlights that the availability of moisture and the dynamics of seasonal weather systems, such as the South Atlantic Convergence Zone (SACZ), are the primary drivers of storm activity in the BHSF (Jayaratne \u0026amp; Kuleshov, \u003cspan class=\"CitationRef\"\u003e2006\u003c/span\u003e; Vogt \u0026amp; Hodanish, \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThis strong seasonal agreement is clearly visualized in Fig. 4. During the summer (DJF), there is significant spatial variation, with minimum lightning records in regions closer to the coast (eastern BHSF) and maximums in the interior, notably in the state of Bahia (Figs. 4a and 4b). In stark contrast, winter (JJA) is the season of lowest electrical activity, with a near-total absence of lightning across much of the BHSF, except for its southern portion (Figs. 4e and 4f). The spring (SON) and autumn (MAM) months clearly represent transitional seasons (Figs. 4c, 4d, 4g, 4h). During the winter dry period, the near-zero lightning and precipitation values in the central basin are attributable to the large-scale reduction in convective activity and atmospheric moisture (Nesbitt et al., \u003cspan class=\"CitationRef\"\u003e2000\u003c/span\u003e; Zipser et al., \u003cspan class=\"CitationRef\"\u003e2006\u003c/span\u003e), reinforcing the influence of climatic and topographic factors on the seasonal distribution of lightning, as noted by Bourscheidt et al. (\u003cspan class=\"CitationRef\"\u003e2009\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eRasmussen and Houze (\u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e) identified and mapped deep and extensive convective cores in South America, with a particular focus on the summer period. A comparison between their results and those of the present study reveals overlaps between some of these convective cores and the lightning hotspots identified in the BHSF. This spatial agreement suggests a strong link between intense convective activity and the occurrence of electrical discharges during the summer in the basin. This correlation is further corroborated by studies from Zipser et al. (\u003cspan class=\"CitationRef\"\u003e2006\u003c/span\u003e) and Romatschke and Houze (\u003cspan class=\"CitationRef\"\u003e2010\u003c/span\u003e), which emphasize the role of deep convection in lightning generation, especially in tropical and subtropical regions. This connection reinforces the physical basis of our statistically-identified hotspots, grounding them in the known meteorology of the region.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\n \u003ch2\u003e3.3 Analysis of Lightning Hotspots\u003c/h2\u003e\n \u003cp\u003eTen hotspots of maximum lightning concentration were identified in the BHSF, all belonging to cluster 4 (Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). The most intense hotspot, located in S\u0026atilde;o Desid\u0026eacute;rio, Bahia, recorded a rate of 39.9 flashes km⁻\u0026sup2; yr⁻\u0026sup1;. Although this value is lower than those of major global hotspots, such as Lake Maracaibo in Venezuela (233 flashes km⁻\u0026sup2; yr⁻\u0026sup1;) and the Congo Basin (205 flashes km⁻\u0026sup2; yr⁻\u0026sup1;) (Albrecht et al., \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e), it is comparable to other highly active regions in South America, such as areas in Colombia that record average rates of up to 70 flashes km⁻\u0026sup2; yr⁻\u0026sup1; (Diaz et al.,\u0026nbsp;\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). This places the BHSF\u0026apos;s most active areas as significant regional centers of intense convective activity.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\n \u003ctable id=\"Tab1\" border=\"1\" class=\"fr-table-selection-hover\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eLightning Hotspots in the BHSF.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRanking\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCity\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eState\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePoint Coordinate\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Land\\:use\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCluster\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eTopography (m)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:flash\\:{km}^{-2}{ano}^{-1}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eS\u0026atilde;o Desid\u0026eacute;rio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12\u0026deg;31\u0026apos;48.0\u0026quot;S 46\u0026deg;07\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAgriculture\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e515,0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e39,9\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLu\u0026iacute;s Eduardo Magalh\u0026atilde;es\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11\u0026deg;55\u0026apos;48.0\u0026quot;S 45\u0026deg;55\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAgriculture\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e542,5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e38,8\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCorrentina\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e13\u0026deg;19\u0026apos;48.0\u0026quot;S 45\u0026deg;37\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAgriculture\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e419,3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e35,4\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLu\u0026iacute;s Eduardo Magalh\u0026atilde;es\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12\u0026deg;25\u0026apos;48.0\u0026quot;S 46\u0026deg;07\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAgriculture\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e517,2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e34,9\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFormosa do Rio Preto\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11\u0026deg;19\u0026apos;48.0\u0026quot;S 46\u0026deg;31\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePasture\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e677,8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e32,0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLu\u0026iacute;s Eduardo Magalh\u0026atilde;es\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12\u0026deg;01\u0026apos;48.0\u0026quot;S 45\u0026deg;55\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eForest Formation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e507,1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e30,5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFormoso\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e14\u0026deg;49\u0026apos;48.0\u0026quot;S 46\u0026deg;25\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eForest Formation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1182,4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e29,7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSanta F\u0026eacute; de Minas\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e16\u0026deg;37\u0026apos;12.0\u0026quot;S 45\u0026deg;13\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eForest Formation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e479,9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e29,6\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBuritis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15\u0026deg;07\u0026apos;48.0\u0026quot;S 46\u0026deg;49\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGrassland\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e454,8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e29,2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTr\u0026ecirc;s Marias\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e18\u0026deg;01\u0026apos;12.0\u0026quot;S 45\u0026deg;01\u0026apos;48.0\u0026quot;W\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eForest Formation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e603,0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e29,2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eThe S\u0026atilde;o Desid\u0026eacute;rio hotspot (Fig. 5) reveals that maximum activity does not occur on the topographic peaks themselves, but on plateau areas near depressions. This pattern suggests that terrain-induced moisture convergence, combined with mesoscale circulations generated by the differential heating between the plateau and the valleys, may be a key mechanism for the initiation and anchoring of convection (Bourscheidt et al., \u003cspan class=\"CitationRef\"\u003e2009\u003c/span\u003e; Whiteman, \u003cspan class=\"CitationRef\"\u003e2000\u003c/span\u003e). Hotspots 7 and 10, in turn, are situated in plains surrounded by steep terrain, indicating that the channeling of moisture flows and the convergence of valley-mountain breezes may be the dominant triggers in these locations (Albrecht et al., \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eLand use also stands out as a modulating factor. The four most active hotspots are located in areas of intensive agriculture (Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). The conversion of native vegetation to agriculture alters the surface energy and moisture balance. Although the relationship is complex, studies indicate that landscape heterogeneity, such as the boundaries between agricultural fields and native vegetation, can create temperature and moisture gradients that induce local circulations and trigger convection (Pielke, \u003cspan class=\"CitationRef\"\u003e2001\u003c/span\u003e; Taylor et al., \u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e). Furthermore, agricultural activities, such as biomass burning, can inject aerosols into the atmosphere, which, under certain moisture conditions, can invigorate cloud updrafts and enhance lightning production (Rosenfeld et al., \u003cspan class=\"CitationRef\"\u003e2008\u003c/span\u003e; Wang et al., \u003cspan class=\"CitationRef\"\u003e2018\u003c/span\u003e). Therefore, the high incidence of lightning in western Bahia appears to be the result of a synergistic interaction between the favorable topography of the Western Bahia Plateau, seasonal moisture availability, and the surface alterations imposed by agricultural land use.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"4. CONCLUSION","content":"\u003cp\u003eThis study concludes that lightning activity in the BHSF is spatially heterogeneous and can be segmented into four distinct regimes. The large-scale distribution is modulated by macroclimatic factors, while intense convective activity is concentrated in 10 well-defined hotspots controlled by local-scale forcings. The temporal behavior of lightning follows a clear pattern, with a diurnal peak in the late afternoon driven by surface heating and a seasonal cycle with a maximum in the summer, strongly coupled with the region's rainy season.\u003c/p\u003e\u003cp\u003eThe lightning hotspots are primarily associated with specific topographic features. The majority are located on the Western Bahia Plateau, where convection is intensified by the interaction of the plateau topography with surrounding valleys and depressions. Orographic uplift and terrain-induced mesoscale circulations are identified as the primary triggering mechanisms. Furthermore, agricultural land use emerges as a key modulating factor, with the most intense hotspots situated directly over cultivated areas. It is concluded that while topography acts as the primary trigger for storm initiation, the surface conditions modified by agriculture likely contribute to the intensification of these storms.\u003c/p\u003e\u003cp\u003eThe primary contribution of this work is the demonstration of the synergistic interaction among topography, land use, and atmospheric dynamics in determining the storm regime of the BHSF. These findings are relevant for improving severe weather warning systems, for electrical and agricultural infrastructure planning, and for a better understanding of hydrometeorological processes in one of Brazil's most important river basins. Looking ahead, future research could employ high-resolution numerical modeling to quantify the relative contribution of each of these factors to the formation of lightning hotspots, an essential step for developing more precise predictive models in a context of ongoing climate and land-use change.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e This work was partially supported by the National Council for Scientific and Technological Development (CNPq) [Grant number 312707/2021-5].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e The authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e All authors contributed to the study conception and design. The methodology, software development, data curation and formal analysis were performed by Samuel Amorim Silva, with contributions from Weber Andrade Gon\u0026ccedil;alves, Lizando Pereira de Abreu and Douglas Leonardo Sales Pedrosa. The first draft of the manuscript was written by Samuel Amorim Silva and all authors commented on previous versions of the manuscript. Supervision was provided by Weber Andrade Gon\u0026ccedil;alves. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e Lightning data are freely available from NASA EARTHDATA (DOI: 10.5067/LIS/LIS/DATA304). CHIRPS precipitation data are freely available from the USGS Climate Hazards Center (CHC) at UC Santa Barbara (https://data.chc.ucsb.edu/products/CHIRPS-2.0/). Key processed datasets generated during this study are available on Zenodo (DOI: 10.5281/zenodo.15588150). MapBiomas data can be downloaded from the MapBiomas Brazil project website (https://brasil.mapbiomas.org/downloads/).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAbreu, L. P. (2023). \u003cem\u003eCaracteriza\u0026ccedil;\u0026atilde;o dos rel\u0026acirc;mpagos ocorridos na regi\u0026atilde;o Nordeste do Brasil, por meio de sensoriamento remoto\u003c/em\u003e [Characterization of lightning in the Northeast region of Brazil using remote sensing] [Doctoral dissertation, Universidade Federal do Rio Grande do Norte].\u003c/li\u003e\n\u003cli\u003eAbreu, L. P., Gon\u0026ccedil;alves, W. A., Mattos, E. V., \u0026amp; Albrecht, R. I. (2020). Assessment of the total lightning flash rate density (FRD) in northeast Brazil (NEB) based on TRMM orbital data from 1998 to 2013. \u003cem\u003eInternational Journal of Applied Earth Observation and Geoinformation\u003c/em\u003e, \u003cem\u003e93\u003c/em\u003e, 102195. https://doi.org/10.1016/j.jag.2020.102195\u003c/li\u003e\n\u003cli\u003eAlbrecht, R. I., Goodman, S. J., Buechler, D. E., Blakeslee, R. J., \u0026amp; Christian, H. J. (2016). Where are the lightning hotspots on Earth? \u003cem\u003eBulletin of the American Meteorological Society\u003c/em\u003e, \u003cem\u003e97\u003c/em\u003e(11), 2051\u0026ndash;2068. https://doi.org/10.1175/BAMS-D-14-00193.1\u003c/li\u003e\n\u003cli\u003eAlbrecht, R. I., Morales, C. A., \u0026amp; Silva Dias, M. A. F. (2011). Electrification of precipitating systems over the Amazon: Physical processes of thunderstorm development. \u003cem\u003eJournal of Geophysical Research: Atmospheres\u003c/em\u003e, \u003cem\u003e116\u003c/em\u003e(D8), D08208. https://doi.org/10.1029/2010JD015401\u003c/li\u003e\n\u003cli\u003eAlvares, C. A., Stape, J. L., Sentelhas, P. C., de Moraes Gon\u0026ccedil;alves, J. L., \u0026amp; Sparovek, G. (2013). K\u0026ouml;ppen\u0026rsquo;s climate classification map for Brazil. \u003cem\u003eMeteorologische Zeitschrift\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(6), 711\u0026ndash;728. https://doi.org/10.1127/0941-2948/2013/0507\u003c/li\u003e\n\u003cli\u003eArthur, D., \u0026amp; Vassilvitskii, S. (2007). K-Means++: The advantages of careful seeding. In \u003cem\u003eProceedings of the eighteenth annual ACM-SIAM symposium on Discrete algorithms\u003c/em\u003e (pp. 1027\u0026ndash;1035). Society for Industrial and Applied Mathematics.\u003c/li\u003e\n\u003cli\u003eBezerra, B. G., da Silva, L. L., dos Santos e Silva, C. M., \u0026amp; de Carvalho, G. G. (2019). Changes of precipitation extremes indices in S\u0026atilde;o Francisco River Basin, Brazil from 1947 to 2012. \u003cem\u003eTheoretical and Applied Climatology\u003c/em\u003e, \u003cem\u003e135\u003c/em\u003e(1\u0026ndash;2), 565\u0026ndash;576. https://doi.org/10.1007/s00704-018-2396-6\u003c/li\u003e\n\u003cli\u003eBoccippio, D. J., Koshak, W. J., \u0026amp; Blakeslee, R. J. (2002). Performance Assessment of the Optical Transient Detector and Lightning Imaging Sensor. Part I: Predicted Diurnal Variability. \u003cem\u003eJournal of Atmospheric and Oceanic Technology\u003c/em\u003e, \u003cem\u003e19\u003c/em\u003e(9), 1318\u0026ndash;1332. https://doi.org/10.1175/1520-0426(2002)019\u0026lt;1318:PAOTOT\u0026gt;2.0.CO;2\u003c/li\u003e\n\u003cli\u003eBourscheidt, V., Pinto, O., Jr., Naccarato, K. P., \u0026amp; Pinto, I. R. C. A. (2009). The influence of topography on the cloud-to-ground lightning density in South Brazil. \u003cem\u003eAtmospheric Research\u003c/em\u003e, \u003cem\u003e91\u003c/em\u003e(2\u0026ndash;4), 508\u0026ndash;513. https://doi.org/10.1016/j.atmosres.2008.06.010\u003c/li\u003e\n\u003cli\u003eBurgesser, R. E., Chameides, W. L., \u0026amp; De-Gaetano, A. T. (2014). Projected increase in lightning strikes in the United States due to global warming. \u003cem\u003eScience\u003c/em\u003e, \u003cem\u003e346\u003c/em\u003e(6211), 851-854. https://doi.org/10.1126/science.1259100\u003c/li\u003e\n\u003cli\u003eCardoso, I., Naccarato, K. P., Pinto, O., Jr., \u0026amp; Pinto, I. R. C. A. (2014). Lightning casualty demographics in Brazil and their implications for safety rules. \u003cem\u003eAtmospheric Research\u003c/em\u003e, \u003cem\u003e135\u0026ndash;136\u003c/em\u003e, 374\u0026ndash;379. https://doi.org/10.1016/j.atmosres.2012.12.006\u003c/li\u003e\n\u003cli\u003eCecil, D. J., Buechler, D. E., \u0026amp; Blakeslee, R. J. (2014). Gridded lightning climatology from TRMM-LIS and OTD: Dataset description. \u003cem\u003eAtmospheric Research\u003c/em\u003e, \u003cem\u003e135\u0026ndash;136\u003c/em\u003e, 404\u0026ndash;414. https://doi.org/10.1016/j.atmosres.2012.06.028\u003c/li\u003e\n\u003cli\u003eCelebi, M. E., Kingravi, H. A., \u0026amp; Vela, P. A. (2013). A comparative study of efficient initialization methods for the k-means clustering algorithm. \u003cem\u003eExpert Systems with Applications\u003c/em\u003e, \u003cem\u003e40\u003c/em\u003e(1), 200\u0026ndash;210. https://doi.org/10.1016/j.eswa.2012.07.021\u003c/li\u003e\n\u003cli\u003eChristian, H. J., Blakeslee, R. J., \u0026amp; Goodman, S. J. (1992). \u003cem\u003eLightning Imaging Sensor (LIS) for the Earth Observing System\u003c/em\u003e (NASA Technical Memorandum 4350). NASA.\u003c/li\u003e\n\u003cli\u003eComit\u0026ecirc; da Bacia Hidrogr\u0026aacute;fica do Rio S\u0026atilde;o Francisco. (2016). \u003cem\u003ePlano de Recursos H\u0026iacute;dricos da Bacia Hidrogr\u0026aacute;fica do Rio S\u0026atilde;o Francisco 2016-2025\u003c/em\u003e [Water Resources Plan for the S\u0026atilde;o Francisco River Basin 2016-2025]. http://cbhsaofrancisco.org.br\u003c/li\u003e\n\u003cli\u003eCosta, M. O., de Matos Quintela, M., de Jesus, R. M., Almeida, A., de Andrade, H., \u0026amp; Fia, F. R. L. (2017). Evaluation of violations in water quality standards in the monitoring network of S\u0026atilde;o Francisco River basin, the third largest in Brazil. \u003cem\u003eEnvironmental Monitoring and Assessment\u003c/em\u003e, \u003cem\u003e189\u003c/em\u003e(11), 590. https://doi.org/10.1007/s10661-017-6266-y\u003c/li\u003e\n\u003cli\u003eDewan, A., Ongee, E. T., Rafiuddin, M., Rahman, M. M., \u0026amp; Mahmood, R. (2018). Spatial and temporal analysis of a 17-year lightning climatology over Bangladesh with LIS data. \u003cem\u003eTheoretical and Applied Climatology\u003c/em\u003e, \u003cem\u003e134\u003c/em\u003e(1\u0026ndash;2), 347\u0026ndash;362. https://doi.org/10.1007/s00704-017-2278-3\u003c/li\u003e\n\u003cli\u003eDiaz, F., Ortiz, D., \u0026amp; Roman, F. (2022). A climatological overview of lightning activity in Colombia using lightning imaging sensor (LIS) data. \u003cem\u003eTheoretical and Applied Climatology\u003c/em\u003e, \u003cem\u003e149\u003c/em\u003e(1), 69\u0026ndash;82. https://doi.org/10.1007/s00704-022-04029-w\u003c/li\u003e\n\u003cli\u003eFernandes, W. A. (2006). \u003cem\u003eCaracter\u0026iacute;sticas dos rel\u0026acirc;mpagos gerados por nuvens de tempestades em ambientes sob a influ\u0026ecirc;ncia das queimadas em Rond\u0026ocirc;nia\u003c/em\u003e [Characteristics of lightning generated by storm clouds in environments under the influence of fires in Rond\u0026ocirc;nia] [Doctoral dissertation, Instituto Nacional de Pesquisas Espaciais].\u003c/li\u003e\n\u003cli\u003eFujisada, H., Bailey, G. B., Kelly, G. G., Hara, S., \u0026amp; Abrams, M. J. (2005). ASTER DEM performance. \u003cem\u003eIEEE Transactions on Geoscience and Remote Sensing\u003c/em\u003e, \u003cem\u003e43\u003c/em\u003e(12), 2707\u0026ndash;2714. https://doi.org/10.1109/TGRS.2005.847924\u003c/li\u003e\n\u003cli\u003eFunk, C., Peterson, P., Landsfeld, M., Pedreros, D., Verdin, J., Shukla, S., Husak, G., Rowland, J., Harrison, L., Hoell, A., \u0026amp; Michaelsen, J. (2014). \u003cem\u003eA quasi-global precipitation time series for drought monitoring\u003c/em\u003e (U.S. Geological Survey Data Series 832). https://doi.org/10.3133/ds832\u003c/li\u003e\n\u003cli\u003eG1. (2013, March 4). \u003cem\u003eRaios causam preju\u0026iacute;zo de R$ 1 bilh\u0026atilde;o \u0026agrave; economia do pa\u0026iacute;s, diz Inpe\u003c/em\u003e [Lightning causes R$ 1 billion loss to the country\u0026apos;s economy, says Inpe]. G1. http://g1.globo.com/brasil/noticia/2013/03/raios-causam-prejuizo-de-r-1-bilhao-economia-do-pais-diz-inpe.html\u003c/li\u003e\n\u003cli\u003eJayaratne, E. R., \u0026amp; Kuleshov, Y. (2006). Geographical and seasonal characteristics of the relationship between lightning ground flash density and rainfall within the continent of Australia. \u003cem\u003eAtmospheric Research\u003c/em\u003e, \u003cem\u003e79\u003c/em\u003e(1), 1\u0026ndash;14. https://doi.org/10.1016/j.atmosres.2005.03.004\u003c/li\u003e\n\u003cli\u003eKodinar\u0026iacute;ya, T. M., \u0026amp; Makwana, P. R. (2013). Review on determining number of Cluster in K-Means Clustering. \u003cem\u003eInternational Journal of Advance Research in Computer Science and Management Studies\u003c/em\u003e, \u003cem\u003e1\u003c/em\u003e(6), 90\u0026ndash;95.\u003c/li\u003e\n\u003cli\u003eKotroni, V., \u0026amp; Lagouvardos, K. (2008). Lightning occurrence in relation with elevation, terrain slope, and vegetation cover in the Mediterranean. \u003cem\u003eJournal of Geophysical Research: Atmospheres\u003c/em\u003e, \u003cem\u003e113\u003c/em\u003e(D21), D21113. https://doi.org/10.1029/2008JD010605\u003c/li\u003e\n\u003cli\u003eMacQueen, J. (1967). Some methods for classification and analysis of multivariate observations. In L. M. Le Cam \u0026amp; J. Neyman (Eds.), \u003cem\u003eProceedings of the fifth Berkeley symposium on mathematical statistics and probability, Volume 1: Theory of statistics\u003c/em\u003e (pp. 281\u0026ndash;297). University of California Press.\u003c/li\u003e\n\u003cli\u003eMapBiomas. (2024). \u003cem\u003eAlgorithm Theoretical Basis Document (ATBD) - Collection 9\u003c/em\u003e (Version 2).\u003c/li\u003e\n\u003cli\u003eMarengo, J. A., Tomasella, J., Alves, L. M., Soares, W. R., \u0026amp; Rodriguez, D. A. (2012). Extreme climatic events in the Amazon basin. \u003cem\u003eTheoretical and Applied Climatology\u003c/em\u003e, \u003cem\u003e107\u003c/em\u003e(1\u0026ndash;2), 73\u0026ndash;85. https://doi.org/10.1007/s00704-011-0465-1\u003c/li\u003e\n\u003cli\u003eMarkowski, P., \u0026amp; Richardson, Y. (2010). \u003cem\u003eMesoscale meteorology in midlatitudes\u003c/em\u003e. John Wiley \u0026amp; Sons.\u003c/li\u003e\n\u003cli\u003eMattos, E. V., Machado, L. A. T., \u0026amp; Gon\u0026ccedil;alves, W. A. (2011). Cloud-to-ground lightning and Mesoscale Convective Systems. \u003cem\u003eAtmospheric Research\u003c/em\u003e, \u003cem\u003e99\u003c/em\u003e(3\u0026ndash;4), 377\u0026ndash;390. https://doi.org/10.1016/j.atmosres.2010.11.007\u003c/li\u003e\n\u003cli\u003eMattos, E. V., da Silva, J. B., Machado, L. A. T., \u0026amp; Albrecht, R. I. (2024). Rel\u0026acirc;mpagos: g\u0026ecirc;nese, tipos e impactos [Lightning: Genesis, types and impacts]. \u003cem\u003eRevista Brasileira de Geografia F\u0026iacute;sica\u003c/em\u003e, \u003cem\u003e17\u003c/em\u003e(5), 3876\u0026ndash;3905. https://doi.org/10.26848/rbgf.v17.5.p3876-3905\u003c/li\u003e\n\u003cli\u003eMondal, S., Das, S., \u0026amp; Maitra, A. (2022). Spatio-temporal variability of lightning climatology and its association with thunderstorm indices over India. \u003cem\u003eTheoretical and Applied Climatology\u003c/em\u003e, \u003cem\u003e149\u003c/em\u003e(3-4), 1195\u0026ndash;1212. https://doi.org/10.1007/s00704-022-04104-2\u003c/li\u003e\n\u003cli\u003eMutti, P. R., de Medeiros, G. A., \u0026amp; de Assis de Souza Filho, F. (2022). Long-term meteorological drought characterization in the S\u0026atilde;o Francisco watershed, Brazil: A climatic water balance approach. \u003cem\u003eInternational Journal of Climatology\u003c/em\u003e, \u003cem\u003e42\u003c/em\u003e(13), 6757\u0026ndash;6775. https://doi.org/10.1002/joc.7701\u003c/li\u003e\n\u003cli\u003eNesbitt, S. W., Zipser, E. J., \u0026amp; Cecil, D. J. (2000). A Census of Precipitation Features in the Tropics Using TRMM: Radar, Ice Scattering, and Lightning Observations. \u003cem\u003eJournal of Climate\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(23), 4087\u0026ndash;4106. https://doi.org/10.1175/1520-0442(2000)013\u0026lt;4087:ACOPFI\u0026gt;2.0.CO;2\u003c/li\u003e\n\u003cli\u003eOda, F. S. S., Naccarato, K. P., Pinto, O., Jr., \u0026amp; Holle, R. L. (2022). An initial assessment of the distribution of total Flash Rate Density (FRD) in Brazil from GOES-16 Geostationary Lightning Mapper (GLM) observations. \u003cem\u003eAtmospheric Research\u003c/em\u003e, \u003cem\u003e270\u003c/em\u003e, 106081. https://doi.org/10.1016/j.atmosres.2022.106081\u003c/li\u003e\n\u003cli\u003ePe\u0026ntilde;a, J. M., Lozano, J. A., \u0026amp; Larra\u0026ntilde;aga, P. (1999). An empirical comparison of four initialization methods for the K-Means algorithm. \u003cem\u003ePattern Recognition Letters\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(10), 1027\u0026ndash;1040. https://doi.org/10.1016/S0167-8655(99)00069-0\u003c/li\u003e\n\u003cli\u003ePereira, R. G., Naccarato, K. P., \u0026amp; Pinto, O., Jr. (2025). Temporal Evolution of Lightning Properties in the Metropolitan Area of S\u0026atilde;o Paulo (MASP) During the CHUVA-Vale Campaign. \u003cem\u003eAtmosphere\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(4), 426. https://doi.org/10.3390/atmos16040426\u003c/li\u003e\n\u003cli\u003ePielke, R. A., Sr. (2001). Influence of the spatial distribution of vegetation and soils on the prediction of cumulus convective rainfall. \u003cem\u003eReviews of Geophysics\u003c/em\u003e, \u003cem\u003e39\u003c/em\u003e(2), 151\u0026ndash;177. https://doi.org/10.1029/1999RG000072\u003c/li\u003e\n\u003cli\u003ePinto Junior, O., \u0026amp; Pinto, I. R. C. A. (2008). Mortes por raios no Brasil: uma nova estimativa e considera\u0026ccedil;\u0026otilde;es sobre a sua distribui\u0026ccedil;\u0026atilde;o, sazonalidade e letalidade [Deaths by lightning in Brazil: A new estimate and considerations about its distribution, seasonality and lethality]. \u003cem\u003eRevista Brasileira de Geof\u0026iacute;sica\u003c/em\u003e, \u003cem\u003e26\u003c/em\u003e(1), 77\u0026ndash;87.\u003c/li\u003e\n\u003cli\u003ePotdar, S. S., Singh, D., \u0026amp; Singh, R. (2025). Influence of land use land cover and topography on lightning distribution over north and north-east Indian region. \u003cem\u003eJournal of Atmospheric and Solar-Terrestrial Physics\u003c/em\u003e, \u003cem\u003e268\u003c/em\u003e, 106460. https://doi.org/10.1016/j.jastp.2025.106460\u003c/li\u003e\n\u003cli\u003eQie, X., Toumi, R., \u0026amp; Yuan, T. (2003). Lightning activities on the Tibetan Plateau as observed by the lightning imaging sensor. \u003cem\u003eJournal of Geophysical Research: Atmospheres\u003c/em\u003e, \u003cem\u003e108\u003c/em\u003e(D17), 4553. https://doi.org/10.1029/2002JD003304\u003c/li\u003e\n\u003cli\u003eRasmussen, K. L., \u0026amp; Houze, R. A., Jr. (2011). Orogenic convection in subtropical South America as seen by the TRMM satellite. \u003cem\u003eMonthly Weather Review\u003c/em\u003e, \u003cem\u003e139\u003c/em\u003e(8), 2399\u0026ndash;2420. https://doi.org/10.1175/MWR-D-10-05006.1\u003c/li\u003e\n\u003cli\u003eReynolds, S. E., Brook, M., \u0026amp; Gourley, M. F. (1957). Thunderstorm Charge Separation. \u003cem\u003eJournal of Meteorology\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(5), 426\u0026ndash;436. https://doi.org/10.1175/1520-0469(1957)014\u0026lt;0426:TCS\u0026gt;2.0.CO;2\u003c/li\u003e\n\u003cli\u003eRoe, G. H. (2005). Orographic Precipitation. \u003cem\u003eAnnual Review of Earth and Planetary Sciences\u003c/em\u003e, \u003cem\u003e33\u003c/em\u003e, 645\u0026ndash;671. https://doi.org/10.1146/annurev.earth.33.092203.122541\u003c/li\u003e\n\u003cli\u003eRomatschke, U., \u0026amp; Houze, R. A., Jr. (2010). Extreme Summer Convection in South America. \u003cem\u003eJournal of Climate\u003c/em\u003e, \u003cem\u003e23\u003c/em\u003e(14), 3761\u0026ndash;3791. https://doi.org/10.1175/2010JCLI3465.1\u003c/li\u003e\n\u003cli\u003eRosenfeld, D., Lohmann, U., Rissman, G. B., Flagan, R. C., Seinfeld, J. H., Andreae, M. O., Asmi, E., \u0026amp; Jurkat, T. (2008). Flood or drought: how do aerosols affect precipitation? \u003cem\u003eScience\u003c/em\u003e, \u003cem\u003e321\u003c/em\u003e(5894), 1309\u0026ndash;1313. https://doi.org/10.1126/science.1160606\u003c/li\u003e\n\u003cli\u003edos Santos, A. P. P., da S. S. de Souza, E. B., \u0026amp; dos Santos, H. L. C. (2024). Lightning under Different Land Use and Cover, and the Influence of Topography in the Caraj\u0026aacute;s Mineral Province, Eastern Amazon. \u003cem\u003eAtmosphere\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(3), 375. https://doi.org/10.3390/atmos15030375\u003c/li\u003e\n\u003cli\u003eSaunders, C. (2008). Charge Separation Mechanisms in Clouds. \u003cem\u003eSpace Science Reviews\u003c/em\u003e, \u003cem\u003e137\u003c/em\u003e(1\u0026ndash;4), 335\u0026ndash;353. https://doi.org/10.1007/s11214-008-9345-0\u003c/li\u003e\n\u003cli\u003eSchneider, L., Eremenko, M., Cuesta, J., Dufour, G., \u0026amp; Foret, G. (2019). Relative impact of aerosol, soil moisture, and orography perturbations on deep convection. \u003cem\u003eAtmospheric Chemistry and Physics\u003c/em\u003e, \u003cem\u003e19\u003c/em\u003e(19), 12343\u0026ndash;12359. https://doi.org/10.5194/acp-19-12343-2019\u003c/li\u003e\n\u003cli\u003eSim\u0026otilde;es, F. T. A. (2023). \u003cem\u003eSensibilidade da vaz\u0026atilde;o m\u0026eacute;dia anual ao clima na bacia hidrogr\u0026aacute;fica do Rio S\u0026atilde;o Francisco\u003c/em\u003e [Sensitivity of the average annual flow to the climate in the S\u0026atilde;o Francisco river basin] [Undergraduate thesis, Universidade de Bras\u0026iacute;lia].\u003c/li\u003e\n\u003cli\u003eTaylor, C. M., Gounou, A., Guichard, F., Harris, P. P., Ellis, R. J., Couvreux, F., \u0026amp; De Kauwe, M. (2011). Frequency of Sahelian storm initiation enhanced over mesoscale soil-moisture patterns. \u003cem\u003eNature Geoscience\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(7), 430\u0026ndash;433. https://doi.org/10.1038/ngeo1173\u003c/li\u003e\n\u003cli\u003eUnfer, G. R., Dias, N. L., Chamecki, M., Martins, L. D., \u0026amp; Gon\u0026ccedil;alves, W. A. (2025). Decoding the Relationship Between Cloud Electrification, Downdrafts, and Surface Ozone in the Amazon Basin. \u003cem\u003eJournal of Geophysical Research: Atmospheres\u003c/em\u003e, \u003cem\u003e130\u003c/em\u003e(4), e2024JD042158. https://doi.org/10.1029/2024JD042158\u003c/li\u003e\n\u003cli\u003eVogt, B. J., \u0026amp; Hodanish, S. J. (2016). A geographical analysis of warm season lightning/landscape interactions across Colorado, USA. \u003cem\u003eApplied Geography\u003c/em\u003e, \u003cem\u003e75\u003c/em\u003e, 93\u0026ndash;103. https://doi.org/10.1016/j.apgeog.2016.08.006\u003c/li\u003e\n\u003cli\u003eWang, Y., Fan, J., Zhang, R., Zhang, D., \u0026amp; Li, P. (2018). Aerosol microphysical effects on lightning activity under different cloud-dominated regimes. \u003cem\u003eAtmospheric Chemistry and Physics\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e(17), 12797\u0026ndash;12809. https://doi.org/10.5194/acp-18-12797-2018\u003c/li\u003e\n\u003cli\u003eWhiteman, C. D. (2000). \u003cem\u003eMountain meteorology: Fundamentals and applications\u003c/em\u003e. Oxford University Press.\u003c/li\u003e\n\u003cli\u003eWilliams, E. R. (1989). The tripole structure of thunderstorms. \u003cem\u003eJournal of Geophysical Research: Atmospheres\u003c/em\u003e, \u003cem\u003e94\u003c/em\u003e(D11), 13151\u0026ndash;13167. https://doi.org/10.1029/JD094iD11p13151\u003c/li\u003e\n\u003cli\u003eWu, J. (2012). Cluster analysis and K-means clustering: An introduction. In J. He, X. Li, \u0026amp; X. Li (Eds.), \u003cem\u003eAdvances in K-means clustering\u003c/em\u003e (pp. 1\u0026ndash;16). Springer.\u003c/li\u003e\n\u003cli\u003eZhang, D., Koshak, W., \u0026amp; Boccippio, D. (2019). Evaluation of the Performance Characteristics of the Lightning Imaging Sensor. \u003cem\u003eJournal of Atmospheric and Oceanic Technology\u003c/em\u003e, \u003cem\u003e36\u003c/em\u003e(6), 1015\u0026ndash;1031. https://doi.org/10.1175/JTECH-D-18-0173.1\u003c/li\u003e\n\u003cli\u003eZipser, E. J., Cecil, D. J., Liu, C., Nesbitt, S. W., \u0026amp; Yorty, D. P. (2006). Where are the most intense thunderstorms on earth? \u003cem\u003eBulletin of the American Meteorological Society\u003c/em\u003e, \u003cem\u003e87\u003c/em\u003e(8), 1057\u0026ndash;1071. https://doi.org/10.1175/BAMS-87-8-1057\u003c/li\u003e\n\u003cli\u003eZoboli, V. G., \u0026amp; da Silva, T. A. (2023). Estudo da incid\u0026ecirc;ncia de nuvens CBS no centro-oeste brasileiro: Uma avalia\u0026ccedil;\u0026atilde;o decenal (2013-2023) a partir das cartas SIGWX PROG [Study of the incidence of CB clouds in the Brazilian midwest: A decennial evaluation (2013-2023) from SIGWX PROG charts]. \u003cem\u003eRevista Brasileira de Avia\u0026ccedil;\u0026atilde;o Civil \u0026amp; Ci\u0026ecirc;ncias Aeron\u0026aacute;uticas\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(5), 178\u0026ndash;201. https://rbac.cia.emnuvens.com.br/revista/article/view/197\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"pure-and-applied-geophysics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"paag","sideBox":"Learn more about [Pure and Applied Geophysics](https://www.springer.com/journal/24)","snPcode":"24","submissionUrl":"https://submission.nature.com/new-submission/24/3","title":"Pure and Applied Geophysics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Cluster, Hotspots, Slope, Remote Sensing","lastPublishedDoi":"10.21203/rs.3.rs-7566962/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7566962/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWith one of the highest rates of lightning activity globally, Brazil faces a significant natural hazard. The S\u0026atilde;o Francisco River Basin (BHSF) represents a key area for studying this phenomenon. This study analyzes the spatiotemporal distribution of lightning activity within the basin, identifies areas of maximum concentration (hotspots), and investigates its relationship with surface and atmospheric drivers. We used lightning data from the Lightning Imaging Sensor (LIS; 1998\u0026ndash;2013), precipitation (CHIRPS), topography (ASTER), and land use (MapBiomas) data, applying the K-means clustering technique for pattern segmentation. Results indicate that hotspots, with flash rates up to 39.9 flashes km⁻\u0026sup2; yr⁻\u0026sup1;, are concentrated in the western portion of the basin, predominantly over plateau areas characterized by agricultural use. The temporal analysis revealed a distinct seasonal cycle, with maximum activity in summer coupled with the rainfall regime, and a diurnal peak in the late afternoon. We conclude that the interaction between topography-induced air uplift and surface alterations from agricultural land use are the primary modulators of the storm regime in the BHSF, offering valuable insights for risk mitigation strategies.\u003c/p\u003e","manuscriptTitle":"Lightning Behavior and Its Relationship with Topography, Precipitation, and Land Use in the São Francisco River Basin","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-11 11:20:15","doi":"10.21203/rs.3.rs-7566962/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-11-10T13:29:29+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-10T10:28:28+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-28T15:56:02+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-28T13:52:35+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"306834833476801609181912321857525633226","date":"2025-10-12T11:20:06+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"51062356658851756148248669605866712749","date":"2025-10-10T19:50:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"193246451914240684969310579464115481661","date":"2025-09-25T06:51:47+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-22T06:09:30+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-11T06:57:20+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-10T07:23:12+00:00","index":"","fulltext":""},{"type":"submitted","content":"Pure and Applied Geophysics","date":"2025-09-08T18:28:53+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"pure-and-applied-geophysics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"paag","sideBox":"Learn more about [Pure and Applied Geophysics](https://www.springer.com/journal/24)","snPcode":"24","submissionUrl":"https://submission.nature.com/new-submission/24/3","title":"Pure and Applied Geophysics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"4f3b3df8-7667-42cb-a204-ca61a79ec7a2","owner":[],"postedDate":"September 11th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-28T09:40:10+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-11 11:20:15","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7566962","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7566962","identity":"rs-7566962","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}