ENSO-driven climate variability and crop yield responses in contrasting Brazilian regions: insights from functional data analysis and machine learning

ENSO-driven climate variability and crop yield responses in contrasting Brazilian regions: insights from functional data analysis and machine learning | 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 ENSO-driven climate variability and crop yield responses in contrasting Brazilian regions: insights from functional data analysis and machine learning Ludmilla Ferreira Justino, David Henriques Matta, Marcos Vinício Cesario dos Santos, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9314260/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 10 You are reading this latest preprint version Abstract The El Niño-Southern Oscillation (ENSO) is a major driver of climate and agricultural productivity variability, as those events modulate significant fraction of rainfall and air temperature interannual variation, key factors directly affecting crop performance. This study aimed to evaluate the impacts of ENSO on rice, common bean, and soybean yields in two periods (1961–1990 and 1991–2019) using a Random Forest modeling approach. The analysis focused on the states of Rio Grande do Sul (RS) and Goiás (GO), Brazil, which represent contrasting climatic regions within the country. Daily data on rainfall and maximum and minimum air temperature (1961–2019) were obtained from 69 weather stations. ENSO phases were classified using the Oceanic Niño Index (ONI), in the Niño 3.4 region, and analyzed through Functional Data Analysis (FDA) and Functional Analysis of Variance (FANOVA) across two periods (1961–1990 and 1991–2019). Observed crop yield data were incorporated into Random Forest models to estimate the relative importance of climatic variables (air temperature and rainfall) and spatial factors (municipality) in determining agricultural productivity. The results demonstrate a clear influence of ENSO on climate variability and crop yields, particularly in Rio Grande do Sul, whereas its effects were less pronounced in Goiás. The El Niño phase generally favored yield increases across all evaluated crops (irrigated rice, common bean, and soybean), while the La Niña phase was frequently associated with yield reductions. During Neutral years, the influence of ENSO was comparatively weaker. A reduction in the relative importance of rainfall was observed in the more recent period (1991–2019), indicating increased rainfall homogeneity over time. In contrast, local factors became more influential in determining crop yield. Nevertheless, ENSO remains a critical factor for explaining yield variability and for supporting agricultural management and decision-making strategies. Random Forest Climate variability Predictive modeling Functional data analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 1 INTRODUCTION Agricultural production constitutes a fundamental pillar of the Brazilian economy, contributing substantially to the trade balance, Gross Domestic Product (GDP), and employment generation. In this context, the state of Rio Grande do Sul stands out as the fourth-largest grain producer in the country, with a production of 33.2 million tons in the 2024/2025 growing season, corresponding to 9.8% of national output. Although soybean ( Glycine max L.; 43%) and irrigated rice ( Oryza sativa L.; 25%) account for the largest share of this production, crops with lower relative participation, such as common bean ( Phaseolus vulgaris L.), representing only 0.2% of state production, hold considerable socioeconomic importance, particularly within family farming systems (Conab, 2025; Feix et al., 2022 ). Furthermore, Rio Grande do Sul is the leading producer of irrigated rice in Brazil, with 8.3 million tons (74.3% of national production), and ranks fourth in soybean production, with 14.3 million tons (8.4% of the national total). Common bean production, predominantly of the “black” commercial type, reaches 56.2 thousand tons, equivalent to 7.1% of Brazilian production (Conab, 2025). Similarly, the state of Goiás plays a strategic role in national agricultural production, ranking as the third-largest grain producer, with 35 million tons in the 2024/2025 season, representing approximately 10% of total Brazilian production. In this state, soybean is the dominant crop (58%), followed by maize (36%), sorghum (4%), and common bean (0.9%) (Conab, 2025). Despite their agricultural relevance, these two states exhibit markedly contrasting climatic conditions. Goiás is located in Central-East South America (CESA), at the core of the South American Monsoon System (Vera et al., 2006 ; Marengo et al., 2012 ), which is characterized by a well-defined rainy season during the austral summer (December–February, DJF) and a dry season during winter (June–August, JJA). The region is also influenced by the South Atlantic Convergence Zone (SACZ) (Kodama, 1992 , 1993 ; Van Der Wiel et al., 2015 ), whose activity significantly intensifies rainfall volumes. In addition, due to its location within a transition zone between northern and southern Brazil, Goiás exhibits sensitivity to teleconnections associated with the El Niño–Southern Oscillation (ENSO) phenomenon (Grimm, 2003 ; Penalba & Rivera, 2016 ; Moura et al., 2019 ; Nóia Júnior & Sentelhas, 2019 ; Matta et al., 2023 ). Conversely, Rio Grande do Sul is part of the Southeast South America (SESA) region, where ENSO-related effects are particularly pronounced (Crespo et al., 2024 ). El Niño events are typically associated with increased rainfall (Medeiros & Oliveira, 2021 ; Arruda et al., 2025 ), a higher frequency of rainy days (Fontana & Almeida, 2002 ), increased cloud cover (Custodio, 2016 ), and elevated air temperatures (Guimarães & Reis, 2012 ). In contrast, La Niña events are generally associated with reduced rainfall in southern Brazil and increased precipitation in the North and Northeast regions, often resulting in drought conditions in Rio Grande do Sul (Lopes et al., 2022 ; Scheibel et al., 2024 ). The impacts of ENSO on agricultural systems are widely documented and vary according to region, crop type, and the specific phase of the phenomenon (Iizumi et al., 2014 ; Cao et al., 2023 ). These effects are particularly critical for rainfed crops, such as soybean (Qian et al., 2020 ) and common bean (Cirino et al., 2015 ), which are highly sensitive to climatic variability and extreme events. Although irrigated rice is less vulnerable to rainfall fluctuations, variations in air temperature and solar radiation may negatively affect grain yield (Qian et al., 2020 ; Singh et al., 2023 ), with potential consequences for supply and market prices, thereby impacting food security (Liu et al., 2023 ). Recent advances have improved the understanding of ENSO mechanisms and their spatiotemporal patterns, as well as their impacts on agricultural systems (Qi et al., 2022 ; Liang et al., 2024 ; Hintz et al., 2025 ; Wu et al., 2025 ). However, important gaps remain in the integrated understanding of crop responses to climatic variability associated with ENSO, particularly in studies based on long-term historical datasets. In this context, advanced analytical approaches, including machine learning techniques, are essential for identifying the climatic variables that most strongly influence crop performance under different ENSO phases. Given this scenario, the present study aimed to evaluate the influence of ENSO on rainfall, air temperature, and the yield of irrigated rice, common bean, and soybean across two distinct periods (1961–1990 and 1991–2019) in the states of Goiás and Rio Grande do Sul, Brazil. 2 MATERIALS AND METHODS 2.1 Description of the Study Area The study area comprises two Brazilian states with contrasting climatic conditions: Rio Grande do Sul (RS) and Goiás (GO), located in the Southern and Central-West regions, respectively (Fig. 1 and Table S1 – Supplementary Material). Rio Grande do Sul is characterized by a humid subtropical climate, predominantly classified as Cfa (humid subtropical climate, 86.7%) and Cfb (temperate oceanic climate, 13.3%). These climate types are defined by the absence of a well-defined dry season and a relatively uniform distribution of rainfall throughout the year, with annual precipitation ranging from 1000 to 1600 mm in the southern portion and from 1600 to 2000 mm in the northern region (Matzenauer et al., 2011 ). The Cfa climate, which predominates across most of the state, occurs at altitudes between 0 and 650 m and is characterized by hot summers, with mean temperatures below 18°C in the coldest month and above 22°C in the warmest month. In contrast, the Cfb climate occurs at higher altitudes (approximately 900 m) and is characterized by milder summers, mean temperatures below 22°C in the warmest month, and frequent frost events. In Goiás, the predominant climate is classified as Aw (94%), corresponding to a tropical climate with a marked dry winter. The state exhibits annual precipitation between 1600 and 1900 mm, altitudes reaching up to 1200 m, and mean annual temperatures ranging from 19 to 20°C (Alvares et al., 2013 ). 2.2 Meteorological Dataset Meteorological stations from the National Institute of Meteorology (INMET) were selected based on the availability of more than 30 years of daily records of maximum and minimum air temperature and/or rainfall, with less than 30% missing or inconsistent data. For air temperature, records were considered inconsistent when Tmax 200 mm) or at least three identical consecutive measurements (excluding zero rainfall). Missing values in both rainfall and temperature series were gap-filled using the BR-DWGD database (Xavier et al., 2022 ). Based on these criteria, a total of 69 stations were selected across the study area: 62 located in Rio Grande do Sul and 7 in Goiás (Fig. 1 and Table S1 ). In Rio Grande do Sul, 16 stations provided temperature data and 46 provided rainfall data, whereas in Goiás, five stations provided temperature data and two provided rainfall data. Rainfall data were aggregated into weekly accumulated values, while maximum and minimum air temperatures were represented by the absolute weekly maximum and minimum values. This procedure was adopted to reduce the influence of daily variability and noise associated with non-rainy days. Subsequently, the dataset was filtered according to the agricultural calendar of each region. In Rio Grande do Sul, the analyzed period comprised 43 weeks (September 1st to June 28th), whereas in Goiás, it comprised 34 weeks (October 1st to May 26th), encompassing the complete crop development cycles of rice, common bean, and soybean. This selection reflects the sowing windows for irrigated rice, rainfed common bean, and rainfed soybean in Rio Grande do Sul, and for rainfed common bean and soybean in Goiás. 2.3 ENSO DATA ENSO phases are typically defined based on sea surface temperature (SST) anomalies and their persistence across the equatorial Pacific Ocean. In this study, monthly SST data from the Climate Prediction Center (CPC) of the National Oceanic and Atmospheric Administration (NOAA) were used for the period 1961–2019. The Oceanic Niño Index (ONI), provided by NOAA, is based on three-month running means of SST anomalies in the Niño 3.4 region and is widely used to identify El Niño and La Niña events. El Niño events are characterized by positive SST anomalies equal to or greater than 0.5°C above the climatological mean, whereas La Niña events correspond to negative anomalies equal to or less than − 0.5°C. Neutral conditions indicate the absence of significant anomalies (NOAA, 2019). For Rio Grande do Sul, the prevailing anomaly for each growing season was calculated as the mean of the quarterly ONI values, considering the SON, OND, NDJ, DJF, JFM, FMA, MAM, and AMJ periods. While for Goiás, the calculation was based on the OND, NDJ, DJF, JFM, FMA, and MAM quarters, thereby encompassing the complete crop development cycles in both regions. Based on these criteria, each growing season was classified according to the ENSO phase (El Niño, La Niña, or Neutral). Climate data (Section 2.2) and ONI values were analyzed across two temporal periods: (a) 1961–1990 and (b) 1991–2019. This temporal segmentation is particularly relevant in the context of ENSO, as it enables the assessment of potential changes in the frequency and cyclic behavior of El Niño and La Niña events across different climatic periods (Li et al., 2024 ). 2.4 Statistical Analysis To evaluate the influence of ENSO on climatic variables (rainfall and air temperature), Functional Data Analysis (FDA) was applied to both temporal periods (1961–1990 and 1991–2019) for the states of Goiás and Rio Grande do Sul. FDA enables the discrimination of ENSO impacts on rainfall and temperature, as well as the assessment of similarity patterns across ENSO phases (La Niña, El Niño, and Neutral) over time. In addition, this approach allows the analysis of groups of estimated functions, facilitating the evaluation of their statistical distribution and the estimation of mean functional curves for the climatic variables of interest. In recent years, FDA has been increasingly applied in meteorology (Beyaztas & Yaseen, 2019 ; Curceac et al., 2019 ; Matta et al., 2023 ), economics (Wang et al., 2021 ), geosciences (Bottazzi & Rossa, 2017 ; Pérez-Plaza et al., 2018 ), and agricultural sciences (Kwak et al., 2016 ; Shah et al., 2018 ; Xu et al., 2018a , 2018b , Justino et al. 2026 ), demonstrating its robustness for analyzing complex temporal patterns. All statistical analyses and graphical outputs were performed using R software (R Core Team, 2025 ). 2.4.1 Classification of Climatic Variables based on ENSO Phases Climatic data for rainfall, maximum temperature, and minimum temperature, corresponding to the states of Rio Grande do Sul and Goiás, the analyzed periods ((a) and (b)), and ENSO phases, were initially processed using the fdata function from the fda.usc R package (Febrero-Bande & Oviedo de la Fuente, 2012) and subsequently converted into functional objects. Following this step, the functional k-means clustering method was applied to each combination of state, period, and ENSO phase to estimate representative mean functional curves for each group. 2.4.2 Comparison of Mean Functional Curve Estimates After defining the representative mean functional curves for each ENSO phase, a Functional Analysis of Variance (FANOVA) was performed to assess whether the estimated functions (mean functional curves for El Niño, La Niña, and Neutral groups) differed statistically. The equality of functional means was tested using the fdANOVA package (Górecki & Smaga, 2017 ). The FP (F-type Permutation) test statistic, an extension of the classical ANOVA F-statistic for functional data, was adopted. This approach is particularly appropriate when variability is concentrated in a limited number of functional modes, thereby improving analytical efficiency while reducing model complexity. A significance level of 5% was assumed for all hypothesis tests. This analytical framework enabled the identification of climatic patterns associated with ENSO phases for absolute maximum temperature, absolute minimum temperature, and weekly accumulated rainfall. FANOVA was initially applied separately to each climatic variable within each period. When statistically significant differences were detected, post hoc multiple comparison tests were conducted between pairs of ENSO phases (El Niño vs. Neutral; El Niño vs. La Niña; Neutral vs. La Niña). To control for Type I error, the Holm–Bonferroni method was applied to sequentially adjust p-values (Holm, 1979 ). It is important to emphasize that both the pairwise comparisons among ENSO phases and the subsequent analysis of climatic variable importance for crop yield (Section 2.4.3) were performed only when FANOVA indicated statistically significant differences among functional means. 2.4.3 Yield Analysis Yield data for irrigated rice, common bean, and soybean in the selected municipalities (Table S1 ) were obtained from the Municipal Agricultural Production database (PAM, https://sidra.ibge.gov.br/tabela/1612 ) of the Brazilian Institute of Geography and Statistics (IBGE). These data represent the average yield for each crop (soybean, rice, and bean) during the growing seasons from September 1st to June 28th in Rio Grande do Sul and from October 1st to May 26th in Goiás. However, yield records for all analyzed crops and both states are only available from 1974 onwards (IBGE, 2020). Therefore, for yield analyses, temporal period (a) was adjusted from 1961–1990 to 1974–1990 to ensure data completeness and historical consistency. 2.4.4 Technological Trend in Crop Yield To remove the technological trend from the yield data, the methodology proposed by Heinemann & Sentelhas ( 2011 ) was applied. A loess regression was used to estimate the temporal trend associated with technological advancements. The estimated technological component was subsequently subtracted from the original yield values, isolating the variability attributable to climatic factors. This procedure was performed independently for each municipality (Table S1 ) and for both temporal periods: (a) 1974–1990 and (b) 1991–2019. An example of this approach applied to soybean yield in Bagé-RS during period (a) is presented in Figure S1 . Consequently, the yield data were adjusted for technological evolution and projected into a current baseline scenario, preserving the climatic variability that constitutes the focus of this study. 2.4.5 Crop Yield Prediction Following the detrending procedure, an unsupervised k -means clustering algorithm was applied to classify crop yields into two categories: high and low (Figure S2). The minimum and maximum thresholds for each group are presented in Table S2. This classification was subsequently used as the response variable in a supervised Random Forest classification model, implemented using the randomForest R package (Liaw & Wiener, 2002 ). Random Forest is an ensemble learning method based on multiple decision trees and is widely applied to both regression and classification problems. The predictor variables consisted of the climatic variables derived from the analyses described in Section 2.4.2. In cases where multiple comparison tests indicated no significant differences among ENSO phases for a given variable and state, these categories were aggregated into a single group. Additionally, municipalities were included as a predictor variable to incorporate spatial information and improve model performance. This approach enabled the evaluation of the relative importance of climatic variables and geographic location in determining yield classification, providing a quantitative measure of each variable’s contribution to yield patterns. For each crop, the Random Forest model was fitted using node splitting based on the Gini index (Sharda et al., 2019 ). Model accuracy was derived from the training dataset and evaluated separately for each crop and temporal period (Table S3). 2.4.6 Importance of Climatic Variables in Crop Yield Following model calibration, SHAP (SHapley Additive exPlanations) analysis was applied to interpret the Random Forest results and quantify the individual contribution of each predictor variable to yield classification. This approach is based on the concept of Shapley values (1953) from cooperative game theory and allows for the identification of both the magnitude and direction of variable effects (Siemers & Bajorath, 2023 ; Zheng et al., 2023 ). SHAP values represent the additive contribution of each explanatory variable, for each observation, relative to the baseline average prediction. Considering the binary response variable (1 for high yield and 0 for low yield), positive SHAP values indicate conditions that increase the probability of high yield, whereas negative values indicate factors associated with low-yield classification. 3 RESULTS AND DISCUSSION 3.1 Functional ANOVA The results of the Functional Analysis of Variance (FANOVA), incorporating the Holm–Bonferroni correction for pairwise comparisons among ENSO phases in the states of Goiás and Rio Grande do Sul, for weekly accumulated rainfall, absolute maximum temperature, and absolute minimum air temperature, are presented in Table 1 . Table 1 Functional Analysis of Variance (FANOVA) for the comparison of ENSO phases (El Niño, La Niña, and Neutral) in Goiás and Rio Grande do Sul for climatic variables (accumulated rainfall, maximum air temperature, and minimum air temperature) across the periods: (a) 1961–1990 and (b) 1991–2019. Significance level: p = 0.05. Hypotheses: H₀: El Niño = Neutral = La Niña; H₁: El Niño ≠ Neutral ≠ La Niña. Variable Period P-value (Holm-Bonferroni correction) Rio Grande do Sul Goiás Maximum Temperature (a) 1961–1990 0.003 0.683 (b) 1991–2019 0.008 0.273 Minimum Temperature (a) 1961–1990 0.012 0.372 (b) 1991–2019 0.000 0.642 Accumulated rainfall (a) 1961–1990 0.000 0.070 (b) 1991–2019 0.000 1.000 For Rio Grande do Sul, the analysis revealed statistically significant differences ( p value < 0.05) among ENSO phases (El Niño, Neutral, and La Niña) for all evaluated climatic variables across both temporal periods ((a) 1961–1990 and (b) 1991–2019). Moreover, the magnitude of statistical significance increased in the more recent period (b), particularly for minimum temperature, indicating that differences among ENSO phases have become more evident in recent decades (Table 1 ). This pattern suggests the presence of an interdecadal oscillation in ENSO amplitude, which modulates the intensity of its climatic teleconnections (Li et al., 2011 ). In contrast, for accumulated rainfall, the influence of ENSO remained consistent across both analyzed periods. In Goiás, however, FANOVA results indicated no statistically significant differences ( p value < 0.05) among ENSO phases for any of the evaluated climatic variables in either temporal period at the 5% significance level (Table 1 ). These findings suggest that climate variability in Goiás is less sensitive to ENSO-related oscillations when compared to Rio Grande do Sul. This result contrasts with Matta et al. ( 2023 ), who, using a shorter climatic dataset period (1990–2011), reported significant differences among ENSO phases in Goiás; however, that study also highlighted a reduction in statistical significance between periods 1990–2000 and 2001–2011. Given that the global analysis for Rio Grande do Sul rejected the null hypothesis of equality among ENSO phases, a pairwise comparison (El Niño vs. Neutral; El Niño vs. La Niña; Neutral vs. La Niña) was subsequently conducted. The results of these comparisons are presented in Table 2 . Table 2 Functional Analysis of Variance (FANOVA) for pairwise comparisons of ENSO phases (El Niño, La Niña, and Neutral) in the state of Rio Grande do Sul for climatic variables (accumulated rainfall, maximum air temperature, and minimum air temperature) across the periods: (a) 1961–1990 and (b) 1991–2019. Significance level: p = 0.05. Hypotheses: (1) El Niño = Neutral; (2) El Niño = La Niña; (3) La Niña = Neutral. Variable Period 1 2 3 H 0 : El Niño = Neutral H 1 : El Niño ≠ Neutral H 0 : El Niño = La Niña H 1 : El Niño ≠ La Niña H 0 : La Niña = Neutral H 1 : La Niña ≠ Neutral p-value Maximum Air Temperature (a) 1961–1990 0.025 0.081 0.003 (b) 1991–2019 0.002 0.260 0.006 Minimum Air Temperature (a) 1961–1990 0.133 0.001 0.009 (b) 1991–2019 0.002 0.000 0.000 Accumulated Rainfall (a) 1961–1990 0.000 0.000 0.004 (b) 1991–2019 0.000 0.000 0.000 For maximum air temperature, statistically significant differences were identified only when comparing the Neutral phase with El Niño and with La Niña in both temporal periods. No significant differences were observed between El Niño and La Niña, indicating that these two phases exert similar effects on maximum temperature in the study region. This finding is consistent with previous studies indicating that the Neutral phase is characterized by atmospheric and oceanic conditions close to the climatological mean, with less pronounced anomalies compared to El Niño and La Niña events (Lin & Qian, 2019 ). Consequently, during Neutral conditions, the effects on climatic variables such as maximum temperature tend to be less pronounced (Feng & Hao, 2021 ). For minimum air temperature, the results revealed a more complex and period-dependent behavior. During the 1961–1990 period (period a), statistically significant differences were primarily associated with the La Niña phase, which differed from both El Niño and Neutral conditions. In contrast, in the more recent period (b) (1991–2019), all pairwise comparisons were statistically significant, indicating that each ENSO phase exerts distinct influences on minimum temperature. Although the influence of ENSO on climatic variables is generally more evident at continental scales than at the regional scale considered in this study, a consistent pattern was observed for accumulated rainfall across all periods, with statistically significant differences among all ENSO phases. This result confirms that ENSO acts as a major modulator of the regional rainfall regime, although it is not the sole controlling factor (Souza et al., 2021 ). In this context, El Niño events are typically associated with increased rainfall in Rio Grande do Sul, whereas La Niña events are frequently linked to rainfall deficits (Arruda et al., 2025 ), with Neutral conditions representing an intermediate state (Lenka et al., 2022 ). Tamaddun et al. ( 2019 ) evaluated the influence of ENSO phases on changes in temperature, precipitation, and potential evapotranspiration in India and reported that El Niño years exert a stronger influence on these variables compared to La Niña and Neutral years. In contrast, Sattar et al. ( 2021 ) observed higher maximum temperatures and reduced rainfall during both El Niño and La Niña events relative to Neutral conditions in Bangladesh. These findings highlight that ENSO impacts are region-specific and depend on both the climatic variable considered and the phase of the phenomenon, being more pronounced in certain regions and under specific conditions. 3.2 Climatic Patterns across ENSO Phases Figure 2 presents the mean functional estimates for accumulated rainfall across the three ENSO phases for the states of Rio Grande do Sul and Goiás, including confidence intervals for the mean curves over weeks 1 to 34 (Goiás) and 1 to 43 (Rio Grande do Sul), for both analyzed periods. In Goiás, during the 1961–1990 period (Fig. 2 A), the accumulation of rainfall exhibited a relatively constant increment up to approximately week 28 (April), reaching an accumulated volume of approximately 1400 mm. This interval was characterized by smooth and nearly linear growth, indicating high regularity in rainfall distribution and reduced interannual variability. From April onwards, a deceleration in accumulated rainfall and an increase in variability were observed. Moreover, the highest rainfall volumes occurred during El Niño years, whereas Neutral conditions remained close to the overall mean of the three phases, and La Niña years exhibited lower accumulated rainfall. In the more recent period (1991–2019; Fig. 2 B), a similar stabilization of the mean rainfall curve after week 28 was observed in Goiás. However, this pattern was accompanied by a reduction of more than 100 mm in total accumulated rainfall, along with increased interannual variability in the final weeks. These changes in rainfall dynamics may be associated with climate change processes, which tend to intensify climate extremes and increase rainfall irregularity (Furtak & Wolińska, 2023 ). In contrast to period (a), during 1991–2019, the highest rainfall volumes were recorded during La Niña years, followed by Neutral conditions, whereas El Niño years presented the lowest accumulated rainfall. It is important to note that changes in ENSO phase frequency contributed to this pattern: period (a) was dominated by El Niño events (five occurrences) relative to La Niña (four), whereas period (b) showed a higher frequency of La Niña events (six) compared to El Niño (four), promoting increased rainfall in some years and drought conditions in others, thereby increasing interannual variability (Cai et al., 2021 ). In Rio Grande do Sul, accumulated rainfall during El Niño years in the 1961–1990 period (Fig. 2 C) was markedly higher than that observed during La Niña and Neutral years. By week 43, El Niño years reached approximately 1700 mm, while Neutral and La Niña years accumulated around 1300 mm and 1350 mm, respectively. The divergence among the curves begins around week 7 (October), coinciding with the onset of higher rainfall volumes in the state (INMET, 2020). Under El Niño conditions, Rio Grande do Sul typically experiences above-average precipitation, whereas La Niña years are associated with rainfall deficits and frequent droughts (Arruda et al., 2025 ). A similar pattern was reported by Teegavarapu & Sharma ( 2021 ), who observed higher precipitation during El Niño years, followed by Neutral and La Niña conditions, in Florida, USA. In contrast, during the more recent period (1991–2019; Fig. 2 D), the differences among ENSO phases were less pronounced, although still statistically significant (Table 2 ). Rainfall patterns during El Niño and La Niña years closely followed those observed under Neutral conditions, suggesting a potential reduction in the intensity or variability of ENSO effects on regional climate over time (Cai et al., 2020 ; Matta et al., 2023 ). According to Li et al. ( 2024 ), ENSO intensity has fluctuated in recent decades, increasing after the mid-1970s but declining during the 21st century. Similarly, An et al. ( 2023 ) argue that the historically strong relationship between ENSO and rainfall has weakened in recent decades. Figure 3 presents the ratio between the mean functional estimates of accumulated rainfall for different pairs of ENSO phases in Rio Grande do Sul. This analysis provides insights into how ENSO phases modulate accumulated rainfall over time, quantifying the relative increase or decrease in precipitation between phases. For the 1961–1990 period (Fig. 3 A), a pronounced positive peak is observed in the El Niño/La Niña ratio, indicating a strong predominance of rainfall during El Niño in the initial weeks. In contrast, the La Niña/Neutral ratio exhibits a negative peak, reflecting reduced rainfall during La Niña in the same interval. The El Niño/Neutral ratio remained relatively stable, without pronounced oscillations, but consistently ranged between 1.3 and 1.5, indicating higher rainfall during El Niño throughout most of the period. From week 15 onwards, all curves showed minimal variation, with La Niña and Neutral phases converging to similar values (ratio close to 1), while El Niño maintained higher rainfall levels. These results corroborate the patterns observed in Fig. 2 C, which also indicate higher precipitation during El Niño and similar behavior between La Niña and Neutral phases. In period (b) (1991–2019; Fig. 3 B), the ratios among ENSO phases remained close to 1 throughout the weeks, without the pronounced peaks observed in the earlier period. This indicates the absence of a clear dominance of any ENSO phase, reinforcing the pattern shown in Fig. 2 D, where rainfall differences among phases were minimal. Figure 4 A and 4 B present the mean functional estimates for maximum air temperature across ENSO phases over weeks 1 to 34 in Goiás for the two analyzed periods ((a) 1961–1990 and (b) 1991–2019). In both periods, maximum temperature exhibited a gradual decrease over time, with higher values at the beginning of the cycle (October) and lower values toward the end of the period (May). During period (a) (Fig. 4 A), maximum temperatures in Goiás started at approximately 34°C and gradually declined to around 30.6°C by the final week. In period (b) (Fig. 4 B), slightly higher initial values were observed, close to 36°C, with final values around 31.5°C. This increase in temperature suggests a warming trend over recent decades, potentially associated with global climate change (Regoto et al., 2021 ; Hofmann et al., 2021 ). Additionally, greater variability was observed between El Niño and La Niña curves, particularly up to approximately week 24 (March). A similar decreasing pattern was observed for minimum air temperature in the final weeks, especially after April (week 28) (Fig. 5 A and 5 B). This behavior is associated with the seasonal transition from autumn to winter, characterized by lower temperatures, shorter photoperiods, and reduced solar radiation (Justino et al., 2025 ). In Rio Grande do Sul, maximum air temperature (Figs. 4 C and 4 D) exhibited marked seasonal variation, with peak values occurring between mid-January and February (weeks 18 and 23), corresponding to the austral summer, regardless of ENSO phase. This period is also associated with increased solar radiation and lower relative humidity (INMET, 2020). During period (a) (1961–1990; Fig. 4 C), maximum temperatures in La Niña years were generally lower than those observed during Neutral conditions. A similar pattern was observed during El Niño years, although with greater variability, with temperatures fluctuating both above and below those recorded under Neutral conditions. In the more recent period (1991–2019; Fig. 4 D), maximum temperatures during El Niño years were frequently higher than those observed under Neutral conditions, particularly after week 25 (February). La Niña years exhibited greater variability but generally remained below Neutral values. Silva et al. ( 2020 ) reported increased air temperatures in Brazil during El Niño years, whereas Neutral and La Niña periods presented more similar conditions. Similarly, Sattar et al. ( 2021 ) observed higher maximum temperatures during both El Niño and La Niña years compared to Neutral conditions in Bangladesh, with more pronounced differences during El Niño events. The ratio analysis of maximum air temperature between ENSO phases in Rio Grande do Sul (Fig. 6 A) indicates that, for most weeks, values remained close to 1, suggesting limited differences among phases. However, greater variability was observed in the initial and final weeks, highlighting the predominance of El Niño over La Niña (ratio > 1) and lower values for La Niña relative to Neutral conditions (ratio < 1). Similarly, in period (b) (1991–2019; Fig. 6 B), the ratio curves remained clustered around 1, indicating subtle differences among ENSO phases. The El Niño/La Niña and El Niño/Neutral ratios were generally slightly above 1, indicating marginally higher maximum temperatures during El Niño. Conversely, the La Niña/Neutral ratio was below 1, indicating lower maximum temperatures during La Niña. These results are consistent with the patterns observed in Figs. 4 C and 4 D. Overall, minimum air temperature curves in Rio Grande do Sul (Figs. 5 C and 5 D) followed a seasonal pattern similar to that observed for maximum temperature, with an increase until approximately week 25 (summer), followed by a decline until week 43 (early winter). During period (a) (1961–1990; Fig. 5 C), minimum temperatures were generally lower during La Niña years, except between weeks 18 and 24. In period (b) (1991–2019; Fig. 5 D), El Niño years exhibited higher minimum temperatures during most weeks, exceeding those observed under Neutral conditions, particularly between weeks 10–19 and after week 25. The ratio analysis for minimum air temperature (Figs. 6 C and 6 D) revealed greater variability compared to maximum temperature. The El Niño/La Niña ratio showed pronounced positive peaks in both periods, indicating higher minimum temperatures during El Niño events. Conversely, the La Niña/Neutral ratio exhibited negative peaks in certain weeks, indicating lower minimum temperatures during La Niña relative to Neutral conditions. These results suggest that differences in minimum temperature among ENSO phases are more irregular and tend to intensify in the final weeks of the cycle, indicating a less stable ENSO influence compared to that observed for maximum temperature. 3.3 CLIMATE VARIABLE AND YIELD Figure 7 presents the mean absolute predictive importance of climatic variables and municipality for the yields of lowland rice, common bean, and soybean in the state of Rio Grande do Sul during periods (a) 1974–1990 and (b) 1991–2019. This analysis was not performed for the state of Goiás because, as described in Section 3.1 (Functional ANOVA), the FANOVA results for that state indicated no significant differences ( p value < 0.05) among ENSO phases (El Niño, La Niña, and Neutral) for any of the evaluated climatic variables in either analyzed period. A higher mean absolute SHAP value indicates a greater influence of a given variable on model output. For lowland rice during period (a) (1974–1990; Fig. 7 A), municipality showed the greatest predictive importance (80%) for yield, followed by accumulated rainfall (10.9%), whereas minimum (5.4%) and maximum (3.6%) air temperatures exhibited comparatively lower influence. When the specific effect of each variable on rice yield is examined (Fig. 8 A), accumulated rainfall during El Niño years (red) contributed mainly to yield increases (positive values), whereas La Niña (blue) and Neutral (gray) conditions were predominantly associated with yield reductions. For minimum temperature, the La Niña phase (blue) exerted a negative effect, while El Niño and Neutral phases (yellow), which did not differ significantly for this variable (Table 2 ) and were therefore evaluated jointly, tended to increase yield or produce no relevant effect (values close to zero). For maximum temperature, the Neutral phase showed a smaller effect on yield (gray), whereas El Niño and La Niña (purple), also evaluated jointly because of the absence of statistical differences, were associated with greater yield variation. In period (b) (1991–2019; Fig. 7 B), a substantial shift in the relative importance of the explanatory variables for rice yield was observed. Municipality remained the dominant predictor (80.3%), followed by minimum temperature (7.1%) and maximum temperature (6.8%). In contrast, rainfall, which ranked as the second most important variable in the earlier period, exhibited the lowest relevance in yield prediction (5.8%). During this period, El Niño (red) and La Niña (blue), whether evaluated individually or jointly (purple), tended to exert positive effects on rice yield regardless of the climatic variable analyzed. Conversely, the Neutral phase (gray) was associated with yield reductions (Fig. 8 B). Although the data used in this study refer to lowland rice, which in principle is less directly affected by rainfall variability, the substantial predictive relevance of rainfall, particularly in the first period, may be explained by its indirect effects on the production system and on other environmental variables, such as air and soil temperature, relative humidity, and solar radiation, all of which influence crop development and the energy balance. For example, high humidity and prolonged leaf wetness may favor the incidence of fungal diseases such as rice blast ( Magnaporthe oryzae ) and brown spot ( Bipolaris oryzae ), thereby compromising grain filling and final yield (Kirtphaiboon et al., 2021 ; Percich et al., 1997 ; Bhandari et al., 2024 ). In addition, rainfall occurring at undesirable times, such as during physiological maturity and harvest, when irrigation should be suspended, may compromise both yield and grain quality (Sohn et al., 2021 ). The high explanatory capacity of municipality in both periods is associated with edaphic, topographic, and management-related factors that vary spatially among production regions. Soil texture, fertility, water-holding capacity, and salinity directly affect nutrient availability and crop development (Ye et al., 2024 ; Omar et al., 2024 ). Topography is another critical factor, as flatter terrain facilitates irrigation management and promotes greater homogeneity in microclimatic conditions (Fan et al., 2020 ). Furthermore, heterogeneity in technological level among rice-producing municipalities affects phytosanitary management, water and fertilizer use efficiency, and the quality of field operations, resulting in substantial differences in yield (Fan & Gan, 2025 ). The greater predictive importance of maximum and minimum air temperatures in the more recent period (Fig. 7 B) indicates the growing impact of extreme thermal conditions associated with climate change, particularly during critical stages of rice development. In general, high temperatures reduce photosynthetic efficiency and increase respiration and transpiration, thereby compromising vegetative growth and nutrient uptake (Xiong et al., 2017 ; Lu et al., 2025 ). During flowering, elevated temperatures reduce the number of spikelets per panicle and increase pollen sterility, thereby decreasing the formation of viable seeds (Yu et al., 2024 ). During grain filling, excessive heat leads to the development of smaller and less dense grains with lower commercial quality (Lu et al., 2025 ). Song et al. ( 2022 ) reported an 8% reduction in rice yield for each 1°C increase in mean air temperature. When this increase occurred during the vegetative stage, the effect on yield was less severe. Xiong et al. ( 2017 ) further observed that the physiological processes underlying rice yield formation are differentially affected by high daytime and nighttime temperatures. This distinction helps explain the marked difference in the predictive importance of maximum and minimum temperatures, which play distinct roles across crop developmental stages. For common bean during period (a) (1974–1990; Fig. 7 C), SHAP importance analysis indicated that the variables with the greatest influence on yield were municipality (85.2%), followed by accumulated rainfall (7.5%), minimum temperature (5.9%), and maximum air temperature (1.4%). As observed for rice, the high importance attributed to municipality may reflect edaphoclimatic, topographic, and management-related differences among production regions, which directly affect crop performance. When the individual effects of the climatic variables on common bean yield are considered (Fig. 8 C) for period (a) (1974–1990), a pattern similar to that observed for rice is evident. With respect to accumulated rainfall, the El Niño phase (red) contributed to yield increases (positive values), whereas La Niña (blue) and Neutral (gray) conditions exerted predominantly negative effects. For minimum temperature, the La Niña phase (blue) showed a negative effect, whereas El Niño and Neutral phases (yellow) positively influenced yield. In contrast, maximum temperature exhibited a comparatively smaller effect on yield, particularly under Neutral ENSO conditions (gray). Among the climatic variables, accumulated rainfall showed substantial predictive importance, indicating a high sensitivity of common bean yield to water availability. The water requirement of common bean ranges from approximately 350 mm during the vegetative stage to 200–350 mm during the reproductive stage (Heinemann et al., 2022 ). In this context, the most critical stage for drought stress extends from pre-flowering to the end of grain filling (Soureshjani et al., 2019 ; Androcioli et al., 2020 ). During flowering, water deficit promotes flower abscission, whereas during grain formation it causes abortion of fertilized ovules within pods (Mathobo et al., 2017 ). In common bean, stress during the reproductive stage causes, among other effects, yield reduction (Rosales-Serna et al., 2004 ; Labastida et al., 2023 ), attenuation of photosynthetic performance (Mladenov et al., 2023 ; Papathanasiou et al., 2022 ; Polania et al., 2022 ), shortening of the crop cycle (Beebe et al., 2013 ; Labastida et al., 2023 ), and reductions in pod and seed formation (Labastida et al., 2023 ; Gonçalves et al., 2022 ). Maximum and minimum air temperatures also contributed to the prediction of common bean yield, although to a lesser extent (Fig. 7 C). A similar pattern was observed in the more recent period (b) (1991–2019; Fig. 7 D), in which the importance of municipality became even more pronounced relative to the first period, accounting for 90.6% of the predictive importance. Rainfall remained the second most important variable (3.7%), followed by minimum (3.3%) and maximum (2.4%) air temperatures. In period (b) (1991–2019), accumulated rainfall and minimum air temperature exerted more positive effects on common bean yield during El Niño years (red) than during La Niña years (blue) (Fig. 8 D). The influence of these variables under Neutral conditions was less pronounced (gray). In contrast, for maximum temperature, the Neutral phase (gray) showed a smaller magnitude of effect on yield, whereas the joint evaluation of El Niño and La Niña (purple) resulted in a predominantly negative impact. Recent decades have been characterized by rising air temperatures, a greater frequency of extreme events, and increasing irregularity and intensification of ENSO phenomena (Yang et al., 2018 ). These factors may have contributed to the increased importance of maximum temperature in yield prediction between the first and the most recent periods. Whereas four La Niña events and two El Niño events occurred between 1974 and 1990, the more recent period (1991–2019) included six La Niña and four El Niño events. This higher frequency of ENSO events contributes to increased thermal variability in Rio Grande do Sul and may result in substantial yield variation. During La Niña episodes, which were the more frequent events in the recent period, elevated temperatures at certain crop stages combined with reduced rainfall (Figs. 2 D, 4 D, and 5 D) in southern Brazil favor both water and heat stress in plants (Lopes et al., 2022 ). For adequate common bean development, the minimum, optimum, and maximum air temperatures are 12°C, 21°C, and 29°C, respectively (Vieira et al., 2006 ). Excessively high temperatures can cause physical, physiological, and biochemical damage to plants, including alterations in protein structure and synthesis, changes in enzymatic activity, injury to membranes and the photosynthetic apparatus, and oxidative stress (Fahad et al., 2017 ; Taiz et al., 2017 ). Near flowering, high temperatures can reduce pollen production and viability (Porch & Jahn, 2001 ; Prasad et al., 2002 ), promote flower and pod abscission (Ofir et al., 1993 ), impair seed formation (Vargas et al., 2021 ), and reduce yield (Soltani et al., 2019 ). Heinemann et al. ( 2025 ) also demonstrated that common bean is particularly vulnerable during the reproductive stage as minimum temperatures increase. For soybean, the predictive importance analysis for yield in Rio Grande do Sul during 1974–1990 (Fig. 7 E) indicated a predominance of municipality, accounting for 77.7% of the mean absolute importance, followed by accumulated rainfall (15.8%), maximum temperature (5.3%), and minimum air temperature (1.3%). As observed for rice and common bean (Figs. 7 A and 7 C), the high contribution of municipality reflects the combined influence of edaphoclimatic, topographic, and management-related factors that vary spatially and condition crop yield. During this period, changes in minimum and maximum air temperatures associated with ENSO phases produced effects of comparatively lower magnitude on soybean yield. In contrast, more substantial effects were observed for accumulated rainfall, particularly during El Niño years (red), when its influence was positive (Fig. 8 E). Soybean is highly sensitive to water stress, especially during flowering and grain filling, stages during which water deficiency may alter enzymatic and hormonal activity (Zhou et al., 2022 ), reduce photosynthetic carbon assimilation (Du et al., 2020a ; Tavares et al., 2022 ), cause tissue damage and oxidative stress (Du et al., 2020b ), impair nutrient assimilation (Kaya et al., 2024 ), reduce biomass accumulation and seed number and weight (Poudel et al., 2023 ), and ultimately compromise yield (Singh et al., 2021 ). In the more recent period ((b) 1991–2019), a substantial increase in the predictive importance of municipality (89%) for soybean yield was observed, accompanied by reductions in the importance of accumulated rainfall (4.2%) and maximum air temperature (3.1%) (Fig. 7 F). This shift relative to the previous period (1974–1990) suggests that local factors have become even more dominant in determining crop yield, which may reflect technological advances, including the adoption of more productive and better-adapted cultivars, as well as improvements in crop management practices. It is important to emphasize that the adoption of such practices reduces crop vulnerability to climatic variability and increases production resilience (Raza et al., 2019 ). Nevertheless, the analysis indicates that, although technological progress and management have gained importance over the decades, the effect of ENSO on water availability remains a key factor for understanding soybean yield variability in the region. The marked reduction in the predictive importance of accumulated rainfall between periods (a) (1974–1990) and (b) (1991–2019), observed across all analyzed crops (rice, common bean, and soybean), is consistent with the climatic analysis presented in Fig. 2 . That analysis showed a narrowing of the rainfall gap among ENSO phases, indicating a more homogeneous precipitation regime in recent decades. During 1974–1990, El Niño events were associated with substantially greater precipitation than the other phases (Fig. 2 C), thereby contributing to greater yield variability linked to water availability. In contrast, during 1991–2019 (Fig. 2 D), although differences among phases remained statistically significant (Table 2 ), total precipitation amounts were more similar, resulting in a reduction in the relative importance of rainfall for yield prediction. Consistent with the pattern observed for 1974–1990, in the more recent period ((b) 1991–2019) (Fig. 8 F), accumulated rainfall during El Niño years (red) favored increases in soybean yield. The most negative effects across all variables were observed during La Niña years (blue), whereas the least pronounced effects occurred under Neutral conditions (gray) (Fig. 8 F). Von Bloh et al. ( 2023 ) reported that soybean yield estimates in the Southern and Northeastern regions of Brazil showed the poorest predictive performance among all regions of the country. This result was attributed to the high interannual yield variability characteristic of these areas, which are strongly influenced by ENSO. Such climatic instability increases uncertainty in predictive models and hinders the identification of consistent yield patterns, particularly in the more recent period, which was characterized by a higher frequency of both La Niña and El Niño events. 4 CONCLUSION ENSO exerts a strong influence on climatic variability and, consequently, on the productivity of lowland rice, common bean, and soybean, particularly in Rio Grande do Sul, where accumulated rainfall and maximum and minimum air temperatures differ significantly among El Niño, La Niña, and Neutral phases. In the state of Goiás, by contrast, the divergence among ENSO phases is less evident. Overall, the El Niño phase tended to favor yield increases across all evaluated crops (lowland rice, common bean, and soybean), whereas the La Niña phase was frequently associated with yield reductions. Under Neutral conditions, the influence of ENSO was less pronounced. In period (a) (1974–1990), the El Niño phase did not differ statistically from the Neutral phase with respect to minimum temperature, nor from the La Niña phase with respect to maximum temperature. In period (b) (1991–2019), El Niño again showed no statistically significant difference from La Niña for maximum temperature. These findings indicate an overlap in thermal patterns between El Niño and other ENSO phases. For all evaluated crops, a reduction in the predictive importance of accumulated rainfall was observed in the more recent period (1991–2019), reflecting a more homogeneous precipitation pattern among ENSO phases. In contrast, local factors, including soil characteristics, topography, technological level, and management practices, assumed an even greater role in determining yield. Despite this shift, ENSO remains a key factor in explaining yield variability, reinforcing the need for agricultural planning and management strategies that are adapted to variable climatic conditions. Declarations AUTHORS' CONTRIBUTION Conceptualization (DHM; ABH) Data Curation (ABH; DCV); Formal Analysis (MVCS; DCV); Investigation (MVCS; DHM; ABH; LFJ); Methodology (MVCS; DHM; ABH; LFJ); Project Administration (ABH)); Resources (ABH); Software (MVCS; LFJ); Supervision (DHM; ABH); Validation (SVC; LFS); Visualization (MVCS); Writing – Original Draft (DHM; ABH; LFS); Writing (SVQ; DHM; ABH; LFS; DVC); Review & Editing (DHM; ABH; LFS; SVC; DVC). ACKNOWLEDGMENTS AB Heinemann acknowledges support from the “Conselho Nacional de Desenvolvimento Científico e Tecnológico” (CNPq Nº 4/2021 - 310209/2021-8). References Alvares CA, Stape JL, Sentelhas PC, Gonçalves JLM, Sparovek G (2013) Koppen’s climate classification map for Brazil. Meteorol Z 22:711–728. https://doi.org/10.1127/0941-2948/2013/0507 An D, Eggeling J, Zhang L, He H, Sapkota A, Wang Y, Gao C (2023) Extreme precipitation patterns in the Asia–Pacific region and its correlation with El Niño-Southern Oscillation (ENSO). Scientific Reports 13: 1-12. https://doi.org/10.1038/s41598-023-38317-0 Androcioli LG, Zeffa DM, Alves DS, Tomaz JP, Moda-Cirino V (2020) Effect of water deficit on morphoagronomic and physiological traits of common bean genotypes with contrasting drought tolerance. Water 12:1–13. https://doi.org/10.3390/w12010217 Arruda AM, Centeno LN, Nunes AB (2025) Relation between major climatic indices and subseasonal precipitation in Rio Grande do Sul State, Brazil. Meteorology 4:1–26. https://doi.org/10.3390/meteorology4010005 Beebe SE, Rao IM, Blair MB, Acosta-Gallegos JA (2013) Phenotyping common beans for adaptation to drought. Front Physiol 4:1–20.https://doi.org/10.3389/fphys.2013.00035 Beyaztas U, Yaseen ZM (2019) Drought interval simulation using functional data analysis. J Hydrol 579:1–13. https://doi.org/10.1016/j.jhydrol.2019.124141 Bhandari R, Khatri S, Acharya B, Gauchan U, Adhikari G, Thapa S (2024) Screening of rice genotypes against brown leaf spot ( Bipolaris oryzae ) (Breda de Haan) shoemaker, under field and laboratory conditions at Banke, Nepal. Discover Plants 1:1–12. https://doi.org/10.1007/s44372-024-00008-z Bottazzi F, Rossa ED (2017) A Functional Data Analysis Approach to Surrogate Modeling in Reservoir and Geomechanics Uncertainty Quantification. Math Geosci 49:517–540. https://doi.org/10.1007/s11004-017-9685-y Cai W, Ng B, Geng T, Wu L, Santoso A, McPhaden MJ (2020) Butterfly effect and a self-modulating El Niño response to global warming. Nature 585:68–73. https://doi.org/10.1038/s41586-020-2641-x Cai W, Santoso A, Collins M, Dewitte B, Karamperidou C, Kug J, Lengaigne M, McPhaden MJ, Stuecker MF, Taschetto AS, Timmermann A, Wu L, Yeh S, Wang G, Ng B, Jia F, Yang Y, Ying J, Zheng X, Bayr T, Brown JR, Capotondi A, Cobb KM, Gan B, Geng T, Ham Y, Jin F, Jo H, Li X, Lin X, McGregor S, Park J, Stein K, Yang K, Zhang L, Zhong W (2021) Changing El Niño–Southern Oscillation in a warming climate. Nat Rev Earth Environ 2:628–644. https://doi.org/10.1038/s43017-021-00199-z Cao J, Zhang S, Tao F, Chen Y, Luo X, Xie J (2023) Forecasting global crop yields based on El Nino Southern Oscillation early signals.Agric Syst 205:1–12. https://doi.org/10.1016/j.agsy.2022.103564 Cirino PH, Féres JG, Braga MJ, Reis E (2015) Assessing the Impacts of ENSO-related weather effects on the Brazilian Agriculture. Procedia Econ Finance 24:146–155. https://doi.org/10.1016/S2212-5671(15)00635-8 Conab - Companhia Nacional de Abastecimento (2025) Acompanhamento da safra brasileira Grãos, 10º levantamento - safra 2024/25. Brasília 12:1-125. https://www.gov.br/conab/pt-br/atuacao/informacoes-agropecuarias/safras/safra-de-graos Acessed 07 august 2025 Crespo F, Fontana D, Schirmbeck J (2024) Effects of a Strong ENSO Phenomenon on Evapotranspiration Flux Dynamics in the State of Rio Grande do Sul (RS), Brazil. Rev Bras Meteorol 39:1–9. http://dx.doi.org/10.1590/0102-77863910034 Curceac S, Ternynck C, Ouarda TBMJ, Chebana F, Niang SD (2019) Short-term air temperature forecasting using nonparametric functional data analysis and SARMA models. Environ Model Softw 111:394–408. https://doi.org/10.1016/j.envsoft.2018.09.017 Custodio MS (2016) Índice de nebulosidade diurna do estado do Rio Grande do Sul: climatologia e impactos do El Niño oscilação sul. Cienc Nat 38:382–392. https://doi.org/10.5902/2179460X18175 Du Y, Zhao Q, Chen L, Yao X, Zhang H, Wu J, Xie F (2020a) Effect of Drought Stress during Soybean R2–R6 Growth Stages on Sucrose Metabolism in Leaf and Seed. Int J Mol Sci 21:1–19. https://doi.org/10.3390/ijms21020618 Du Y, Zhao Q, Chen L, Yao X, Xie F (2020b) Effect of Drought Stress at Reproductive Stages on Growth and Nitrogen Metabolism in Soybean. Agronomy 10:1–21. https://doi.org/10.3390/agronomy10020302 Fahad S, Bajwa AA, Nazir U, Anjum SA, Farooq A, Zohaib A, Sadia S, Nasim W, Adkins S, Saud S, Ihsan MZ, Alharby H, Wu C, Wang D, Huang J (2017) Crop Production under Drought and Heat Stress: Plant Responses and Management Options. Front Plant Sci 8:1–16. https://doi.org/10.3389/fpls.2017.01147 Fan B, Tao W, Qin G, Hopkins I, Zhang Y, Wang Q, Lin H, Guo L (2020) Soil micro-climate variation in relation to slope aspect, position, and curvature in a forested catchment. Agric For Meteorol 290:1–17. https://doi.org/10.1016/j.agrformet.2020.107999 Fan D, Gan W (2025) Does the adoption of Internet digital technologies improve farmers’ fertilizer management performance? An empirical analysis based on the China Rural Revitalization Survey. Int J Agric Sustain 23:1–18. https://doi.org/10.1080/14735903.2025.2467619 Febrero-Bande M, De la Fuente MO (2012) Statistical computing in functional data analysis: The R Package fda.usc. J Stat Softw 51:1–28. https://doi.org/10.18637/jss.v051.i04 Feix RD, Leusin Júnior S, Borges BK, Pessoa ML (2022) Painel do Agronegócio do Rio Grande do Sul – 2022. Porto Alegre: Secretaria de Planejamento, Governança e Gestão, p 78 Feng S, Hao Z (2021) Quantitative contribution of ENSO to precipitation-temperature dependence and associated compound dry and hot events. Atmos Res 260:1–7. https://doi.org/10.1016/j.atmosres.2021.105695 Fontana DC, Almeida TS (2002) Climatologia do número de dias com precipitação pluvial no Estado do Rio Grande do Sul. Rev Bras Agrometeorol 10:135–145 Furtak K, Wolińska A (2023) The impact of extreme weather events as a consequence of climate change on the soil moisture and on the quality of the soil environment and agriculture - A review. Catena 231:1–15. https://doi.org/10.1016/j.catena.2023.107378 Gonçalves GMC, Gonçalves JGR, Carvalho Paulino JF, Almeida CP, Carbonell SAM, Chiorato AF (2022) Water deficit on the physiological, morphoagronomic, and technological traits of carioca common bean genotypes. Sci Agric 79:1–10. https://doi.org/10.1590/1678-992X-2021-0016 Górecki T, Smaga L (2017) fdANOVA: an R software package for analysis of variance for univariate and multivariate functional data.Comput Stat 34:571–597.https://doi.org/10.1007/s00180-018-0842-7 Grimm AM (2003) The El Niño impact on the summer monsoon in Brazil: regional processes versus remote influences. J Clim 16:263–280. https://doi.org/10.1175/1520-0442(2003)0162.0.CO;2 Guimarães DP, Reis R (2012) Impactos do fenômeno ENOS sobre a temperatura no Brasil. Rev Espinhaco 1:34–40. https://doi.org/10.5281/zenodo.3968883 Heinemann AB, Sentelhas PC (2011) Environmental group identification for upland rice production in central Brazil. Sci Agric 68:540–547. https://doi.org/10.1590/S0103-90162011000500005 Heinemann AB, Costa-Neto G, Fritsche-Neto R, Matta DH, Fernandes IK (2022) Enviromic prediction is useful to define the limits of climate adaptation: A case study of common bean in Brazil. Field Crops Res 286:1–19. https://doi.org/10.1016/j.fcr.2022.108628 Heinemann AB, Matta DH, Stone LF, Costa-Neto G, Resende RT, Gonçalves PAO, Justino LF (2025) Envirotyping-informed mixed models to study the climatic drivers and yield seasonal variation for common beans in Brazil. Eur J Agron 171:1–15. https://doi.org/10.1016/j.eja.2025.127821 Hintz G, Carcedo A, Almeida LF, Corassa G, Horbe T, Pott L, Schwalbert R, Hefley T, Prasad PVV, Ciampitti I (2025) Climate-adaptative management strategies for soybean production under ENSO scenarios in Southern Brazil: An in-silico analysis of crop failure risk. Agric Syst 222:1–11. https://doi.org/10.1016/j.agsy.2024.104153 Hofmann GS, Cardoso MF, Alves RJV, Weber EJ, Barbosa AA, Toledo PM, Pontual FB, Salles LO, Hasenack H, Cordeiro JLP, Aquino FE, Oliveira LFB (2021) The Brazilian Cerrado is becoming hotter and drier. Glob Change Biol 27:1–14. https://doi.org/10.1111/gcb.15712 Holm S (1979) A simple sequentially rejective multiple test procedure. Scand J Stat 6:65–70 IBGE - Instituto Brasileiro de Geografia e Estatística (2020) Produção Agrícola Municipal 2020 https://sidra.ibge.gov.br/pesquisa/pam/tabelas Acessed 03 april 2024 Iizumi T, Luo J, Challinor AJ, Sakurai G, Yokozawa M, Sakuma H, Brown ME, Yamagata T (2014) Impacts of El Niño Southern Oscillation on the global yields of major crops. Nat Commun 5:1–7. https://doi.org/10.1038/ncomms4712 INMET - Instituto Nacional de Meteorologia (2020) Normais Climatológicas. https://clima.inmet.gov.br Acessed 13 august 2025 Justino LF, Heinemann AB, Matta DH, Stone LF, Gonçalves PAO, Silva SC (2025) Spatio-temporal dynamics of water stress for common bean production in Goiás, Brazil. Theor Appl Climatol 156:1–21. https://doi.org/10.1007/s00704-025-05444-9 Justino LF, Heinemann AB, Matta DH, Stone LF, Andrade FW, Silva SC (2026) Optimizing Sowing Calendars for Climate-Resilient Common Bean Production in Central-Southern Brazil: A Functional Data Analysis Approach. Resources 15:1-24. https://doi.org/10.3390/resources15030040 Kaya C, Akin S, Sarioglu A, Ashraf M, Alyemeni MN, Ahmad P (2024) Enhancement of soybean tolerance to water stress through regulation of nitrogen and antioxidant defence mechanisms mediated by the synergistic role of salicylic acid and thiourea. Plant Physiol Biochem 207:1-14.https://doi.org/10.1016/j.plaphy.2023.108320 Kirtphaiboon S, Humphries U, Khan A, Yusuf A (2021) Model of rice blast disease under tropical climate conditions. Chaos Solitons Fractals 143:1–8. https://doi.org/10.1016/j.chaos.2020.110530 Kodama Y (1992) Large-scale common features of subtropical precipitation zones (the Baiu Frontal Zone, the SPCZ, and the SACZ) Part I: characteristics of subtropical frontal zones. J Meteorol Soc Jpn 70:813–836. https://doi.org/10.2151/jmsj1965.70.4_813 Kodama Y (1993) Large-scale common features of sub-tropical convergence zones (the Baiu Frontal Zone, the SPCZ, and the SACZ) Part II: conditions of the circulations for generating the STCZs. J Meteorol Soc Jpn 71:581–610. https://doi.org/10.2151/jmsj1965.71.5_581 Kwak IY, Moore CR, Spalding EP, Broman KW (2016) Mapping quantitative trait loci underlying function-valued traits using functional principal component analysis and multi-trait mapping. G3: Genes, Genomes, Genetics 6:79–86. https://doi.org/10.1534/g3.115.024133 Labastida D, Ingvarsson PK, Rendón-Anaya M (2023) Dissecting the genetic basis of drought responses in common bean using natural variation. Front Plant Sci 14:1–12. https://doi.org/10.3389/fpls.2023.1143873 Lenka S, Devi R, Joseph CM, Gouda KC (2022) Effect of large-scale oceanic and atmospheric processes on the Indian summer monsoon. Theor Appl Climatol 147:1561–1576. https://doi.org/10.1007/s00704-021-03896-3 Li J, Xie S, Cook E, Huang G, D’Arrigo R, Liu F, Ma J, Zheng X (2011) Interdecadal modulation of El Niño amplitude during the past millennium. Nat Clim Change 1:114–118. https://doi.org/10.1038/nclimate1086 Li X, Yu J, Ding R (2024) El Niño-La Niña Asymmetries in the Changes of ENSO Complexities and Dynamics Since 1990. Geophys Res Lett 51:1–9. https://doi.org/10.1029/2023GL106395 Liang X, Hu Z, You X (2024) ENSONet: a data-driven ENSO forecasting model with concise spatial location learning parameters and temporal embedding. Clim Dyn 62:4081–4098. https://doi.org/10.1007/s00382-024-07119-z Liaw A, Wiener M (2002) Classification and Regression by randomForest. R News 2:18–22. https://CRAN.R-project.org/doc/Rnews/ Lin J, Qian T (2019) Impacts of the ENSO Lifecycle on Stratospheric Ozone and Temperature. Geophys Res Lett 46:10646–10658. https://doi.org/10.1029/2019GL083697 Liu Y, Cai W, Lin X, Zhang Y (2023) Nonlinear El Niño impacts on the global economy under climate change.Nat Commun 14:1–11. https://doi.org/10.1038/s41467-023-41551-9 Lopes AB, Andreoli RV, Souza RAF, Ceron WL, Kayano MT, Canchala T, Moraes DS (2022) Multiyear La Niña effects on the precipitation in South America. Int J Climatol 42:1–16. https://doi.org/10.1002/joc.7847 Lu F, Feng B, Chen L, Qiu J, Wei X (2025) How Does Rice Cope with High-Temperature Stress During Its Growth and Development, Especially at the Grain-Filling Stage? Agronomy 15:1–16. https://doi.org/10.3390/agronomy15030623 Marengo JA, Liebmann B, Grimm AM, Misra V, Silva Dias PL, Cavalcanti IFA, Carvalho LMV, Berbery EH, Ambrizzi T, Vera CS, Saulo AC, Nogues-Paegle J, Zipser E, Seth A, Alves LM (2012) Recent developments on the South American monsoon system. Int J Climatol 32:1–21. https://doi.org/10.1002/joc.2254 Mathobo R, Marais D, Steyn JM (2017) The effect of drought stress on yield, leaf gaseous exchange and chlorophyll fluorescence of dry beans ( Phaseolus vulgaris L.). Agric Water Manag 180:118–125. https://doi.org/10.1016/j.agwat.2016.11.005 Matta DH, Coelho CAS, Santos LL, Stone LF, Heinemann AB (2023) Analysis of Goiás State rainfall and temperature similarity patterns during the El Niño Southern Oscillation phenomenon phases across the years. Theor Appl Climatol 153:1013–1031. https://doi.org/10.1007/s00704-023-04503-3 Matzenauer R, Radin B, Almeida IR (2011) Atlas Climático do Rio Grande do Sul. Porto Alegre: Fepagro, p. 186 Medeiros FJ, Oliveira CP (2021) Dynamical Aspects of the Recent Strong El Niño Events and Its Climate Impacts in Northeast Brazil. Pure Appl Geophys 178:2315–2332. https://doi.org/10.1007/s00024-021-02758-3 Mladenov P, Aziz S, Topalova E, Renaut J, Planchon S, Raina A, Tomlekova N (2023) Physiological Responses of Common Bean Genotypes to Drought Stress. Agronomy 13:1–14. https://doi.org/10.3390/agronomy13041022 Moura MM, Santos AR, Pezzopane JEM, Alexandre RS, Silva SF, Pimentel SM, Andrade MSS, Silva FGR, Branco ERF, Moreira TR, Silva RG, Carvalho JR (2019) Relation of El Niño and La Niña phenomena to precipitation, evapotranspiration and temperature in the Amazon basin. Sci Total Environ 651:1639–1651. https://doi.org/10.1016/j.scitotenv.2018.09.242 NOAA - National Oceanic and Atmospheric Administration (2020) Historical ENSO episodes (1950–present): cold and warm episodes by 791 season. National Weather Service, Climate Prediction Center. https://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears_ERSSTv3b.shtml Acessed 28 july 2025 Nóia Júnior RS, Sentelhas PC (2019) Soybean-maize succession in Brazil: impacts of sowing dates on climate variability, yield and economic profitability. Eur J Agron 103:140–151. https://doi.org/10.1016/j.eja.2018.12.008 Ofir M, Gross Y, Bangerth F, Kigel J (1993) High temperature effects on pod and seed production as related to hormone levels and abscission of reproductive structures in common bean ( Phaseolus vulgaris L.). Sci Hortic 55:201–211. https://doi.org/10.1016/0304-4238(93)90032-L Omar MM, Massawe BHJ, Shitindi MJ, Pedersen O, Meliyo JL, Fue KG (2024) Assessment of salt-affected soil in selected rice irrigation schemes in Tanzania: understanding salt types for optimizing management approaches. Front Soil Sci 22:1–14. https://doi.org/10.3389/fsoil.2024.1372838 Papathanasiou F, Ninou E, Mylonas I, Baxevanos D, Papadopoulou F, Avdikos I, Sistanis I, Koskosidis A, Vlachostergios DN, Stefanou S, Tigka E, Kargiotidou A (2022) The Evaluation of Common Bean ( Phaseolus vulgaris L.) Genotypes under Water Stress Based on Physiological and Agronomic Parameters. Plants 11:1–15. https://doi.org/10.3390/plants11182432 Penalba OC, Rivera JA (2016) Precipitation response to El Niño/La Niña events in southern South America - emphasis in regional drought occurrences. Adv Geosci 42:1–14. https://doi.org/10.5194/adgeo-42-1-2016 Percich JA, Nyvall RF, Malvick DK, Kohls CL (1997) Interaction of temperature and moisture on infection of wild rice by Bipolaris oryzae in the growth chamber. Plant Dis 81:1193–1195. https://doi.org/10.1094/PDIS.1997.81.10.1193 Pérez-Plaza S, Fernández-Palacín F, Berrocoso M, Páez R, Rosado B (2018) Analysis of a GPS Network Based on Functional Data Analysis. Math Geosci 50:659–677. https://doi.org/10.1007/s11004-018-9731-4 Polania JA, Salazar-Chavarría V, Gonzalez-Lemes I, Acosta-Maspons A, Chater CCC, Covarrubias AA (2022) Contrasting Phaseolus Crop Water Use Patterns and Stomatal Dynamics in Response to Terminal Drought. Front Plant Sci:1–16. https://doi.org/10.3389/fpls.2022.894657 Poudel S, Vennam RR, Shrestha A, Reddy KR, Wijewardane NK, Reddy KN, Bheemanahalli R (2023) Resilience of soybean cultivars to drought stress during flowering and early-seed setting stages. Sci Rep 13:1–13. https://doi.org/10.1038/s41598-023-28354-0 Porch TG, Jahn M (2001) Effects of high-temperature stress on microsporogenesis in heat-sensitive and heat-tolerant genotypes of Phaseolus vulgaris . Plant Cell Environ 24:723–731. https://doi.org/10.1046/j.1365-3040.2001.00716.x Prasad PVV, Boote KJ, Allen H, Thomas JMG (2002) Effects of elevated temperature and carbon dioxide on seed-set and yield of kidney bean ( Phaseolus vulgaris L.). Glob Change Biol 8:710–721. https://doi.org/10.1046/j.1365-2486.2002.00508.x Qi Z, Cui C, Jiang Y, Chen Y, Ju J, Guo N (2022) Changes in the spatial and temporal characteristics of China’s arid region in the background of ENSO. Sci Rep 12:1–16. https://doi.org/10.1038/s41598-022-21712-4 Qian Y, Zhao J, Zheng S, Cao Y, Xue L (2020) Risk assessment of the global crop loss in ENSO events. Phys Chem Earth 116:1–10.https://doi.org/10.1016/j.pce.2020.102845 R Core Team (2025) R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. https://www.R-project.org/ Raza A, Razzaq A, Mehmood SS, Zou X, Zhang X, Lv Y, Xu J (2019) Impact of Climate Change on Crops Adaptation and Strategies to Tackle Its Outcome: A Review. Plants 8:1–29. https://doi.org/10.3390/plants8020034 Regoto P, Dereczynski C, Chou SC, Bazzanela AC (2021) Observed changes in air temperature and precipitation extremes over Brazil. Int J Climatol 41:1–18. https://doi.org/10.1002/joc.7119 Rosales-Serna R, Kohashi-Shibata J, Acosta-Gallegos JA, Trejo-López C, Ortiz-Cereceres J, Kelly JD (2004) Biomass distribution, maturity acceleration and yield in drought-stressed common bean cultivars. Field Crops Res 85:203–211. https://doi.org/10.1016/S0378-4290(03)00161-8 Sattar MA, Mia S, Shanta AA, Biswas AKMAA, Ludwig F (2021) Remote Impacts from El Niño and La Niña on Climate Variables and Major Crops Production in Coastal Bangladesh. Atmosphere 12:1–19. https://doi.org/10.3390/atmos12111449 Scheibel CH, Nascimento AB, Araújo Júnior GN, Almeida ACS, Silva TGF, Silva JLP, Silva Junior FB, Farias JA, Santos JPAS, Oliveira-Júnior JF, Silva JLB, João FM, Deus AS, Teodoro I, Oliveira HFE, Silva MV (2024) Characterization of Water Bodies through Hydro-Physical Indices and Anthropogenic Effects in the Eastern Northeast of Brazil. Climate 12:1–37. https://doi.org/10.3390/cli12090150 Shah DA, De Wolf ED, Paul PA, Madden LV (2018) Functional data analysis of weather variables linked to Fusarium head blight epidemics in the United States. Phytopathology 109:96–110. https://doi.org/10.1094/PHYTO-11-17-0386-R Shapley LS (1953) A value for n-person games. In: Kuhn H, Tucker A (eds) Contributions to the Theory of Games II. Princeton: Princeton University Press, pp 307–317 Sharda R, Voß S, Suthaharan S (2019) Machine Learning Models and Algorithms for Big Data Classification: Thinking with Examples for Effective Learning. New York: Springer, p. 359 Siemers FM, Bajorath J (2023) Differences in learning characteristics between support vector machine and random forest models for compound classification revealed by Shapley value analysis. Sci Rep 13:1–12.https://doi.org/10.1038/s41598-023-33215-x Silva KA, Rolim GS, Valeriano TTB, Moraes JRSC (2020) Influence of El Niño and La Niña on coffee yield in the main coffee-producing regions of Brazil. Theor Appl Climatol 139:1019–1029. https://doi.org/10.1007/s00704-019-03039-9 Singh S, Sah S, Singh RN (2023) The 2023-24 El Niño event and its possible global consequences on food security with emphasis on India. Food Secur 15:1431–1436. https://doi.org/10.1007/s12571-023-01419-8 Singh SK, Reddy VR, Devi MJ, Timlin DJ (2021) Impact of water stress under ambient and elevated carbon dioxide across three temperature regimes on soybean canopy gas exchange and productivity. Sci Rep 11:1–13. https://doi.org/10.1038/s41598-021-96037-9 Sohn SI, Pandian S, Kumar TS, Zoclanclounon YAB, Muthuramalingam P, Shilpha J, Satish L, Ramesh M (2021) Seed Dormancy and Pre-Harvest Sprouting in Rice—An Updated Overview. Int J Mol Sci 22:1–22. https://doi.org/10.3390/ijms222111804 Soltani A, Weraduwage SM, Sharkey TD, Lowry DB (2019) Elevated temperatures cause loss of seed set in common bean ( Phaseolus vulgaris L.) potentially through the disruption of source-sink relationships. BMC Genomics 20:1–18. https://doi.org/10.1186/s12864-019-5669-2 Song Y, Wang C, Linderholm HW, Fu Y, Cai W, Xu J, Zhuang L, Wu M, Shi Y, Wang G, Chen D (2022) The negative impact of increasing temperatures on rice yields in southern China. Sci Total Environ 820:1–7.https://doi.org/10.1016/j.scitotenv.2022.153262 Soureshjani HK, Nezami A, Kafi M, Tadayon M (2019) Responses of two common bean ( Phaseolus vulgaris ) genotypes to deficit irrigation. Agric Water Manag 213:270–279.https://doi.org/10.1016/j.agwat.2018.09.038 Souza IP, Andreoli RV, Kayano MT, Vargas FF, Cerón WL, Martins JA, Freitas E, Souza RAF (2021) Seasonal precipitation variability modes over South America associated to El Niño-Southern Oscillation (ENSO) and non-ENSO components during the 1951–2016 period. Int J Climatol 41:1–18. https://doi.org/10.1002/joc.7075 Taiz L, Zeiger E, Moller I, Murphy A (2017) Fisiologia e desenvolvimento vegetal, 6th ed. Porto Alegre: Artmed, p. 888 Tamaddun KA, Kalra A, Bernardez M, Ahmad S (2019) Effects of ENSO on Temperature, Precipitation, and Potential Evapotranspiration of North India’s Monsoon: An Analysis of Trend and Entropy. Water 11:1–21. https://doi.org/10.3390/w11020189 Tavares CJ, Ribeiro Junior WQ, Ramos MLG, Pereira LF, Casari RACN, Pereira AF, Sousa CAF, Silva AR, Neto SPS, Mertz-Henning LM (2022) Water Stress Alters Morphophysiological, Grain Quality and Vegetation Indices of Soybean Cultivars. Plants 11:1–17. https://doi.org/10.3390/plants11040559 Teegavarapu RSV, Sharma PJ (2021) Influences of climate variability on regional precipitation and temperature associations. Hydrol Sci J 66:2395–2414. https://doi.org/10.1080/02626667.2021.1994976 Van der Wiel K, Matthews AJ, Stevens DP, Joshi MM (2015) A dynamical framework for the origin of the diagonal South Pacific and South Atlantic Convergence Zones. Q J R Meteorol Soc 141:1997–2010. https://doi.org/10.1002/qj.2508 Vargas Y, Mayor-Duran VM, Buendia HF, Ruiz-Guzman H, Raatz B (2021) Physiological and genetic characterization of heat stress effects in a common bean RIL population. PLoS One 16:1–23.https://doi.org/10.1371/journal.pone.0249859 Vera C, Higgins W, Amador J, Ambrizzi T, Garreaud R, Gochis D, Gutzler D, Lettenmaier D, Marengo J, Mechoso CR, Nogues-Paegle J, Dias PLS, Zhang C (2006) Toward a unified view of the American monsoon systems. J Clim 19:4977–5000.https://doi.org/10.1175/JCLI3896.1 Vieira C, Paula Junior TJ, Borém A (eds) (2006) Feijão, 2nd edn. Viçosa: UFV, pp. 13-18 Von Bloh M, Nóia Júnior RS, Wangerpohl X, Saltik AO, Haller V, Kaiser L, Asseng S (2023) Machine learning for soybean yield forecasting in Brazil. Agric For Meteorol 341:1–12. https://doi.org/10.1016/j.agrformet.2023.109670 Wang D, Li X, Tian S, He L, Xu Y, Wang X (2021) Quantifying the dynamics between environmental information disclosure and firms’ financial performance using functional data analysis. Sustain Prod Consum 28:192–205. https://doi.org/10.1016/j.spc.2021.03.026 Wu X, Jiang D, Yang Y (2025) ENSO exacerbated the impact of compound dry–hot events on maize yield over China during 1961–2020. Glob Planet Change 251:1–10. https://doi.org/10.1016/j.gloplacha.2025.104828 Xavier AC, Scanlon BR, King CW, Alves AI (2022) New Improved Brazilian Daily Weather Gridded Data (1961–2020). Int J Climatol 42:8390-8404. https://doi.org/10.1002/joc.7731 Xiong D, Ling X, Huang J, Peng S (2017) Meta-analysis and dose-response analysis of high temperature effects on rice yield and quality. Environ Exp Bot 141:1–9. https://doi.org/10.1016/j.envexpbot.2017.06.00 Xu Y, Li Y, Nettleton D (2018a) Nested Hierarchical Functional Data Modeling and Inference for the Analysis of Functional Plant Phenotypes. J Am Stat Assoc 113:593–606. https://doi.org/10.1080/01621459.2017.1366907 Xu Y, Qiu Y, Schnable JC (2018b) Functional Modeling of Plant Growth Dynamics. Plant Phenome J 1:1–10. https://doi.org/10.2135/tppj2017.09.0007 Yang S, Li Z, Yu J, Hu X, Dong W, He S (2018) El Niño–Southern Oscillation and its impact in the changing climate. Natl Sci Rev 5:840–857. https://doi.org/10.1093/nsr/nwy046 Ye C, Zheng G, Tao Y, Xu Y, Chu G, Xu C, Chen S, Liu Y, Zhang X, Wang D (2024) Effect of Soil Texture on Soil Nutrient Status and Rice Nutrient Absorption in Paddy Soils. Agronomy 14:1–17. https://doi.org/10.3390/agronomy14061339 Yu J, Du T, Zhang P, Ma Z, Chen X, Cao J, Li H, Li T, Zhu Y, Xu F, Hu Q, Liu G, Li G, Wei H (2024) Impacts of High Temperatures on the Growth and Development of Rice and Measures for Heat Tolerance Regulation: A Review. Agronomy 14:1–24. https://doi.org/10.3390/agronomy14122811 Zheng G, Zhang Y, Yue X, Li K (2023) Interpretable prediction of thermal sensation for elderly people based on data sampling, machine learning and SHapley Additive exPlanations (SHAP). Build Environ 242:1–15. https://doi.org/10.1016/j.buildenv.2023.110602 Zhou Q, Li Y, Wang X, Yan C, Ma C, Liu J, Dong S (2022) Effects of Different Drought Degrees on Physiological Characteristics and Endogenous Hormones of Soybean. Plants 11:1–17. https://doi.org/10.3390/plants11172282 Additional Declarations No competing interests reported. Supplementary Files ENSOSuplementarv12.docx Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 16 May, 2026 Reviews received at journal 13 May, 2026 Reviews received at journal 04 May, 2026 Reviewers agreed at journal 15 Apr, 2026 Reviewers agreed at journal 10 Apr, 2026 Reviewers agreed at journal 10 Apr, 2026 Reviewers invited by journal 10 Apr, 2026 Editor assigned by journal 07 Apr, 2026 Submission checks completed at journal 07 Apr, 2026 First submitted to journal 03 Apr, 2026 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-9314260","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":623898611,"identity":"14225d20-c571-4b9a-a9d6-b9de6629642a","order_by":0,"name":"Ludmilla Ferreira Justino","email":"","orcid":"","institution":"Universidade Federal de Goiás","correspondingAuthor":false,"prefix":"","firstName":"Ludmilla","middleName":"Ferreira","lastName":"Justino","suffix":""},{"id":623898612,"identity":"0989a02f-2beb-44f5-9644-36ec6c03eb58","order_by":1,"name":"David Henriques Matta","email":"","orcid":"","institution":"Universidade Federal de Goiás","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"Henriques","lastName":"Matta","suffix":""},{"id":623898613,"identity":"65046ec9-d2fe-40f8-8754-0c4cfe8a1d3c","order_by":2,"name":"Marcos Vinício Cesario dos Santos","email":"","orcid":"","institution":"Universidade Federal de Goiás","correspondingAuthor":false,"prefix":"","firstName":"Marcos","middleName":"Vinício Cesario dos","lastName":"Santos","suffix":""},{"id":623898614,"identity":"a7c6e8b5-1632-4a45-969d-7026f615d7dd","order_by":3,"name":"Luís Fernando Stone","email":"","orcid":"","institution":"Brazilian Agricultural Research Corporation","correspondingAuthor":false,"prefix":"","firstName":"Luís","middleName":"Fernando","lastName":"Stone","suffix":""},{"id":623898615,"identity":"b2420da8-9783-42c8-ba0f-2c03a346594e","order_by":4,"name":"Daniel Castro","email":"","orcid":"","institution":"Brazilian Agricultural Research Corporation","correspondingAuthor":false,"prefix":"","firstName":"Daniel","middleName":"","lastName":"Castro","suffix":""},{"id":623898616,"identity":"aa2682b9-a199-4a90-8c0d-6f0cf71c167d","order_by":5,"name":"Santiago Vianna","email":"","orcid":"","institution":"Brazilian Agricultural Research Corporation","correspondingAuthor":false,"prefix":"","firstName":"Santiago","middleName":"","lastName":"Vianna","suffix":""},{"id":623898617,"identity":"336c0b79-1787-42a5-b395-684ce869f130","order_by":6,"name":"Alexandre Bryan Heinemann","email":"data:image/png;base64,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","orcid":"","institution":"Brazilian Agricultural Research Corporation","correspondingAuthor":true,"prefix":"","firstName":"Alexandre","middleName":"Bryan","lastName":"Heinemann","suffix":""}],"badges":[],"createdAt":"2026-04-03 15:09:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9314260/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9314260/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107362300,"identity":"65e7f06f-afce-42be-8c2e-4f1a291e57c7","added_by":"auto","created_at":"2026-04-20 18:40:02","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1338799,"visible":true,"origin":"","legend":"\u003cp\u003eGeographical distribution of weather stations according to climate classification (A and B) and altitude (C and D) in the states of Goiás (A and C) and Rio Grande do Sul (B and D).\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/2190a37a60fa091ace0ee26a.png"},{"id":107486728,"identity":"df967067-3db9-4966-9cb2-9bf5b08d8ba3","added_by":"auto","created_at":"2026-04-22 02:38:47","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":164443,"visible":true,"origin":"","legend":"\u003cp\u003eMean functional estimates of accumulated rainfall across ENSO phases (El Niño, La Niña, and Neutral) in the states of Goiás (A, B) and Rio Grande do Sul (C, D) for the periods 1961–1990 (A, C) and 1991–2019 (B, D). In Goiás (A, B), the green curve represents the mean functional estimate considering all ENSO phases (non-significant differences). The Neutral phase curve includes the 95% confidence interval (CI) around the functional mean.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/f816d5c5e93ec9debdcc7679.png"},{"id":107362302,"identity":"1b366717-2ed2-4c38-97c5-a6bf2c47bd99","added_by":"auto","created_at":"2026-04-20 18:40:02","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":102498,"visible":true,"origin":"","legend":"\u003cp\u003eRatio of the mean functional estimates of accumulated rainfall between ENSO phases (El Niño/La Niña – black; El Niño/Neutral – red; La Niña/Neutral – blue) for the periods 1961–1990 (A) and 1991–2019 (B).\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/9e4dc5f406da86e3b9aae406.png"},{"id":107362305,"identity":"32410b99-a0af-4e1c-b2f6-3dee26b39bcd","added_by":"auto","created_at":"2026-04-20 18:40:02","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":156303,"visible":true,"origin":"","legend":"\u003cp\u003eMean functional estimates of maximum air temperature across ENSO phases (El Niño, La Niña, and Neutral) in the states of Goiás (A, B) and Rio Grande do Sul (C, D) for the periods 1961–1990 (A, C) and 1991–2019 (B, D). In Goiás (A, B), the green curve represents the mean functional estimate considering all ENSO phases (non-significant differences). The Neutral phase curve includes the 95% confidence interval (CI) around the functional mean.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/3a037446b257c0e5906ddb79.png"},{"id":107362307,"identity":"8a9a4115-2f26-4f1e-9228-af50040f0500","added_by":"auto","created_at":"2026-04-20 18:40:02","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":151404,"visible":true,"origin":"","legend":"\u003cp\u003eMean functional estimates of minimum air temperature across ENSO phases (El Niño, La Niña, and Neutral) in the states of Goiás (A, B) and Rio Grande do Sul (C, D) for the periods 1961–1990 (A, C) and 1991–2019 (B, D). In Goiás (A, B), the green curve represents the mean functional estimate considering all ENSO phases (non-significant differences). The Neutral phase curve includes the 95% confidence interval (CI) around the functional mean.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/0380770470e920bf84651b00.png"},{"id":107486660,"identity":"e87e69bf-5e3b-448e-a467-c53d4232a65e","added_by":"auto","created_at":"2026-04-22 02:38:36","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":206843,"visible":true,"origin":"","legend":"\u003cp\u003eRatio of the mean functional estimates of maximum air temperature (A and B) and minimum air temperature (C and D) between ENSO phases (El Niño/La Niña – black; El Niño/Neutral – red; La Niña/Neutral – blue) for the periods 1961–1990 (A) and 1991–2019 (B).\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/15b8ece1608417ebe176b829.png"},{"id":107362304,"identity":"2c000ef0-41b1-4a7d-9a19-d3993916479a","added_by":"auto","created_at":"2026-04-20 18:40:02","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":259621,"visible":true,"origin":"","legend":"\u003cp\u003eMean absolute predictive importance of climatic variables and municipality for the yields of lowland rice (A, B), common bean (C, D), and soybean (E, F) in the state of Rio Grande do Sul across the periods 1974–1990 (A, C, E) and 1991–2019 (B, D, F), estimated using SHAP values.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/3be3c00b81a7a68d50a67d9b.png"},{"id":107486764,"identity":"9901496c-e33a-4979-8b50-2b6cd725fe51","added_by":"auto","created_at":"2026-04-22 02:38:54","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":196593,"visible":true,"origin":"","legend":"\u003cp\u003eSHAP values quantifying the contribution of each observation of accumulated rainfall, maximum air temperature, and minimum air temperature to the overall mean prediction of the \u003cem\u003eRandom Forest\u003c/em\u003e yield model for lowland rice (A, B), common bean (C, D), and soybean (E, F) in the state of Rio Grande do Sul across the periods 1974–1990 (A, C, E) and 1991–2019 (B, D, F). Colors represent ENSO phases evaluated separately for each climatic variable, either individually (Neutral – gray; El Niño – red; La Niña – blue) or grouped into statistically similar clusters (El Niño and Neutral – yellow; El Niño and La Niña – purple), according to the significance of ENSO phases (see Table 2).\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/9611c5fd0fa2dc3fcb0901b4.png"},{"id":107705119,"identity":"fbd6ce2b-5ed8-4a15-b4d1-2f8ee4aead20","added_by":"auto","created_at":"2026-04-24 09:08:22","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2826235,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/881f5692-cdc6-4ed1-a920-3bc3e5b15e37.pdf"},{"id":107362299,"identity":"28c40266-2cd7-4eae-8390-a3f4bc11b0d7","added_by":"auto","created_at":"2026-04-20 18:40:02","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":3601629,"visible":true,"origin":"","legend":"","description":"","filename":"ENSOSuplementarv12.docx","url":"https://assets-eu.researchsquare.com/files/rs-9314260/v1/b471975a8d8bb5ae77a9f6bb.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"ENSO-driven climate variability and crop yield responses in contrasting Brazilian regions: insights from functional data analysis and machine learning","fulltext":[{"header":"1 INTRODUCTION","content":"\u003cp\u003eAgricultural production constitutes a fundamental pillar of the Brazilian economy, contributing substantially to the trade balance, Gross Domestic Product (GDP), and employment generation. In this context, the state of Rio Grande do Sul stands out as the fourth-largest grain producer in the country, with a production of 33.2\u0026nbsp;million tons in the 2024/2025 growing season, corresponding to 9.8% of national output. Although soybean (\u003cem\u003eGlycine max\u003c/em\u003e L.; 43%) and irrigated rice (\u003cem\u003eOryza sativa\u003c/em\u003e L.; 25%) account for the largest share of this production, crops with lower relative participation, such as common bean (\u003cem\u003ePhaseolus vulgaris\u003c/em\u003e L.), representing only 0.2% of state production, hold considerable socioeconomic importance, particularly within family farming systems (Conab, 2025; Feix et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Furthermore, Rio Grande do Sul is the leading producer of irrigated rice in Brazil, with 8.3\u0026nbsp;million tons (74.3% of national production), and ranks fourth in soybean production, with 14.3\u0026nbsp;million tons (8.4% of the national total). Common bean production, predominantly of the \u0026ldquo;black\u0026rdquo; commercial type, reaches 56.2 thousand tons, equivalent to 7.1% of Brazilian production (Conab, 2025).\u003c/p\u003e \u003cp\u003eSimilarly, the state of Goi\u0026aacute;s plays a strategic role in national agricultural production, ranking as the third-largest grain producer, with 35\u0026nbsp;million tons in the 2024/2025 season, representing approximately 10% of total Brazilian production. In this state, soybean is the dominant crop (58%), followed by maize (36%), sorghum (4%), and common bean (0.9%) (Conab, 2025).\u003c/p\u003e \u003cp\u003eDespite their agricultural relevance, these two states exhibit markedly contrasting climatic conditions. Goi\u0026aacute;s is located in Central-East South America (CESA), at the core of the South American Monsoon System (Vera et al., \u003cspan citationid=\"CR102\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Marengo et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), which is characterized by a well-defined rainy season during the austral summer (December\u0026ndash;February, DJF) and a dry season during winter (June\u0026ndash;August, JJA). The region is also influenced by the South Atlantic Convergence Zone (SACZ) (Kodama, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e1992\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e1993\u003c/span\u003e; Van Der Wiel et al., \u003cspan citationid=\"CR100\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), whose activity significantly intensifies rainfall volumes. In addition, due to its location within a transition zone between northern and southern Brazil, Goi\u0026aacute;s exhibits sensitivity to teleconnections associated with the El Ni\u0026ntilde;o\u0026ndash;Southern Oscillation (ENSO) phenomenon (Grimm, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Penalba \u0026amp; Rivera, \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Moura et al., \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; N\u0026oacute;ia J\u0026uacute;nior \u0026amp; Sentelhas, \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Matta et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eConversely, Rio Grande do Sul is part of the Southeast South America (SESA) region, where ENSO-related effects are particularly pronounced (Crespo et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). El Ni\u0026ntilde;o events are typically associated with increased rainfall (Medeiros \u0026amp; Oliveira, \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Arruda et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), a higher frequency of rainy days (Fontana \u0026amp; Almeida, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2002\u003c/span\u003e), increased cloud cover (Custodio, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), and elevated air temperatures (Guimar\u0026atilde;es \u0026amp; Reis, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). In contrast, La Ni\u0026ntilde;a events are generally associated with reduced rainfall in southern Brazil and increased precipitation in the North and Northeast regions, often resulting in drought conditions in Rio Grande do Sul (Lopes et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Scheibel et al., \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe impacts of ENSO on agricultural systems are widely documented and vary according to region, crop type, and the specific phase of the phenomenon (Iizumi et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Cao et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These effects are particularly critical for rainfed crops, such as soybean (Qian et al., \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and common bean (Cirino et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), which are highly sensitive to climatic variability and extreme events. Although irrigated rice is less vulnerable to rainfall fluctuations, variations in air temperature and solar radiation may negatively affect grain yield (Qian et al., \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Singh et al., \u003cspan citationid=\"CR89\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), with potential consequences for supply and market prices, thereby impacting food security (Liu et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eRecent advances have improved the understanding of ENSO mechanisms and their spatiotemporal patterns, as well as their impacts on agricultural systems (Qi et al., \u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Liang et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Hintz et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Wu et al., \u003cspan citationid=\"CR106\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). However, important gaps remain in the integrated understanding of crop responses to climatic variability associated with ENSO, particularly in studies based on long-term historical datasets. In this context, advanced analytical approaches, including machine learning techniques, are essential for identifying the climatic variables that most strongly influence crop performance under different ENSO phases.\u003c/p\u003e \u003cp\u003eGiven this scenario, the present study aimed to evaluate the influence of ENSO on rainfall, air temperature, and the yield of irrigated rice, common bean, and soybean across two distinct periods (1961\u0026ndash;1990 and 1991\u0026ndash;2019) in the states of Goi\u0026aacute;s and Rio Grande do Sul, Brazil.\u003c/p\u003e"},{"header":"2 MATERIALS AND METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 \u003cspan type=\"SmallCaps\" class=\"SmallCaps\" name=\"Emphasis\"\u003eDescription of the Study Area\u003c/span\u003e\u003c/h2\u003e \u003cp\u003eThe study area comprises two Brazilian states with contrasting climatic conditions: Rio Grande do Sul (RS) and Goi\u0026aacute;s (GO), located in the Southern and Central-West regions, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e \u0026ndash; Supplementary Material).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eRio Grande do Sul is characterized by a humid subtropical climate, predominantly classified as Cfa (humid subtropical climate, 86.7%) and Cfb (temperate oceanic climate, 13.3%). These climate types are defined by the absence of a well-defined dry season and a relatively uniform distribution of rainfall throughout the year, with annual precipitation ranging from 1000 to 1600 mm in the southern portion and from 1600 to 2000 mm in the northern region (Matzenauer et al., \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). The Cfa climate, which predominates across most of the state, occurs at altitudes between 0 and 650 m and is characterized by hot summers, with mean temperatures below 18\u0026deg;C in the coldest month and above 22\u0026deg;C in the warmest month. In contrast, the Cfb climate occurs at higher altitudes (approximately 900 m) and is characterized by milder summers, mean temperatures below 22\u0026deg;C in the warmest month, and frequent frost events.\u003c/p\u003e \u003cp\u003eIn Goi\u0026aacute;s, the predominant climate is classified as Aw (94%), corresponding to a tropical climate with a marked dry winter. The state exhibits annual precipitation between 1600 and 1900 mm, altitudes reaching up to 1200 m, and mean annual temperatures ranging from 19 to 20\u0026deg;C (Alvares et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 \u003cspan type=\"SmallCaps\" class=\"SmallCaps\" name=\"Emphasis\"\u003eMeteorological Dataset\u003c/span\u003e\u003c/h2\u003e \u003cp\u003eMeteorological stations from the National Institute of Meteorology (INMET) were selected based on the availability of more than 30 years of daily records of maximum and minimum air temperature and/or rainfall, with less than 30% missing or inconsistent data.\u003c/p\u003e \u003cp\u003eFor air temperature, records were considered inconsistent when Tmax\u0026thinsp;\u0026lt;\u0026thinsp;Tmin or when at least three consecutive observations were identical. For rainfall, inconsistencies included extreme daily values (\u0026gt;\u0026thinsp;200 mm) or at least three identical consecutive measurements (excluding zero rainfall). Missing values in both rainfall and temperature series were gap-filled using the BR-DWGD database (Xavier et al., \u003cspan citationid=\"CR107\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBased on these criteria, a total of 69 stations were selected across the study area: 62 located in Rio Grande do Sul and 7 in Goi\u0026aacute;s (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). In Rio Grande do Sul, 16 stations provided temperature data and 46 provided rainfall data, whereas in Goi\u0026aacute;s, five stations provided temperature data and two provided rainfall data.\u003c/p\u003e \u003cp\u003eRainfall data were aggregated into weekly accumulated values, while maximum and minimum air temperatures were represented by the absolute weekly maximum and minimum values. This procedure was adopted to reduce the influence of daily variability and noise associated with non-rainy days. Subsequently, the dataset was filtered according to the agricultural calendar of each region. In Rio Grande do Sul, the analyzed period comprised 43 weeks (September 1st to June 28th), whereas in Goi\u0026aacute;s, it comprised 34 weeks (October 1st to May 26th), encompassing the complete crop development cycles of rice, common bean, and soybean. This selection reflects the sowing windows for irrigated rice, rainfed common bean, and rainfed soybean in Rio Grande do Sul, and for rainfed common bean and soybean in Goi\u0026aacute;s.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 \u003cspan type=\"SmallCaps\" class=\"SmallCaps\" name=\"Emphasis\"\u003eENSO DATA\u003c/span\u003e\u003c/h2\u003e \u003cp\u003eENSO phases are typically defined based on sea surface temperature (SST) anomalies and their persistence across the equatorial Pacific Ocean. In this study, monthly SST data from the Climate Prediction Center (CPC) of the National Oceanic and Atmospheric Administration (NOAA) were used for the period 1961\u0026ndash;2019. The Oceanic Ni\u0026ntilde;o Index (ONI), provided by NOAA, is based on three-month running means of SST anomalies in the Ni\u0026ntilde;o 3.4 region and is widely used to identify El Ni\u0026ntilde;o and La Ni\u0026ntilde;a events. El Ni\u0026ntilde;o events are characterized by positive SST anomalies equal to or greater than 0.5\u0026deg;C above the climatological mean, whereas La Ni\u0026ntilde;a events correspond to negative anomalies equal to or less than \u0026minus;\u0026thinsp;0.5\u0026deg;C. Neutral conditions indicate the absence of significant anomalies (NOAA, 2019).\u003c/p\u003e \u003cp\u003eFor Rio Grande do Sul, the prevailing anomaly for each growing season was calculated as the mean of the quarterly ONI values, considering the SON, OND, NDJ, DJF, JFM, FMA, MAM, and AMJ periods. While for Goi\u0026aacute;s, the calculation was based on the OND, NDJ, DJF, JFM, FMA, and MAM quarters, thereby encompassing the complete crop development cycles in both regions. Based on these criteria, each growing season was classified according to the ENSO phase (El Ni\u0026ntilde;o, La Ni\u0026ntilde;a, or Neutral).\u003c/p\u003e \u003cp\u003eClimate data (Section 2.2) and ONI values were analyzed across two temporal periods: (a) 1961\u0026ndash;1990 and (b) 1991\u0026ndash;2019. This temporal segmentation is particularly relevant in the context of ENSO, as it enables the assessment of potential changes in the frequency and cyclic behavior of El Ni\u0026ntilde;o and La Ni\u0026ntilde;a events across different climatic periods (Li et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 \u003cspan type=\"SmallCaps\" class=\"SmallCaps\" name=\"Emphasis\"\u003eStatistical Analysis\u003c/span\u003e\u003c/h2\u003e \u003cp\u003eTo evaluate the influence of ENSO on climatic variables (rainfall and air temperature), Functional Data Analysis (FDA) was applied to both temporal periods (1961\u0026ndash;1990 and 1991\u0026ndash;2019) for the states of Goi\u0026aacute;s and Rio Grande do Sul. FDA enables the discrimination of ENSO impacts on rainfall and temperature, as well as the assessment of similarity patterns across ENSO phases (La Ni\u0026ntilde;a, El Ni\u0026ntilde;o, and Neutral) over time. In addition, this approach allows the analysis of groups of estimated functions, facilitating the evaluation of their statistical distribution and the estimation of mean functional curves for the climatic variables of interest.\u003c/p\u003e \u003cp\u003eIn recent years, FDA has been increasingly applied in meteorology (Beyaztas \u0026amp; Yaseen, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Curceac et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Matta et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), economics (Wang et al., \u003cspan citationid=\"CR105\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), geosciences (Bottazzi \u0026amp; Rossa, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; P\u0026eacute;rez-Plaza et al., \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), and agricultural sciences (Kwak et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Shah et al., \u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Xu et al., \u003cspan citationid=\"CR109\" class=\"CitationRef\"\u003e2018a\u003c/span\u003e, \u003cspan citationid=\"CR110\" class=\"CitationRef\"\u003e2018b\u003c/span\u003e, Justino et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2026\u003c/span\u003e), demonstrating its robustness for analyzing complex temporal patterns.\u003c/p\u003e \u003cp\u003eAll statistical analyses and graphical outputs were performed using R software (R Core Team, \u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e2.4.1 Classification of Climatic Variables based on ENSO Phases\u003c/h2\u003e \u003cp\u003eClimatic data for rainfall, maximum temperature, and minimum temperature, corresponding to the states of Rio Grande do Sul and Goi\u0026aacute;s, the analyzed periods ((a) and (b)), and ENSO phases, were initially processed using the fdata function from the fda.usc R package (Febrero-Bande \u0026amp; Oviedo de la Fuente, 2012) and subsequently converted into functional objects.\u003c/p\u003e \u003cp\u003eFollowing this step, the functional k-means clustering method was applied to each combination of state, period, and ENSO phase to estimate representative mean functional curves for each group.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003e2.4.2 Comparison of Mean Functional Curve Estimates\u003c/h2\u003e \u003cp\u003eAfter defining the representative mean functional curves for each ENSO phase, a Functional Analysis of Variance (FANOVA) was performed to assess whether the estimated functions (mean functional curves for El Ni\u0026ntilde;o, La Ni\u0026ntilde;a, and Neutral groups) differed statistically. The equality of functional means was tested using the fdANOVA package (G\u0026oacute;recki \u0026amp; Smaga, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe FP (F-type Permutation) test statistic, an extension of the classical ANOVA F-statistic for functional data, was adopted. This approach is particularly appropriate when variability is concentrated in a limited number of functional modes, thereby improving analytical efficiency while reducing model complexity. A significance level of 5% was assumed for all hypothesis tests.\u003c/p\u003e \u003cp\u003eThis analytical framework enabled the identification of climatic patterns associated with ENSO phases for absolute maximum temperature, absolute minimum temperature, and weekly accumulated rainfall. FANOVA was initially applied separately to each climatic variable within each period. When statistically significant differences were detected, \u003cem\u003epost hoc\u003c/em\u003e multiple comparison tests were conducted between pairs of ENSO phases (El Ni\u0026ntilde;o vs. Neutral; El Ni\u0026ntilde;o vs. La Ni\u0026ntilde;a; Neutral vs. La Ni\u0026ntilde;a).\u003c/p\u003e \u003cp\u003eTo control for Type I error, the Holm\u0026ndash;Bonferroni method was applied to sequentially adjust p-values (Holm, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e1979\u003c/span\u003e). It is important to emphasize that both the pairwise comparisons among ENSO phases and the subsequent analysis of climatic variable importance for crop yield (Section 2.4.3) were performed only when FANOVA indicated statistically significant differences among functional means.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e2.4.3 Yield Analysis\u003c/h2\u003e \u003cp\u003eYield data for irrigated rice, common bean, and soybean in the selected municipalities (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e) were obtained from the Municipal Agricultural Production database (PAM, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://sidra.ibge.gov.br/tabela/1612\u003c/span\u003e\u003cspan address=\"https://sidra.ibge.gov.br/tabela/1612\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) of the Brazilian Institute of Geography and Statistics (IBGE). These data represent the average yield for each crop (soybean, rice, and bean) during the growing seasons from September 1st to June 28th in Rio Grande do Sul and from October 1st to May 26th in Goi\u0026aacute;s.\u003c/p\u003e \u003cp\u003eHowever, yield records for all analyzed crops and both states are only available from 1974 onwards (IBGE, 2020). Therefore, for yield analyses, temporal period (a) was adjusted from 1961\u0026ndash;1990 to 1974\u0026ndash;1990 to ensure data completeness and historical consistency.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e2.4.4 Technological Trend in Crop Yield\u003c/h2\u003e \u003cp\u003eTo remove the technological trend from the yield data, the methodology proposed by Heinemann \u0026amp; Sentelhas (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) was applied. A \u003cem\u003eloess\u003c/em\u003e regression was used to estimate the temporal trend associated with technological advancements. The estimated technological component was subsequently subtracted from the original yield values, isolating the variability attributable to climatic factors.\u003c/p\u003e \u003cp\u003eThis procedure was performed independently for each municipality (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e) and for both temporal periods: (a) 1974\u0026ndash;1990 and (b) 1991\u0026ndash;2019. An example of this approach applied to soybean yield in Bag\u0026eacute;-RS during period (a) is presented in Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e. Consequently, the yield data were adjusted for technological evolution and projected into a current baseline scenario, preserving the climatic variability that constitutes the focus of this study.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e \u003ch2\u003e2.4.5 Crop Yield Prediction\u003c/h2\u003e \u003cp\u003eFollowing the detrending procedure, an unsupervised \u003cem\u003ek\u003c/em\u003e-means clustering algorithm was applied to classify crop yields into two categories: high and low (Figure S2). The minimum and maximum thresholds for each group are presented in Table S2. This classification was subsequently used as the response variable in a supervised \u003cem\u003eRandom Forest\u003c/em\u003e classification model, implemented using the randomForest R package (Liaw \u0026amp; Wiener, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2002\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cem\u003eRandom Forest\u003c/em\u003e is an ensemble learning method based on multiple decision trees and is widely applied to both regression and classification problems. The predictor variables consisted of the climatic variables derived from the analyses described in Section 2.4.2. In cases where multiple comparison tests indicated no significant differences among ENSO phases for a given variable and state, these categories were aggregated into a single group.\u003c/p\u003e \u003cp\u003eAdditionally, municipalities were included as a predictor variable to incorporate spatial information and improve model performance. This approach enabled the evaluation of the relative importance of climatic variables and geographic location in determining yield classification, providing a quantitative measure of each variable\u0026rsquo;s contribution to yield patterns.\u003c/p\u003e \u003cp\u003eFor each crop, the \u003cem\u003eRandom Forest\u003c/em\u003e model was fitted using node splitting based on the Gini index (Sharda et al., \u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Model accuracy was derived from the training dataset and evaluated separately for each crop and temporal period (Table S3).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section3\"\u003e \u003ch2\u003e2.4.6 Importance of Climatic Variables in Crop Yield\u003c/h2\u003e \u003cp\u003eFollowing model calibration, SHAP (SHapley Additive exPlanations) analysis was applied to interpret the \u003cem\u003eRandom Forest\u003c/em\u003e results and quantify the individual contribution of each predictor variable to yield classification. This approach is based on the concept of Shapley values (1953) from cooperative game theory and allows for the identification of both the magnitude and direction of variable effects (Siemers \u0026amp; Bajorath, \u003cspan citationid=\"CR87\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Zheng et al., \u003cspan citationid=\"CR114\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSHAP values represent the additive contribution of each explanatory variable, for each observation, relative to the baseline average prediction. Considering the binary response variable (1 for high yield and 0 for low yield), positive SHAP values indicate conditions that increase the probability of high yield, whereas negative values indicate factors associated with low-yield classification.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"3 RESULTS AND DISCUSSION","content":"\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.1 \u003cspan type=\"SmallCaps\" class=\"SmallCaps\" name=\"Emphasis\"\u003eFunctional ANOVA\u003c/span\u003e\u003c/h2\u003e \u003cp\u003eThe results of the Functional Analysis of Variance (FANOVA), incorporating the Holm\u0026ndash;Bonferroni correction for pairwise comparisons among ENSO phases in the states of Goi\u0026aacute;s and Rio Grande do Sul, for weekly accumulated rainfall, absolute maximum temperature, and absolute minimum air temperature, are presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eFunctional Analysis of Variance (FANOVA) for the comparison of ENSO phases (El Ni\u0026ntilde;o, La Ni\u0026ntilde;a, and Neutral) in Goi\u0026aacute;s and Rio Grande do Sul for climatic variables (accumulated rainfall, maximum air temperature, and minimum air temperature) across the periods: (a) 1961\u0026ndash;1990 and (b) 1991\u0026ndash;2019. Significance level: p\u0026thinsp;=\u0026thinsp;0.05. Hypotheses: H₀: El Ni\u0026ntilde;o\u0026thinsp;=\u0026thinsp;Neutral\u0026thinsp;=\u0026thinsp;La Ni\u0026ntilde;a; H₁: El Ni\u0026ntilde;o\u0026thinsp;\u0026ne;\u0026thinsp;Neutral\u0026thinsp;\u0026ne;\u0026thinsp;La Ni\u0026ntilde;a.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ePeriod\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003e\u003cem\u003eP-value\u003c/em\u003e (Holm-Bonferroni correction)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRio Grande do Sul\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGoi\u0026aacute;s\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eMaximum Temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(a) 1961\u0026ndash;1990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.683\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(b) 1991\u0026ndash;2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.273\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eMinimum Temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(a) 1961\u0026ndash;1990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.372\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(b) 1991\u0026ndash;2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.642\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eAccumulated rainfall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(a) 1961\u0026ndash;1990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.070\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(b) 1991\u0026ndash;2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eFor Rio Grande do Sul, the analysis revealed statistically significant differences (\u003cem\u003ep\u003c/em\u003e value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) among ENSO phases (El Ni\u0026ntilde;o, Neutral, and La Ni\u0026ntilde;a) for all evaluated climatic variables across both temporal periods ((a) 1961\u0026ndash;1990 and (b) 1991\u0026ndash;2019). Moreover, the magnitude of statistical significance increased in the more recent period (b), particularly for minimum temperature, indicating that differences among ENSO phases have become more evident in recent decades (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). This pattern suggests the presence of an interdecadal oscillation in ENSO amplitude, which modulates the intensity of its climatic teleconnections (Li et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). In contrast, for accumulated rainfall, the influence of ENSO remained consistent across both analyzed periods.\u003c/p\u003e \u003cp\u003eIn Goi\u0026aacute;s, however, FANOVA results indicated no statistically significant differences (\u003cem\u003ep\u003c/em\u003e value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) among ENSO phases for any of the evaluated climatic variables in either temporal period at the 5% significance level (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). These findings suggest that climate variability in Goi\u0026aacute;s is less sensitive to ENSO-related oscillations when compared to Rio Grande do Sul. This result contrasts with Matta et al. (\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), who, using a shorter climatic dataset period (1990\u0026ndash;2011), reported significant differences among ENSO phases in Goi\u0026aacute;s; however, that study also highlighted a reduction in statistical significance between periods 1990\u0026ndash;2000 and 2001\u0026ndash;2011.\u003c/p\u003e \u003cp\u003eGiven that the global analysis for Rio Grande do Sul rejected the null hypothesis of equality among ENSO phases, a pairwise comparison (El Ni\u0026ntilde;o vs. Neutral; El Ni\u0026ntilde;o vs. La Ni\u0026ntilde;a; Neutral vs. La Ni\u0026ntilde;a) was subsequently conducted. The results of these comparisons are presented in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eFunctional Analysis of Variance (FANOVA) for pairwise comparisons of ENSO phases (El Ni\u0026ntilde;o, La Ni\u0026ntilde;a, and Neutral) in the state of Rio Grande do Sul for climatic variables (accumulated rainfall, maximum air temperature, and minimum air temperature) across the periods: (a) 1961\u0026ndash;1990 and (b) 1991\u0026ndash;2019. Significance level: p\u0026thinsp;=\u0026thinsp;0.05. Hypotheses: (1) El Ni\u0026ntilde;o\u0026thinsp;=\u0026thinsp;Neutral; (2) El Ni\u0026ntilde;o\u0026thinsp;=\u0026thinsp;La Ni\u0026ntilde;a; (3) La Ni\u0026ntilde;a\u0026thinsp;=\u0026thinsp;Neutral.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ePeriod\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eH\u003csub\u003e0\u003c/sub\u003e: El Ni\u0026ntilde;o\u0026thinsp;=\u0026thinsp;Neutral\u003c/p\u003e \u003cp\u003eH\u003csub\u003e1\u003c/sub\u003e: El Ni\u0026ntilde;o\u0026thinsp;\u0026ne;\u0026thinsp;Neutral\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eH\u003csub\u003e0\u003c/sub\u003e: El Ni\u0026ntilde;o\u0026thinsp;=\u0026thinsp;La Ni\u0026ntilde;a\u003c/p\u003e \u003cp\u003eH\u003csub\u003e1\u003c/sub\u003e: El Ni\u0026ntilde;o\u0026thinsp;\u0026ne;\u0026thinsp;La Ni\u0026ntilde;a\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eH\u003csub\u003e0\u003c/sub\u003e: La Ni\u0026ntilde;a\u0026thinsp;=\u0026thinsp;Neutral\u003c/p\u003e \u003cp\u003eH\u003csub\u003e1\u003c/sub\u003e: La Ni\u0026ntilde;a\u0026thinsp;\u0026ne;\u0026thinsp;Neutral\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c5\" namest=\"c3\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eMaximum Air Temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(a) 1961\u0026ndash;1990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.081\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(b) 1991\u0026ndash;2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.260\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eMinimum Air Temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(a) 1961\u0026ndash;1990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.133\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.009\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(b) 1991\u0026ndash;2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eAccumulated Rainfall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(a) 1961\u0026ndash;1990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(b) 1991\u0026ndash;2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eFor maximum air temperature, statistically significant differences were identified only when comparing the Neutral phase with El Ni\u0026ntilde;o and with La Ni\u0026ntilde;a in both temporal periods. No significant differences were observed between El Ni\u0026ntilde;o and La Ni\u0026ntilde;a, indicating that these two phases exert similar effects on maximum temperature in the study region. This finding is consistent with previous studies indicating that the Neutral phase is characterized by atmospheric and oceanic conditions close to the climatological mean, with less pronounced anomalies compared to El Ni\u0026ntilde;o and La Ni\u0026ntilde;a events (Lin \u0026amp; Qian, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Consequently, during Neutral conditions, the effects on climatic variables such as maximum temperature tend to be less pronounced (Feng \u0026amp; Hao, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor minimum air temperature, the results revealed a more complex and period-dependent behavior. During the 1961\u0026ndash;1990 period (period a), statistically significant differences were primarily associated with the La Ni\u0026ntilde;a phase, which differed from both El Ni\u0026ntilde;o and Neutral conditions. In contrast, in the more recent period (b) (1991\u0026ndash;2019), all pairwise comparisons were statistically significant, indicating that each ENSO phase exerts distinct influences on minimum temperature.\u003c/p\u003e \u003cp\u003eAlthough the influence of ENSO on climatic variables is generally more evident at continental scales than at the regional scale considered in this study, a consistent pattern was observed for accumulated rainfall across all periods, with statistically significant differences among all ENSO phases. This result confirms that ENSO acts as a major modulator of the regional rainfall regime, although it is not the sole controlling factor (Souza et al., \u003cspan citationid=\"CR95\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In this context, El Ni\u0026ntilde;o events are typically associated with increased rainfall in Rio Grande do Sul, whereas La Ni\u0026ntilde;a events are frequently linked to rainfall deficits (Arruda et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), with Neutral conditions representing an intermediate state (Lenka et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTamaddun et al. (\u003cspan citationid=\"CR97\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) evaluated the influence of ENSO phases on changes in temperature, precipitation, and potential evapotranspiration in India and reported that El Ni\u0026ntilde;o years exert a stronger influence on these variables compared to La Ni\u0026ntilde;a and Neutral years. In contrast, Sattar et al. (\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) observed higher maximum temperatures and reduced rainfall during both El Ni\u0026ntilde;o and La Ni\u0026ntilde;a events relative to Neutral conditions in Bangladesh. These findings highlight that ENSO impacts are region-specific and depend on both the climatic variable considered and the phase of the phenomenon, being more pronounced in certain regions and under specific conditions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.2 \u003cspan type=\"SmallCaps\" class=\"SmallCaps\" name=\"Emphasis\"\u003eClimatic Patterns across ENSO Phases\u003c/span\u003e\u003c/h2\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e presents the mean functional estimates for accumulated rainfall across the three ENSO phases for the states of Rio Grande do Sul and Goi\u0026aacute;s, including confidence intervals for the mean curves over weeks 1 to 34 (Goi\u0026aacute;s) and 1 to 43 (Rio Grande do Sul), for both analyzed periods.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn Goi\u0026aacute;s, during the 1961\u0026ndash;1990 period (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA), the accumulation of rainfall exhibited a relatively constant increment up to approximately week 28 (April), reaching an accumulated volume of approximately 1400 mm. This interval was characterized by smooth and nearly linear growth, indicating high regularity in rainfall distribution and reduced interannual variability. From April onwards, a deceleration in accumulated rainfall and an increase in variability were observed. Moreover, the highest rainfall volumes occurred during El Ni\u0026ntilde;o years, whereas Neutral conditions remained close to the overall mean of the three phases, and La Ni\u0026ntilde;a years exhibited lower accumulated rainfall.\u003c/p\u003e \u003cp\u003eIn the more recent period (1991\u0026ndash;2019; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB), a similar stabilization of the mean rainfall curve after week 28 was observed in Goi\u0026aacute;s. However, this pattern was accompanied by a reduction of more than 100 mm in total accumulated rainfall, along with increased interannual variability in the final weeks. These changes in rainfall dynamics may be associated with climate change processes, which tend to intensify climate extremes and increase rainfall irregularity (Furtak \u0026amp; Wolińska, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In contrast to period (a), during 1991\u0026ndash;2019, the highest rainfall volumes were recorded during La Ni\u0026ntilde;a years, followed by Neutral conditions, whereas El Ni\u0026ntilde;o years presented the lowest accumulated rainfall. It is important to note that changes in ENSO phase frequency contributed to this pattern: period (a) was dominated by El Ni\u0026ntilde;o events (five occurrences) relative to La Ni\u0026ntilde;a (four), whereas period (b) showed a higher frequency of La Ni\u0026ntilde;a events (six) compared to El Ni\u0026ntilde;o (four), promoting increased rainfall in some years and drought conditions in others, thereby increasing interannual variability (Cai et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn Rio Grande do Sul, accumulated rainfall during El Ni\u0026ntilde;o years in the 1961\u0026ndash;1990 period (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC) was markedly higher than that observed during La Ni\u0026ntilde;a and Neutral years. By week 43, El Ni\u0026ntilde;o years reached approximately 1700 mm, while Neutral and La Ni\u0026ntilde;a years accumulated around 1300 mm and 1350 mm, respectively. The divergence among the curves begins around week 7 (October), coinciding with the onset of higher rainfall volumes in the state (INMET, 2020). Under El Ni\u0026ntilde;o conditions, Rio Grande do Sul typically experiences above-average precipitation, whereas La Ni\u0026ntilde;a years are associated with rainfall deficits and frequent droughts (Arruda et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). A similar pattern was reported by Teegavarapu \u0026amp; Sharma (\u003cspan citationid=\"CR99\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), who observed higher precipitation during El Ni\u0026ntilde;o years, followed by Neutral and La Ni\u0026ntilde;a conditions, in Florida, USA.\u003c/p\u003e \u003cp\u003eIn contrast, during the more recent period (1991\u0026ndash;2019; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD), the differences among ENSO phases were less pronounced, although still statistically significant (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Rainfall patterns during El Ni\u0026ntilde;o and La Ni\u0026ntilde;a years closely followed those observed under Neutral conditions, suggesting a potential reduction in the intensity or variability of ENSO effects on regional climate over time (Cai et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Matta et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). According to Li et al. (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), ENSO intensity has fluctuated in recent decades, increasing after the mid-1970s but declining during the 21st century. Similarly, An et al. (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) argue that the historically strong relationship between ENSO and rainfall has weakened in recent decades.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e presents the ratio between the mean functional estimates of accumulated rainfall for different pairs of ENSO phases in Rio Grande do Sul. This analysis provides insights into how ENSO phases modulate accumulated rainfall over time, quantifying the relative increase or decrease in precipitation between phases.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFor the 1961\u0026ndash;1990 period (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA), a pronounced positive peak is observed in the El Ni\u0026ntilde;o/La Ni\u0026ntilde;a ratio, indicating a strong predominance of rainfall during El Ni\u0026ntilde;o in the initial weeks. In contrast, the La Ni\u0026ntilde;a/Neutral ratio exhibits a negative peak, reflecting reduced rainfall during La Ni\u0026ntilde;a in the same interval. The El Ni\u0026ntilde;o/Neutral ratio remained relatively stable, without pronounced oscillations, but consistently ranged between 1.3 and 1.5, indicating higher rainfall during El Ni\u0026ntilde;o throughout most of the period. From week 15 onwards, all curves showed minimal variation, with La Ni\u0026ntilde;a and Neutral phases converging to similar values (ratio close to 1), while El Ni\u0026ntilde;o maintained higher rainfall levels. These results corroborate the patterns observed in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC, which also indicate higher precipitation during El Ni\u0026ntilde;o and similar behavior between La Ni\u0026ntilde;a and Neutral phases.\u003c/p\u003e \u003cp\u003eIn period (b) (1991\u0026ndash;2019; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB), the ratios among ENSO phases remained close to 1 throughout the weeks, without the pronounced peaks observed in the earlier period. This indicates the absence of a clear dominance of any ENSO phase, reinforcing the pattern shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD, where rainfall differences among phases were minimal.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB present the mean functional estimates for maximum air temperature across ENSO phases over weeks 1 to 34 in Goi\u0026aacute;s for the two analyzed periods ((a) 1961\u0026ndash;1990 and (b) 1991\u0026ndash;2019). In both periods, maximum temperature exhibited a gradual decrease over time, with higher values at the beginning of the cycle (October) and lower values toward the end of the period (May).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDuring period (a) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA), maximum temperatures in Goi\u0026aacute;s started at approximately 34\u0026deg;C and gradually declined to around 30.6\u0026deg;C by the final week. In period (b) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB), slightly higher initial values were observed, close to 36\u0026deg;C, with final values around 31.5\u0026deg;C. This increase in temperature suggests a warming trend over recent decades, potentially associated with global climate change (Regoto et al., \u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Hofmann et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Additionally, greater variability was observed between El Ni\u0026ntilde;o and La Ni\u0026ntilde;a curves, particularly up to approximately week 24 (March).\u003c/p\u003e \u003cp\u003eA similar decreasing pattern was observed for minimum air temperature in the final weeks, especially after April (week 28) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). This behavior is associated with the seasonal transition from autumn to winter, characterized by lower temperatures, shorter photoperiods, and reduced solar radiation (Justino et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn Rio Grande do Sul, maximum air temperature (Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD) exhibited marked seasonal variation, with peak values occurring between mid-January and February (weeks 18 and 23), corresponding to the austral summer, regardless of ENSO phase. This period is also associated with increased solar radiation and lower relative humidity (INMET, 2020).\u003c/p\u003e \u003cp\u003eDuring period (a) (1961\u0026ndash;1990; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC), maximum temperatures in La Ni\u0026ntilde;a years were generally lower than those observed during Neutral conditions. A similar pattern was observed during El Ni\u0026ntilde;o years, although with greater variability, with temperatures fluctuating both above and below those recorded under Neutral conditions.\u003c/p\u003e \u003cp\u003eIn the more recent period (1991\u0026ndash;2019; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD), maximum temperatures during El Ni\u0026ntilde;o years were frequently higher than those observed under Neutral conditions, particularly after week 25 (February). La Ni\u0026ntilde;a years exhibited greater variability but generally remained below Neutral values. Silva et al. (\u003cspan citationid=\"CR88\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) reported increased air temperatures in Brazil during El Ni\u0026ntilde;o years, whereas Neutral and La Ni\u0026ntilde;a periods presented more similar conditions. Similarly, Sattar et al. (\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) observed higher maximum temperatures during both El Ni\u0026ntilde;o and La Ni\u0026ntilde;a years compared to Neutral conditions in Bangladesh, with more pronounced differences during El Ni\u0026ntilde;o events.\u003c/p\u003e \u003cp\u003eThe ratio analysis of maximum air temperature between ENSO phases in Rio Grande do Sul (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA) indicates that, for most weeks, values remained close to 1, suggesting limited differences among phases. However, greater variability was observed in the initial and final weeks, highlighting the predominance of El Ni\u0026ntilde;o over La Ni\u0026ntilde;a (ratio\u0026thinsp;\u0026gt;\u0026thinsp;1) and lower values for La Ni\u0026ntilde;a relative to Neutral conditions (ratio\u0026thinsp;\u0026lt;\u0026thinsp;1).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSimilarly, in period (b) (1991\u0026ndash;2019; Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB), the ratio curves remained clustered around 1, indicating subtle differences among ENSO phases. The El Ni\u0026ntilde;o/La Ni\u0026ntilde;a and El Ni\u0026ntilde;o/Neutral ratios were generally slightly above 1, indicating marginally higher maximum temperatures during El Ni\u0026ntilde;o. Conversely, the La Ni\u0026ntilde;a/Neutral ratio was below 1, indicating lower maximum temperatures during La Ni\u0026ntilde;a. These results are consistent with the patterns observed in Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD.\u003c/p\u003e \u003cp\u003eOverall, minimum air temperature curves in Rio Grande do Sul (Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD) followed a seasonal pattern similar to that observed for maximum temperature, with an increase until approximately week 25 (summer), followed by a decline until week 43 (early winter).\u003c/p\u003e \u003cp\u003eDuring period (a) (1961\u0026ndash;1990; Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC), minimum temperatures were generally lower during La Ni\u0026ntilde;a years, except between weeks 18 and 24. In period (b) (1991\u0026ndash;2019; Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD), El Ni\u0026ntilde;o years exhibited higher minimum temperatures during most weeks, exceeding those observed under Neutral conditions, particularly between weeks 10\u0026ndash;19 and after week 25.\u003c/p\u003e \u003cp\u003eThe ratio analysis for minimum air temperature (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC and \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD) revealed greater variability compared to maximum temperature. The El Ni\u0026ntilde;o/La Ni\u0026ntilde;a ratio showed pronounced positive peaks in both periods, indicating higher minimum temperatures during El Ni\u0026ntilde;o events. Conversely, the La Ni\u0026ntilde;a/Neutral ratio exhibited negative peaks in certain weeks, indicating lower minimum temperatures during La Ni\u0026ntilde;a relative to Neutral conditions. These results suggest that differences in minimum temperature among ENSO phases are more irregular and tend to intensify in the final weeks of the cycle, indicating a less stable ENSO influence compared to that observed for maximum temperature.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.3 \u003cspan type=\"SmallCaps\" class=\"SmallCaps\" name=\"Emphasis\"\u003eCLIMATE VARIABLE AND YIELD\u003c/span\u003e\u003c/h2\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e presents the mean absolute predictive importance of climatic variables and municipality for the yields of lowland rice, common bean, and soybean in the state of Rio Grande do Sul during periods (a) 1974\u0026ndash;1990 and (b) 1991\u0026ndash;2019. This analysis was not performed for the state of Goi\u0026aacute;s because, as described in Section 3.1 (Functional ANOVA), the FANOVA results for that state indicated no significant differences (\u003cem\u003ep\u003c/em\u003e value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) among ENSO phases (El Ni\u0026ntilde;o, La Ni\u0026ntilde;a, and Neutral) for any of the evaluated climatic variables in either analyzed period.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eA higher mean absolute SHAP value indicates a greater influence of a given variable on model output. For lowland rice during period (a) (1974\u0026ndash;1990; Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA), municipality showed the greatest predictive importance (80%) for yield, followed by accumulated rainfall (10.9%), whereas minimum (5.4%) and maximum (3.6%) air temperatures exhibited comparatively lower influence. When the specific effect of each variable on rice yield is examined (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA), accumulated rainfall during El Ni\u0026ntilde;o years (red) contributed mainly to yield increases (positive values), whereas La Ni\u0026ntilde;a (blue) and Neutral (gray) conditions were predominantly associated with yield reductions. For minimum temperature, the La Ni\u0026ntilde;a phase (blue) exerted a negative effect, while El Ni\u0026ntilde;o and Neutral phases (yellow), which did not differ significantly for this variable (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) and were therefore evaluated jointly, tended to increase yield or produce no relevant effect (values close to zero). For maximum temperature, the Neutral phase showed a smaller effect on yield (gray), whereas El Ni\u0026ntilde;o and La Ni\u0026ntilde;a (purple), also evaluated jointly because of the absence of statistical differences, were associated with greater yield variation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn period (b) (1991\u0026ndash;2019; Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB), a substantial shift in the relative importance of the explanatory variables for rice yield was observed. Municipality remained the dominant predictor (80.3%), followed by minimum temperature (7.1%) and maximum temperature (6.8%). In contrast, rainfall, which ranked as the second most important variable in the earlier period, exhibited the lowest relevance in yield prediction (5.8%). During this period, El Ni\u0026ntilde;o (red) and La Ni\u0026ntilde;a (blue), whether evaluated individually or jointly (purple), tended to exert positive effects on rice yield regardless of the climatic variable analyzed. Conversely, the Neutral phase (gray) was associated with yield reductions (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003eAlthough the data used in this study refer to lowland rice, which in principle is less directly affected by rainfall variability, the substantial predictive relevance of rainfall, particularly in the first period, may be explained by its indirect effects on the production system and on other environmental variables, such as air and soil temperature, relative humidity, and solar radiation, all of which influence crop development and the energy balance. For example, high humidity and prolonged leaf wetness may favor the incidence of fungal diseases such as rice blast (\u003cem\u003eMagnaporthe oryzae\u003c/em\u003e) and brown spot (\u003cem\u003eBipolaris oryzae\u003c/em\u003e), thereby compromising grain filling and final yield (Kirtphaiboon et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Percich et al., \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e1997\u003c/span\u003e; Bhandari et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In addition, rainfall occurring at undesirable times, such as during physiological maturity and harvest, when irrigation should be suspended, may compromise both yield and grain quality (Sohn et al., \u003cspan citationid=\"CR91\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe high explanatory capacity of municipality in both periods is associated with edaphic, topographic, and management-related factors that vary spatially among production regions. Soil texture, fertility, water-holding capacity, and salinity directly affect nutrient availability and crop development (Ye et al., \u003cspan citationid=\"CR112\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Omar et al., \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Topography is another critical factor, as flatter terrain facilitates irrigation management and promotes greater homogeneity in microclimatic conditions (Fan et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Furthermore, heterogeneity in technological level among rice-producing municipalities affects phytosanitary management, water and fertilizer use efficiency, and the quality of field operations, resulting in substantial differences in yield (Fan \u0026amp; Gan, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe greater predictive importance of maximum and minimum air temperatures in the more recent period (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB) indicates the growing impact of extreme thermal conditions associated with climate change, particularly during critical stages of rice development. In general, high temperatures reduce photosynthetic efficiency and increase respiration and transpiration, thereby compromising vegetative growth and nutrient uptake (Xiong et al., \u003cspan citationid=\"CR108\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Lu et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). During flowering, elevated temperatures reduce the number of spikelets per panicle and increase pollen sterility, thereby decreasing the formation of viable seeds (Yu et al., \u003cspan citationid=\"CR113\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). During grain filling, excessive heat leads to the development of smaller and less dense grains with lower commercial quality (Lu et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Song et al. (\u003cspan citationid=\"CR93\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) reported an 8% reduction in rice yield for each 1\u0026deg;C increase in mean air temperature. When this increase occurred during the vegetative stage, the effect on yield was less severe. Xiong et al. (\u003cspan citationid=\"CR108\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) further observed that the physiological processes underlying rice yield formation are differentially affected by high daytime and nighttime temperatures. This distinction helps explain the marked difference in the predictive importance of maximum and minimum temperatures, which play distinct roles across crop developmental stages.\u003c/p\u003e \u003cp\u003eFor common bean during period (a) (1974\u0026ndash;1990; Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC), SHAP importance analysis indicated that the variables with the greatest influence on yield were municipality (85.2%), followed by accumulated rainfall (7.5%), minimum temperature (5.9%), and maximum air temperature (1.4%). As observed for rice, the high importance attributed to municipality may reflect edaphoclimatic, topographic, and management-related differences among production regions, which directly affect crop performance.\u003c/p\u003e \u003cp\u003eWhen the individual effects of the climatic variables on common bean yield are considered (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC) for period (a) (1974\u0026ndash;1990), a pattern similar to that observed for rice is evident. With respect to accumulated rainfall, the El Ni\u0026ntilde;o phase (red) contributed to yield increases (positive values), whereas La Ni\u0026ntilde;a (blue) and Neutral (gray) conditions exerted predominantly negative effects. For minimum temperature, the La Ni\u0026ntilde;a phase (blue) showed a negative effect, whereas El Ni\u0026ntilde;o and Neutral phases (yellow) positively influenced yield. In contrast, maximum temperature exhibited a comparatively smaller effect on yield, particularly under Neutral ENSO conditions (gray).\u003c/p\u003e \u003cp\u003eAmong the climatic variables, accumulated rainfall showed substantial predictive importance, indicating a high sensitivity of common bean yield to water availability. The water requirement of common bean ranges from approximately 350 mm during the vegetative stage to 200\u0026ndash;350 mm during the reproductive stage (Heinemann et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). In this context, the most critical stage for drought stress extends from pre-flowering to the end of grain filling (Soureshjani et al., \u003cspan citationid=\"CR94\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Androcioli et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). During flowering, water deficit promotes flower abscission, whereas during grain formation it causes abortion of fertilized ovules within pods (Mathobo et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In common bean, stress during the reproductive stage causes, among other effects, yield reduction (Rosales-Serna et al., \u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; Labastida et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), attenuation of photosynthetic performance (Mladenov et al., \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Papathanasiou et al., \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Polania et al., \u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), shortening of the crop cycle (Beebe et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Labastida et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), and reductions in pod and seed formation (Labastida et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Gon\u0026ccedil;alves et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMaximum and minimum air temperatures also contributed to the prediction of common bean yield, although to a lesser extent (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC). A similar pattern was observed in the more recent period (b) (1991\u0026ndash;2019; Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD), in which the importance of municipality became even more pronounced relative to the first period, accounting for 90.6% of the predictive importance. Rainfall remained the second most important variable (3.7%), followed by minimum (3.3%) and maximum (2.4%) air temperatures.\u003c/p\u003e \u003cp\u003eIn period (b) (1991\u0026ndash;2019), accumulated rainfall and minimum air temperature exerted more positive effects on common bean yield during El Ni\u0026ntilde;o years (red) than during La Ni\u0026ntilde;a years (blue) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eD). The influence of these variables under Neutral conditions was less pronounced (gray). In contrast, for maximum temperature, the Neutral phase (gray) showed a smaller magnitude of effect on yield, whereas the joint evaluation of El Ni\u0026ntilde;o and La Ni\u0026ntilde;a (purple) resulted in a predominantly negative impact.\u003c/p\u003e \u003cp\u003eRecent decades have been characterized by rising air temperatures, a greater frequency of extreme events, and increasing irregularity and intensification of ENSO phenomena (Yang et al., \u003cspan citationid=\"CR111\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). These factors may have contributed to the increased importance of maximum temperature in yield prediction between the first and the most recent periods. Whereas four La Ni\u0026ntilde;a events and two El Ni\u0026ntilde;o events occurred between 1974 and 1990, the more recent period (1991\u0026ndash;2019) included six La Ni\u0026ntilde;a and four El Ni\u0026ntilde;o events. This higher frequency of ENSO events contributes to increased thermal variability in Rio Grande do Sul and may result in substantial yield variation. During La Ni\u0026ntilde;a episodes, which were the more frequent events in the recent period, elevated temperatures at certain crop stages combined with reduced rainfall (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD, \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD, and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD) in southern Brazil favor both water and heat stress in plants (Lopes et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor adequate common bean development, the minimum, optimum, and maximum air temperatures are 12\u0026deg;C, 21\u0026deg;C, and 29\u0026deg;C, respectively (Vieira et al., \u003cspan citationid=\"CR103\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). Excessively high temperatures can cause physical, physiological, and biochemical damage to plants, including alterations in protein structure and synthesis, changes in enzymatic activity, injury to membranes and the photosynthetic apparatus, and oxidative stress (Fahad et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Taiz et al., \u003cspan citationid=\"CR96\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Near flowering, high temperatures can reduce pollen production and viability (Porch \u0026amp; Jahn, \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2001\u003c/span\u003e; Prasad et al., \u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e2002\u003c/span\u003e), promote flower and pod abscission (Ofir et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e1993\u003c/span\u003e), impair seed formation (Vargas et al., \u003cspan citationid=\"CR101\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and reduce yield (Soltani et al., \u003cspan citationid=\"CR92\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Heinemann et al. (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2025\u003c/span\u003e) also demonstrated that common bean is particularly vulnerable during the reproductive stage as minimum temperatures increase.\u003c/p\u003e \u003cp\u003eFor soybean, the predictive importance analysis for yield in Rio Grande do Sul during 1974\u0026ndash;1990 (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eE) indicated a predominance of municipality, accounting for 77.7% of the mean absolute importance, followed by accumulated rainfall (15.8%), maximum temperature (5.3%), and minimum air temperature (1.3%). As observed for rice and common bean (Figs.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA and \u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC), the high contribution of municipality reflects the combined influence of edaphoclimatic, topographic, and management-related factors that vary spatially and condition crop yield.\u003c/p\u003e \u003cp\u003eDuring this period, changes in minimum and maximum air temperatures associated with ENSO phases produced effects of comparatively lower magnitude on soybean yield. In contrast, more substantial effects were observed for accumulated rainfall, particularly during El Ni\u0026ntilde;o years (red), when its influence was positive (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eE). Soybean is highly sensitive to water stress, especially during flowering and grain filling, stages during which water deficiency may alter enzymatic and hormonal activity (Zhou et al., \u003cspan citationid=\"CR115\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), reduce photosynthetic carbon assimilation (Du et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2020a\u003c/span\u003e; Tavares et al., \u003cspan citationid=\"CR98\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), cause tissue damage and oxidative stress (Du et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2020b\u003c/span\u003e), impair nutrient assimilation (Kaya et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), reduce biomass accumulation and seed number and weight (Poudel et al., \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), and ultimately compromise yield (Singh et al., \u003cspan citationid=\"CR90\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn the more recent period ((b) 1991\u0026ndash;2019), a substantial increase in the predictive importance of municipality (89%) for soybean yield was observed, accompanied by reductions in the importance of accumulated rainfall (4.2%) and maximum air temperature (3.1%) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eF). This shift relative to the previous period (1974\u0026ndash;1990) suggests that local factors have become even more dominant in determining crop yield, which may reflect technological advances, including the adoption of more productive and better-adapted cultivars, as well as improvements in crop management practices. It is important to emphasize that the adoption of such practices reduces crop vulnerability to climatic variability and increases production resilience (Raza et al., \u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Nevertheless, the analysis indicates that, although technological progress and management have gained importance over the decades, the effect of ENSO on water availability remains a key factor for understanding soybean yield variability in the region.\u003c/p\u003e \u003cp\u003eThe marked reduction in the predictive importance of accumulated rainfall between periods (a) (1974\u0026ndash;1990) and (b) (1991\u0026ndash;2019), observed across all analyzed crops (rice, common bean, and soybean), is consistent with the climatic analysis presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. That analysis showed a narrowing of the rainfall gap among ENSO phases, indicating a more homogeneous precipitation regime in recent decades. During 1974\u0026ndash;1990, El Ni\u0026ntilde;o events were associated with substantially greater precipitation than the other phases (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC), thereby contributing to greater yield variability linked to water availability. In contrast, during 1991\u0026ndash;2019 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD), although differences among phases remained statistically significant (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), total precipitation amounts were more similar, resulting in a reduction in the relative importance of rainfall for yield prediction.\u003c/p\u003e \u003cp\u003eConsistent with the pattern observed for 1974\u0026ndash;1990, in the more recent period ((b) 1991\u0026ndash;2019) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eF), accumulated rainfall during El Ni\u0026ntilde;o years (red) favored increases in soybean yield. The most negative effects across all variables were observed during La Ni\u0026ntilde;a years (blue), whereas the least pronounced effects occurred under Neutral conditions (gray) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eF).\u003c/p\u003e \u003cp\u003eVon Bloh et al. (\u003cspan citationid=\"CR104\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) reported that soybean yield estimates in the Southern and Northeastern regions of Brazil showed the poorest predictive performance among all regions of the country. This result was attributed to the high interannual yield variability characteristic of these areas, which are strongly influenced by ENSO. Such climatic instability increases uncertainty in predictive models and hinders the identification of consistent yield patterns, particularly in the more recent period, which was characterized by a higher frequency of both La Ni\u0026ntilde;a and El Ni\u0026ntilde;o events.\u003c/p\u003e \u003c/div\u003e"},{"header":"4 CONCLUSION","content":"\u003cp\u003eENSO exerts a strong influence on climatic variability and, consequently, on the productivity of lowland rice, common bean, and soybean, particularly in Rio Grande do Sul, where accumulated rainfall and maximum and minimum air temperatures differ significantly among El Ni\u0026ntilde;o, La Ni\u0026ntilde;a, and Neutral phases. In the state of Goi\u0026aacute;s, by contrast, the divergence among ENSO phases is less evident. Overall, the El Ni\u0026ntilde;o phase tended to favor yield increases across all evaluated crops (lowland rice, common bean, and soybean), whereas the La Ni\u0026ntilde;a phase was frequently associated with yield reductions. Under Neutral conditions, the influence of ENSO was less pronounced.\u003c/p\u003e \u003cp\u003eIn period (a) (1974\u0026ndash;1990), the El Ni\u0026ntilde;o phase did not differ statistically from the Neutral phase with respect to minimum temperature, nor from the La Ni\u0026ntilde;a phase with respect to maximum temperature. In period (b) (1991\u0026ndash;2019), El Ni\u0026ntilde;o again showed no statistically significant difference from La Ni\u0026ntilde;a for maximum temperature. These findings indicate an overlap in thermal patterns between El Ni\u0026ntilde;o and other ENSO phases.\u003c/p\u003e \u003cp\u003eFor all evaluated crops, a reduction in the predictive importance of accumulated rainfall was observed in the more recent period (1991\u0026ndash;2019), reflecting a more homogeneous precipitation pattern among ENSO phases. In contrast, local factors, including soil characteristics, topography, technological level, and management practices, assumed an even greater role in determining yield. Despite this shift, ENSO remains a key factor in explaining yield variability, reinforcing the need for agricultural planning and management strategies that are adapted to variable climatic conditions.\u003c/p\u003e "},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAUTHORS\u0026apos; CONTRIBUTION\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization (DHM; ABH) Data Curation (ABH; DCV); Formal Analysis (MVCS; DCV); Investigation (MVCS; DHM; ABH; LFJ); Methodology (MVCS; DHM; ABH; LFJ); Project Administration (ABH)); Resources (ABH); Software (MVCS; LFJ); Supervision (DHM; ABH); Validation (SVC; LFS); Visualization (MVCS); Writing \u0026ndash; Original Draft (DHM; ABH; LFS); Writing (SVQ; DHM; ABH; LFS; DVC); Review \u0026amp; Editing (DHM; ABH; LFS; SVC; DVC).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eACKNOWLEDGMENTS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAB Heinemann acknowledges support from the \u0026ldquo;Conselho Nacional de Desenvolvimento Cient\u0026iacute;fico e Tecnol\u0026oacute;gico\u0026rdquo; (CNPq N\u0026ordm; 4/2021 - 310209/2021-8).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAlvares CA, Stape JL, Sentelhas PC, Gon\u0026ccedil;alves JLM, Sparovek G (2013) Koppen\u0026rsquo;s climate classification map for Brazil. Meteorol Z 22:711\u0026ndash;728. \u003cu\u003ehttps://doi.org/10.1127/0941-2948/2013/0507\u003c/u\u003e\u003c/li\u003e\n\u003cli\u003eAn D, Eggeling J, Zhang L, He H, Sapkota A, Wang Y, Gao C (2023) Extreme precipitation patterns in the Asia\u0026ndash;Pacific region and its correlation with El Ni\u0026ntilde;o-Southern Oscillation (ENSO). Scientific Reports 13: 1-12. https://doi.org/10.1038/s41598-023-38317-0\u003c/li\u003e\n\u003cli\u003eAndrocioli LG, Zeffa DM, Alves DS, Tomaz JP, Moda-Cirino V (2020) Effect of water deficit on morphoagronomic and physiological traits of common bean genotypes with contrasting drought tolerance. Water 12:1\u0026ndash;13. https://doi.org/10.3390/w12010217\u003c/li\u003e\n\u003cli\u003eArruda AM, Centeno LN, Nunes AB (2025) Relation between major climatic indices and subseasonal precipitation in Rio Grande do Sul State, Brazil. Meteorology 4:1\u0026ndash;26. \u003cu\u003ehttps://doi.org/10.3390/meteorology4010005\u003c/u\u003e\u003c/li\u003e\n\u003cli\u003eBeebe SE, Rao IM, Blair MB, Acosta-Gallegos JA (2013) Phenotyping common beans for adaptation to drought. Front Physiol 4:1\u0026ndash;20.https://doi.org/10.3389/fphys.2013.00035\u003c/li\u003e\n\u003cli\u003eBeyaztas U, Yaseen ZM (2019) Drought interval simulation using functional data analysis. J Hydrol 579:1\u0026ndash;13. https://doi.org/10.1016/j.jhydrol.2019.124141\u003c/li\u003e\n\u003cli\u003eBhandari R, Khatri S, Acharya B, Gauchan U, Adhikari G, Thapa S (2024) Screening of rice genotypes against brown leaf spot (\u003cem\u003eBipolaris oryzae\u003c/em\u003e) (Breda de Haan) shoemaker, under field and laboratory conditions at Banke, Nepal. Discover Plants 1:1\u0026ndash;12. https://doi.org/10.1007/s44372-024-00008-z\u003c/li\u003e\n\u003cli\u003eBottazzi F, Rossa ED (2017) A Functional Data Analysis Approach to Surrogate Modeling in Reservoir and Geomechanics Uncertainty Quantification. Math Geosci 49:517\u0026ndash;540. https://doi.org/10.1007/s11004-017-9685-y\u003c/li\u003e\n\u003cli\u003eCai W, Ng B, Geng T, Wu L, Santoso A, McPhaden MJ (2020) Butterfly effect and a self-modulating El Ni\u0026ntilde;o response to global warming. Nature 585:68\u0026ndash;73. https://doi.org/10.1038/s41586-020-2641-x\u003c/li\u003e\n\u003cli\u003eCai W, Santoso A, Collins M, Dewitte B, Karamperidou C, Kug J, Lengaigne M, McPhaden MJ, Stuecker MF, Taschetto AS, Timmermann A, Wu L, Yeh S, Wang G, Ng B, Jia F, Yang Y, Ying J, Zheng X, Bayr T, Brown JR, Capotondi A, Cobb KM, Gan B, Geng T, Ham Y, Jin F, Jo H, Li X, Lin X, McGregor S, Park J, Stein K, Yang K, Zhang L, Zhong W (2021) Changing El Ni\u0026ntilde;o\u0026ndash;Southern Oscillation in a warming climate. Nat Rev Earth Environ 2:628\u0026ndash;644. https://doi.org/10.1038/s43017-021-00199-z\u003c/li\u003e\n\u003cli\u003eCao J, Zhang S, Tao F, Chen Y, Luo X, Xie J (2023) Forecasting global crop yields based on El Nino Southern Oscillation early signals.Agric Syst 205:1\u0026ndash;12. https://doi.org/10.1016/j.agsy.2022.103564\u003c/li\u003e\n\u003cli\u003eCirino PH, F\u0026eacute;res JG, Braga MJ, Reis E (2015) Assessing the Impacts of ENSO-related weather effects on the Brazilian Agriculture. Procedia Econ Finance 24:146\u0026ndash;155. https://doi.org/10.1016/S2212-5671(15)00635-8\u003c/li\u003e\n\u003cli\u003eConab - Companhia Nacional de Abastecimento (2025) Acompanhamento da safra brasileira Gr\u0026atilde;os, 10\u0026ordm; levantamento - safra 2024/25. Bras\u0026iacute;lia 12:1-125. https://www.gov.br/conab/pt-br/atuacao/informacoes-agropecuarias/safras/safra-de-graos Acessed 07 august 2025\u003c/li\u003e\n\u003cli\u003eCrespo F, Fontana D, Schirmbeck J (2024) Effects of a Strong ENSO Phenomenon on Evapotranspiration Flux Dynamics in the State of Rio Grande do Sul (RS), Brazil. Rev Bras Meteorol 39:1\u0026ndash;9. http://dx.doi.org/10.1590/0102-77863910034\u003c/li\u003e\n\u003cli\u003eCurceac S, Ternynck C, Ouarda TBMJ, Chebana F, Niang SD (2019) Short-term air temperature forecasting using nonparametric functional data analysis and SARMA models. Environ Model Softw 111:394\u0026ndash;408. https://doi.org/10.1016/j.envsoft.2018.09.017\u003c/li\u003e\n\u003cli\u003eCustodio MS (2016) \u0026Iacute;ndice de nebulosidade diurna do estado do Rio Grande do Sul: climatologia e impactos do El Ni\u0026ntilde;o oscila\u0026ccedil;\u0026atilde;o sul. Cienc Nat 38:382\u0026ndash;392. https://doi.org/10.5902/2179460X18175\u003c/li\u003e\n\u003cli\u003eDu Y, Zhao Q, Chen L, Yao X, Zhang H, Wu J, Xie F (2020a) Effect of Drought Stress during Soybean R2\u0026ndash;R6 Growth Stages on Sucrose Metabolism in Leaf and Seed. Int J Mol Sci 21:1\u0026ndash;19. https://doi.org/10.3390/ijms21020618\u003c/li\u003e\n\u003cli\u003eDu Y, Zhao Q, Chen L, Yao X, Xie F (2020b) Effect of Drought Stress at Reproductive Stages on Growth and Nitrogen Metabolism in Soybean. Agronomy 10:1\u0026ndash;21. https://doi.org/10.3390/agronomy10020302\u003c/li\u003e\n\u003cli\u003eFahad S, Bajwa AA, Nazir U, Anjum SA, Farooq A, Zohaib A, Sadia S, Nasim W, Adkins S, Saud S, Ihsan MZ, Alharby H, Wu C, Wang D, Huang J (2017) Crop Production under Drought and Heat Stress: Plant Responses and Management Options. Front Plant Sci 8:1\u0026ndash;16. https://doi.org/10.3389/fpls.2017.01147\u003c/li\u003e\n\u003cli\u003eFan B, Tao W, Qin G, Hopkins I, Zhang Y, Wang Q, Lin H, Guo L (2020) Soil micro-climate variation in relation to slope aspect, position, and curvature in a forested catchment. Agric For Meteorol 290:1\u0026ndash;17. https://doi.org/10.1016/j.agrformet.2020.107999\u003c/li\u003e\n\u003cli\u003eFan D, Gan W (2025) Does the adoption of Internet digital technologies improve farmers\u0026rsquo; fertilizer management performance? An empirical analysis based on the China Rural Revitalization Survey. Int J Agric Sustain 23:1\u0026ndash;18. https://doi.org/10.1080/14735903.2025.2467619\u003c/li\u003e\n\u003cli\u003eFebrero-Bande M, De la Fuente MO (2012) Statistical computing in functional data analysis: The R Package fda.usc. J Stat Softw 51:1\u0026ndash;28. https://doi.org/10.18637/jss.v051.i04\u003c/li\u003e\n\u003cli\u003eFeix RD, Leusin J\u0026uacute;nior S, Borges BK, Pessoa ML (2022) Painel do Agroneg\u0026oacute;cio do Rio Grande do Sul \u0026ndash; 2022. Porto Alegre: Secretaria de Planejamento, Governan\u0026ccedil;a e Gest\u0026atilde;o, p 78\u003c/li\u003e\n\u003cli\u003eFeng S, Hao Z (2021) Quantitative contribution of ENSO to precipitation-temperature dependence and associated compound dry and hot events. Atmos Res 260:1\u0026ndash;7. https://doi.org/10.1016/j.atmosres.2021.105695\u003c/li\u003e\n\u003cli\u003eFontana DC, Almeida TS (2002) Climatologia do n\u0026uacute;mero de dias com precipita\u0026ccedil;\u0026atilde;o pluvial no Estado do Rio Grande do Sul. Rev Bras Agrometeorol 10:135\u0026ndash;145\u003c/li\u003e\n\u003cli\u003eFurtak K, Wolińska A (2023) The impact of extreme weather events as a consequence of climate change on the soil moisture and on the quality of the soil environment and agriculture - A review. Catena 231:1\u0026ndash;15. https://doi.org/10.1016/j.catena.2023.107378\u003c/li\u003e\n\u003cli\u003eGon\u0026ccedil;alves GMC, Gon\u0026ccedil;alves JGR, Carvalho Paulino JF, Almeida CP, Carbonell SAM, Chiorato AF (2022) Water deficit on the physiological, morphoagronomic, and technological traits of carioca common bean genotypes. Sci Agric 79:1\u0026ndash;10. https://doi.org/10.1590/1678-992X-2021-0016\u003c/li\u003e\n\u003cli\u003eG\u0026oacute;recki T, Smaga L (2017) fdANOVA: an R software package for analysis of variance for univariate and multivariate functional data.Comput Stat 34:571\u0026ndash;597.https://doi.org/10.1007/s00180-018-0842-7\u003c/li\u003e\n\u003cli\u003eGrimm AM (2003) The El Ni\u0026ntilde;o impact on the summer monsoon in Brazil: regional processes versus remote influences. J Clim 16:263\u0026ndash;280. https://doi.org/10.1175/1520-0442(2003)016\u0026lt;0263:TENIOT\u0026gt;2.0.CO;2\u003c/li\u003e\n\u003cli\u003eGuimar\u0026atilde;es DP, Reis R (2012) Impactos do fen\u0026ocirc;meno ENOS sobre a temperatura no Brasil. Rev Espinhaco 1:34\u0026ndash;40. https://doi.org/10.5281/zenodo.3968883\u003c/li\u003e\n\u003cli\u003eHeinemann AB, Sentelhas PC (2011) Environmental group identification for upland rice production in central Brazil. Sci Agric 68:540\u0026ndash;547. https://doi.org/10.1590/S0103-90162011000500005\u003c/li\u003e\n\u003cli\u003eHeinemann AB, Costa-Neto G, Fritsche-Neto R, Matta DH, Fernandes IK (2022) Enviromic prediction is useful to define the limits of climate adaptation: A case study of common bean in Brazil. Field Crops Res 286:1\u0026ndash;19. https://doi.org/10.1016/j.fcr.2022.108628\u003c/li\u003e\n\u003cli\u003eHeinemann AB, Matta DH, Stone LF, Costa-Neto G, Resende RT, Gon\u0026ccedil;alves PAO, Justino LF (2025) Envirotyping-informed mixed models to study the climatic drivers and yield seasonal variation for common beans in Brazil. Eur J Agron 171:1\u0026ndash;15. https://doi.org/10.1016/j.eja.2025.127821\u003c/li\u003e\n\u003cli\u003eHintz G, Carcedo A, Almeida LF, Corassa G, Horbe T, Pott L, Schwalbert R, Hefley T, Prasad PVV, Ciampitti I (2025) Climate-adaptative management strategies for soybean production under ENSO scenarios in Southern Brazil: An in-silico analysis of crop failure risk. Agric Syst 222:1\u0026ndash;11. https://doi.org/10.1016/j.agsy.2024.104153\u003c/li\u003e\n\u003cli\u003eHofmann GS, Cardoso MF, Alves RJV, Weber EJ, Barbosa AA, Toledo PM, Pontual FB, Salles LO, Hasenack H, Cordeiro JLP, Aquino FE, Oliveira LFB (2021) The Brazilian Cerrado is becoming hotter and drier. Glob Change Biol 27:1\u0026ndash;14. https://doi.org/10.1111/gcb.15712\u003c/li\u003e\n\u003cli\u003eHolm S (1979) A simple sequentially rejective multiple test procedure. Scand J Stat 6:65\u0026ndash;70\u003c/li\u003e\n\u003cli\u003eIBGE - Instituto Brasileiro de Geografia e Estat\u0026iacute;stica (2020) Produ\u0026ccedil;\u0026atilde;o Agr\u0026iacute;cola Municipal 2020 \u003cu\u003ehttps://sidra.ibge.gov.br/pesquisa/pam/tabelas\u003c/u\u003e Acessed 03 april 2024\u003c/li\u003e\n\u003cli\u003eIizumi T, Luo J, Challinor AJ, Sakurai G, Yokozawa M, Sakuma H, Brown ME, Yamagata T (2014) Impacts of El Ni\u0026ntilde;o Southern Oscillation on the global yields of major crops. Nat Commun 5:1\u0026ndash;7. https://doi.org/10.1038/ncomms4712\u003c/li\u003e\n\u003cli\u003eINMET - Instituto Nacional de Meteorologia (2020) Normais Climatol\u0026oacute;gicas. https://clima.inmet.gov.br Acessed 13 august 2025\u003c/li\u003e\n\u003cli\u003eJustino LF, Heinemann AB, Matta DH, Stone LF, Gon\u0026ccedil;alves PAO, Silva SC (2025) Spatio-temporal dynamics of water stress for common bean production in Goi\u0026aacute;s, Brazil. Theor Appl Climatol 156:1\u0026ndash;21. https://doi.org/10.1007/s00704-025-05444-9\u003c/li\u003e\n\u003cli\u003eJustino LF, Heinemann AB, Matta DH, Stone LF, Andrade FW, Silva SC (2026) Optimizing Sowing Calendars for Climate-Resilient Common Bean Production in Central-Southern Brazil: A Functional Data Analysis Approach. Resources 15:1-24. \u003cu\u003ehttps://doi.org/10.3390/resources15030040\u003c/u\u003e \u003c/li\u003e\n\u003cli\u003eKaya C, Akin S, Sarioglu A, Ashraf M, Alyemeni MN, Ahmad P (2024) Enhancement of soybean tolerance to water stress through regulation of nitrogen and antioxidant defence mechanisms mediated by the synergistic role of salicylic acid and thiourea. Plant Physiol Biochem 207:1-14.https://doi.org/10.1016/j.plaphy.2023.108320\u003c/li\u003e\n\u003cli\u003eKirtphaiboon S, Humphries U, Khan A, Yusuf A (2021) Model of rice blast disease under tropical climate conditions. Chaos Solitons Fractals 143:1\u0026ndash;8. https://doi.org/10.1016/j.chaos.2020.110530\u003c/li\u003e\n\u003cli\u003eKodama Y (1992) Large-scale common features of subtropical precipitation zones (the Baiu Frontal Zone, the SPCZ, and the SACZ) Part I: characteristics of subtropical frontal zones. J Meteorol Soc Jpn 70:813\u0026ndash;836. https://doi.org/10.2151/jmsj1965.70.4_813\u003c/li\u003e\n\u003cli\u003eKodama Y (1993) Large-scale common features of sub-tropical convergence zones (the Baiu Frontal Zone, the SPCZ, and the SACZ) Part II: conditions of the circulations for generating the STCZs. J Meteorol Soc Jpn 71:581\u0026ndash;610. https://doi.org/10.2151/jmsj1965.71.5_581\u003c/li\u003e\n\u003cli\u003eKwak IY, Moore CR, Spalding EP, Broman KW (2016) Mapping quantitative trait loci underlying function-valued traits using functional principal component analysis and multi-trait mapping. G3: Genes, Genomes, Genetics 6:79\u0026ndash;86. https://doi.org/10.1534/g3.115.024133\u003c/li\u003e\n\u003cli\u003eLabastida D, Ingvarsson PK, Rend\u0026oacute;n-Anaya M (2023) Dissecting the genetic basis of drought responses in common bean using natural variation. Front Plant Sci 14:1\u0026ndash;12. https://doi.org/10.3389/fpls.2023.1143873\u003c/li\u003e\n\u003cli\u003eLenka S, Devi R, Joseph CM, Gouda KC (2022) Effect of large-scale oceanic and atmospheric processes on the Indian summer monsoon. Theor Appl Climatol 147:1561\u0026ndash;1576. https://doi.org/10.1007/s00704-021-03896-3\u003c/li\u003e\n\u003cli\u003eLi J, Xie S, Cook E, Huang G, D\u0026rsquo;Arrigo R, Liu F, Ma J, Zheng X (2011) Interdecadal modulation of El Ni\u0026ntilde;o amplitude during the past millennium. Nat Clim Change 1:114\u0026ndash;118. https://doi.org/10.1038/nclimate1086\u003c/li\u003e\n\u003cli\u003eLi X, Yu J, Ding R (2024) El Ni\u0026ntilde;o-La Ni\u0026ntilde;a Asymmetries in the Changes of ENSO Complexities and Dynamics Since 1990. Geophys Res Lett 51:1\u0026ndash;9. https://doi.org/10.1029/2023GL106395\u003c/li\u003e\n\u003cli\u003eLiang X, Hu Z, You X (2024) ENSONet: a data-driven ENSO forecasting model with concise spatial location learning parameters and temporal embedding. Clim Dyn 62:4081\u0026ndash;4098. https://doi.org/10.1007/s00382-024-07119-z\u003c/li\u003e\n\u003cli\u003eLiaw A, Wiener M (2002) Classification and Regression by randomForest. R News 2:18\u0026ndash;22. https://CRAN.R-project.org/doc/Rnews/\u003c/li\u003e\n\u003cli\u003eLin J, Qian T (2019) Impacts of the ENSO Lifecycle on Stratospheric Ozone and Temperature. Geophys Res Lett 46:10646\u0026ndash;10658. https://doi.org/10.1029/2019GL083697\u003c/li\u003e\n\u003cli\u003eLiu Y, Cai W, Lin X, Zhang Y (2023) Nonlinear El Ni\u0026ntilde;o impacts on the global economy under climate change.Nat Commun 14:1\u0026ndash;11. https://doi.org/10.1038/s41467-023-41551-9\u003c/li\u003e\n\u003cli\u003eLopes AB, Andreoli RV, Souza RAF, Ceron WL, Kayano MT, Canchala T, Moraes DS (2022) Multiyear La Ni\u0026ntilde;a effects on the precipitation in South America. Int J Climatol 42:1\u0026ndash;16. https://doi.org/10.1002/joc.7847\u003c/li\u003e\n\u003cli\u003eLu F, Feng B, Chen L, Qiu J, Wei X (2025) How Does Rice Cope with High-Temperature Stress During Its Growth and Development, Especially at the Grain-Filling Stage? Agronomy 15:1\u0026ndash;16. https://doi.org/10.3390/agronomy15030623\u003c/li\u003e\n\u003cli\u003eMarengo JA, Liebmann B, Grimm AM, Misra V, Silva Dias PL, Cavalcanti IFA, Carvalho LMV, Berbery EH, Ambrizzi T, Vera CS, Saulo AC, Nogues-Paegle J, Zipser E, Seth A, Alves LM (2012) Recent developments on the South American monsoon system. Int J Climatol 32:1\u0026ndash;21. https://doi.org/10.1002/joc.2254\u003c/li\u003e\n\u003cli\u003eMathobo R, Marais D, Steyn JM (2017) The effect of drought stress on yield, leaf gaseous exchange and chlorophyll fluorescence of dry beans (\u003cem\u003ePhaseolus\u003c/em\u003e \u003cem\u003evulgaris\u003c/em\u003e L.). Agric Water Manag 180:118\u0026ndash;125. https://doi.org/10.1016/j.agwat.2016.11.005\u003c/li\u003e\n\u003cli\u003eMatta DH, Coelho CAS, Santos LL, Stone LF, Heinemann AB (2023) Analysis of Goi\u0026aacute;s State rainfall and temperature similarity patterns during the El Ni\u0026ntilde;o Southern Oscillation phenomenon phases across the years. Theor Appl Climatol 153:1013\u0026ndash;1031. https://doi.org/10.1007/s00704-023-04503-3\u003c/li\u003e\n\u003cli\u003eMatzenauer R, Radin B, Almeida IR (2011) Atlas Clim\u0026aacute;tico do Rio Grande do Sul. Porto Alegre: Fepagro, p. 186\u003c/li\u003e\n\u003cli\u003eMedeiros FJ, Oliveira CP (2021) Dynamical Aspects of the Recent Strong El Ni\u0026ntilde;o Events and Its Climate Impacts in Northeast Brazil. Pure Appl Geophys 178:2315\u0026ndash;2332. https://doi.org/10.1007/s00024-021-02758-3 \u003c/li\u003e\n\u003cli\u003eMladenov P, Aziz S, Topalova E, Renaut J, Planchon S, Raina A, Tomlekova N (2023) Physiological Responses of Common Bean Genotypes to Drought Stress. Agronomy 13:1\u0026ndash;14. https://doi.org/10.3390/agronomy13041022\u003c/li\u003e\n\u003cli\u003eMoura MM, Santos AR, Pezzopane JEM, Alexandre RS, Silva SF, Pimentel SM, Andrade MSS, Silva FGR, Branco ERF, Moreira TR, Silva RG, Carvalho JR (2019) Relation of El Ni\u0026ntilde;o and La Ni\u0026ntilde;a phenomena to precipitation, evapotranspiration and temperature in the Amazon basin. Sci Total Environ 651:1639\u0026ndash;1651. https://doi.org/10.1016/j.scitotenv.2018.09.242\u003c/li\u003e\n\u003cli\u003eNOAA - National Oceanic and Atmospheric Administration (2020) Historical ENSO episodes (1950\u0026ndash;present): cold and warm episodes by 791 season. National Weather Service, Climate Prediction Center. https://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears_ERSSTv3b.shtml Acessed 28 july 2025\u003c/li\u003e\n\u003cli\u003eN\u0026oacute;ia J\u0026uacute;nior RS, Sentelhas PC (2019) Soybean-maize succession in Brazil: impacts of sowing dates on climate variability, yield and economic profitability. Eur J Agron 103:140\u0026ndash;151. https://doi.org/10.1016/j.eja.2018.12.008\u003c/li\u003e\n\u003cli\u003eOfir M, Gross Y, Bangerth F, Kigel J (1993) High temperature effects on pod and seed production as related to hormone levels and abscission of reproductive structures in common bean (\u003cem\u003ePhaseolus vulgaris\u003c/em\u003e L.). Sci Hortic 55:201\u0026ndash;211. https://doi.org/10.1016/0304-4238(93)90032-L\u003c/li\u003e\n\u003cli\u003eOmar MM, Massawe BHJ, Shitindi MJ, Pedersen O, Meliyo JL, Fue KG (2024) Assessment of salt-affected soil in selected rice irrigation schemes in Tanzania: understanding salt types for optimizing management approaches. Front Soil Sci 22:1\u0026ndash;14. https://doi.org/10.3389/fsoil.2024.1372838\u003c/li\u003e\n\u003cli\u003ePapathanasiou F, Ninou E, Mylonas I, Baxevanos D, Papadopoulou F, Avdikos I, Sistanis I, Koskosidis A, Vlachostergios DN, Stefanou S, Tigka E, Kargiotidou A (2022) The Evaluation of Common Bean (\u003cem\u003ePhaseolus vulgaris\u003c/em\u003e L.) Genotypes under Water Stress Based on Physiological and Agronomic Parameters. Plants 11:1\u0026ndash;15. https://doi.org/10.3390/plants11182432\u003c/li\u003e\n\u003cli\u003ePenalba OC, Rivera JA (2016) Precipitation response to El Ni\u0026ntilde;o/La Ni\u0026ntilde;a events in southern South America - emphasis in regional drought occurrences. Adv Geosci 42:1\u0026ndash;14. https://doi.org/10.5194/adgeo-42-1-2016\u003c/li\u003e\n\u003cli\u003ePercich JA, Nyvall RF, Malvick DK, Kohls CL (1997) Interaction of temperature and moisture on infection of wild rice by Bipolaris oryzae in the growth chamber. Plant Dis 81:1193\u0026ndash;1195. https://doi.org/10.1094/PDIS.1997.81.10.1193\u003c/li\u003e\n\u003cli\u003eP\u0026eacute;rez-Plaza S, Fern\u0026aacute;ndez-Palac\u0026iacute;n F, Berrocoso M, P\u0026aacute;ez R, Rosado B (2018) Analysis of a GPS Network Based on Functional Data Analysis. Math Geosci 50:659\u0026ndash;677. https://doi.org/10.1007/s11004-018-9731-4\u003c/li\u003e\n\u003cli\u003ePolania JA, Salazar-Chavarr\u0026iacute;a V, Gonzalez-Lemes I, Acosta-Maspons A, Chater CCC, Covarrubias AA (2022) Contrasting Phaseolus Crop Water Use Patterns and Stomatal Dynamics in Response to Terminal Drought. Front Plant Sci:1\u0026ndash;16. https://doi.org/10.3389/fpls.2022.894657\u003c/li\u003e\n\u003cli\u003ePoudel S, Vennam RR, Shrestha A, Reddy KR, Wijewardane NK, Reddy KN, Bheemanahalli R (2023) Resilience of soybean cultivars to drought stress during flowering and early-seed setting stages. Sci Rep 13:1\u0026ndash;13. https://doi.org/10.1038/s41598-023-28354-0\u003c/li\u003e\n\u003cli\u003ePorch TG, Jahn M (2001) Effects of high-temperature stress on microsporogenesis in heat-sensitive and heat-tolerant genotypes of \u003cem\u003ePhaseolus vulgaris\u003c/em\u003e. Plant Cell Environ 24:723\u0026ndash;731. https://doi.org/10.1046/j.1365-3040.2001.00716.x\u003c/li\u003e\n\u003cli\u003ePrasad PVV, Boote KJ, Allen H, Thomas JMG (2002) Effects of elevated temperature and carbon dioxide on seed-set and yield of kidney bean (\u003cem\u003ePhaseolus vulgaris\u003c/em\u003e L.). Glob Change Biol 8:710\u0026ndash;721. https://doi.org/10.1046/j.1365-2486.2002.00508.x\u003c/li\u003e\n\u003cli\u003eQi Z, Cui C, Jiang Y, Chen Y, Ju J, Guo N (2022) Changes in the spatial and temporal characteristics of China\u0026rsquo;s arid region in the background of ENSO. Sci Rep 12:1\u0026ndash;16. https://doi.org/10.1038/s41598-022-21712-4\u003c/li\u003e\n\u003cli\u003eQian Y, Zhao J, Zheng S, Cao Y, Xue L (2020) Risk assessment of the global crop loss in ENSO events. Phys Chem Earth 116:1\u0026ndash;10.https://doi.org/10.1016/j.pce.2020.102845\u003c/li\u003e\n\u003cli\u003eR Core Team (2025) R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. https://www.R-project.org/ \u003c/li\u003e\n\u003cli\u003eRaza A, Razzaq A, Mehmood SS, Zou X, Zhang X, Lv Y, Xu J (2019) Impact of Climate Change on Crops Adaptation and Strategies to Tackle Its Outcome: A Review. Plants 8:1\u0026ndash;29. https://doi.org/10.3390/plants8020034\u003c/li\u003e\n\u003cli\u003eRegoto P, Dereczynski C, Chou SC, Bazzanela AC (2021) Observed changes in air temperature and precipitation extremes over Brazil. Int J Climatol 41:1\u0026ndash;18. https://doi.org/10.1002/joc.7119\u003c/li\u003e\n\u003cli\u003eRosales-Serna R, Kohashi-Shibata J, Acosta-Gallegos JA, Trejo-L\u0026oacute;pez C, Ortiz-Cereceres J, Kelly JD (2004) Biomass distribution, maturity acceleration and yield in drought-stressed common bean cultivars. Field Crops Res 85:203\u0026ndash;211. https://doi.org/10.1016/S0378-4290(03)00161-8\u003c/li\u003e\n\u003cli\u003eSattar MA, Mia S, Shanta AA, Biswas AKMAA, Ludwig F (2021) Remote Impacts from El Ni\u0026ntilde;o and La Ni\u0026ntilde;a on Climate Variables and Major Crops Production in Coastal Bangladesh. Atmosphere 12:1\u0026ndash;19. https://doi.org/10.3390/atmos12111449\u003c/li\u003e\n\u003cli\u003eScheibel CH, Nascimento AB, Ara\u0026uacute;jo J\u0026uacute;nior GN, Almeida ACS, Silva TGF, Silva JLP, Silva Junior FB, Farias JA, Santos JPAS, Oliveira-J\u0026uacute;nior JF, Silva JLB, Jo\u0026atilde;o FM, Deus AS, Teodoro I, Oliveira HFE, Silva MV (2024) Characterization of Water Bodies through Hydro-Physical Indices and Anthropogenic Effects in the Eastern Northeast of Brazil. Climate 12:1\u0026ndash;37. https://doi.org/10.3390/cli12090150\u003c/li\u003e\n\u003cli\u003eShah DA, De Wolf ED, Paul PA, Madden LV (2018) Functional data analysis of weather variables linked to Fusarium head blight epidemics in the United States. Phytopathology 109:96\u0026ndash;110. https://doi.org/10.1094/PHYTO-11-17-0386-R\u003c/li\u003e\n\u003cli\u003eShapley LS (1953) A value for n-person games. In: Kuhn H, Tucker A (eds) Contributions to the Theory of Games II. Princeton: Princeton University Press, pp 307\u0026ndash;317\u003c/li\u003e\n\u003cli\u003eSharda R, Vo\u0026szlig; S, Suthaharan S (2019) Machine Learning Models and Algorithms for Big Data Classification: Thinking with Examples for Effective Learning. New York: Springer, p. 359\u003c/li\u003e\n\u003cli\u003eSiemers FM, Bajorath J (2023) Differences in learning characteristics between support vector machine and random forest models for compound classification revealed by Shapley value analysis. Sci Rep 13:1\u0026ndash;12.https://doi.org/10.1038/s41598-023-33215-x\u003c/li\u003e\n\u003cli\u003eSilva KA, Rolim GS, Valeriano TTB, Moraes JRSC (2020) Influence of El Ni\u0026ntilde;o and La Ni\u0026ntilde;a on coffee yield in the main coffee-producing regions of Brazil. Theor Appl Climatol 139:1019\u0026ndash;1029. https://doi.org/10.1007/s00704-019-03039-9\u003c/li\u003e\n\u003cli\u003eSingh S, Sah S, Singh RN (2023) The 2023-24 El Ni\u0026ntilde;o event and its possible global consequences on food security with emphasis on India. Food Secur 15:1431\u0026ndash;1436. https://doi.org/10.1007/s12571-023-01419-8\u003c/li\u003e\n\u003cli\u003eSingh SK, Reddy VR, Devi MJ, Timlin DJ (2021) Impact of water stress under ambient and elevated carbon dioxide across three temperature regimes on soybean canopy gas exchange and productivity. Sci Rep 11:1\u0026ndash;13. https://doi.org/10.1038/s41598-021-96037-9\u003c/li\u003e\n\u003cli\u003eSohn SI, Pandian S, Kumar TS, Zoclanclounon YAB, Muthuramalingam P, Shilpha J, Satish L, Ramesh M (2021) Seed Dormancy and Pre-Harvest Sprouting in Rice\u0026mdash;An Updated Overview. Int J Mol Sci 22:1\u0026ndash;22. https://doi.org/10.3390/ijms222111804\u003c/li\u003e\n\u003cli\u003eSoltani A, Weraduwage SM, Sharkey TD, Lowry DB (2019) Elevated temperatures cause loss of seed set in common bean (\u003cem\u003ePhaseolus vulgaris\u003c/em\u003e L.) potentially through the disruption of source-sink relationships. BMC Genomics 20:1\u0026ndash;18. https://doi.org/10.1186/s12864-019-5669-2\u003c/li\u003e\n\u003cli\u003eSong Y, Wang C, Linderholm HW, Fu Y, Cai W, Xu J, Zhuang L, Wu M, Shi Y, Wang G, Chen D (2022) The negative impact of increasing temperatures on rice yields in southern China. Sci Total Environ 820:1\u0026ndash;7.https://doi.org/10.1016/j.scitotenv.2022.153262\u003c/li\u003e\n\u003cli\u003eSoureshjani HK, Nezami A, Kafi M, Tadayon M (2019) Responses of two common bean (\u003cem\u003ePhaseolus\u003c/em\u003e \u003cem\u003evulgaris\u003c/em\u003e) genotypes to deficit irrigation. Agric Water Manag 213:270\u0026ndash;279.https://doi.org/10.1016/j.agwat.2018.09.038\u003c/li\u003e\n\u003cli\u003eSouza IP, Andreoli RV, Kayano MT, Vargas FF, Cer\u0026oacute;n WL, Martins JA, Freitas E, Souza RAF (2021) Seasonal precipitation variability modes over South America associated to El Ni\u0026ntilde;o-Southern Oscillation (ENSO) and non-ENSO components during the 1951\u0026ndash;2016 period. Int J Climatol 41:1\u0026ndash;18. https://doi.org/10.1002/joc.7075\u003c/li\u003e\n\u003cli\u003eTaiz L, Zeiger E, Moller I, Murphy A (2017) Fisiologia e desenvolvimento vegetal, 6th ed. Porto Alegre: Artmed, p. 888\u003c/li\u003e\n\u003cli\u003eTamaddun KA, Kalra A, Bernardez M, Ahmad S (2019) Effects of ENSO on Temperature, Precipitation, and Potential Evapotranspiration of North India\u0026rsquo;s Monsoon: An Analysis of Trend and Entropy. Water 11:1\u0026ndash;21. https://doi.org/10.3390/w11020189\u003c/li\u003e\n\u003cli\u003eTavares CJ, Ribeiro Junior WQ, Ramos MLG, Pereira LF, Casari RACN, Pereira AF, Sousa CAF, Silva AR, Neto SPS, Mertz-Henning LM (2022) Water Stress Alters Morphophysiological, Grain Quality and Vegetation Indices of Soybean Cultivars. Plants 11:1\u0026ndash;17. https://doi.org/10.3390/plants11040559\u003c/li\u003e\n\u003cli\u003eTeegavarapu RSV, Sharma PJ (2021) Influences of climate variability on regional precipitation and temperature associations. Hydrol Sci J 66:2395\u0026ndash;2414. https://doi.org/10.1080/02626667.2021.1994976\u003c/li\u003e\n\u003cli\u003eVan der Wiel K, Matthews AJ, Stevens DP, Joshi MM (2015) A dynamical framework for the origin of the diagonal South Pacific and South Atlantic Convergence Zones. Q J R Meteorol Soc 141:1997\u0026ndash;2010. https://doi.org/10.1002/qj.2508\u003c/li\u003e\n\u003cli\u003eVargas Y, Mayor-Duran VM, Buendia HF, Ruiz-Guzman H, Raatz B (2021) Physiological and genetic characterization of heat stress effects in a common bean RIL population. PLoS One 16:1\u0026ndash;23.https://doi.org/10.1371/journal.pone.0249859\u003c/li\u003e\n\u003cli\u003eVera C, Higgins W, Amador J, Ambrizzi T, Garreaud R, Gochis D, Gutzler D, Lettenmaier D, Marengo J, Mechoso CR, Nogues-Paegle J, Dias PLS, Zhang C (2006) Toward a unified view of the American monsoon systems. J Clim 19:4977\u0026ndash;5000.https://doi.org/10.1175/JCLI3896.1\u003c/li\u003e\n\u003cli\u003eVieira C, Paula Junior TJ, Bor\u0026eacute;m A (eds) (2006) Feij\u0026atilde;o, 2nd edn. Vi\u0026ccedil;osa: UFV, pp. 13-18\u003c/li\u003e\n\u003cli\u003eVon Bloh M, N\u0026oacute;ia J\u0026uacute;nior RS, Wangerpohl X, Saltik AO, Haller V, Kaiser L, Asseng S (2023) Machine learning for soybean yield forecasting in Brazil. Agric For Meteorol 341:1\u0026ndash;12. https://doi.org/10.1016/j.agrformet.2023.109670\u003c/li\u003e\n\u003cli\u003eWang D, Li X, Tian S, He L, Xu Y, Wang X (2021) Quantifying the dynamics between environmental information disclosure and firms\u0026rsquo; financial performance using functional data analysis. Sustain Prod Consum 28:192\u0026ndash;205. https://doi.org/10.1016/j.spc.2021.03.026\u003c/li\u003e\n\u003cli\u003eWu X, Jiang D, Yang Y (2025) ENSO exacerbated the impact of compound dry\u0026ndash;hot events on maize yield over China during 1961\u0026ndash;2020. Glob Planet Change 251:1\u0026ndash;10. https://doi.org/10.1016/j.gloplacha.2025.104828\u003c/li\u003e\n\u003cli\u003eXavier AC, Scanlon BR, King CW, Alves AI (2022) New Improved Brazilian Daily Weather Gridded Data (1961\u0026ndash;2020). Int J Climatol 42:8390-8404. https://doi.org/10.1002/joc.7731\u003c/li\u003e\n\u003cli\u003eXiong D, Ling X, Huang J, Peng S (2017) Meta-analysis and dose-response analysis of high temperature effects on rice yield and quality. Environ Exp Bot 141:1\u0026ndash;9. https://doi.org/10.1016/j.envexpbot.2017.06.00\u003c/li\u003e\n\u003cli\u003eXu Y, Li Y, Nettleton D (2018a) Nested Hierarchical Functional Data Modeling and Inference for the Analysis of Functional Plant Phenotypes. J Am Stat Assoc 113:593\u0026ndash;606. https://doi.org/10.1080/01621459.2017.1366907\u003c/li\u003e\n\u003cli\u003eXu Y, Qiu Y, Schnable JC (2018b) Functional Modeling of Plant Growth Dynamics. Plant Phenome J 1:1\u0026ndash;10. https://doi.org/10.2135/tppj2017.09.0007\u003c/li\u003e\n\u003cli\u003eYang S, Li Z, Yu J, Hu X, Dong W, He S (2018) El Ni\u0026ntilde;o\u0026ndash;Southern Oscillation and its impact in the changing climate. Natl Sci Rev 5:840\u0026ndash;857. https://doi.org/10.1093/nsr/nwy046\u003c/li\u003e\n\u003cli\u003eYe C, Zheng G, Tao Y, Xu Y, Chu G, Xu C, Chen S, Liu Y, Zhang X, Wang D (2024) Effect of Soil Texture on Soil Nutrient Status and Rice Nutrient Absorption in Paddy Soils. Agronomy 14:1\u0026ndash;17. https://doi.org/10.3390/agronomy14061339\u003c/li\u003e\n\u003cli\u003eYu J, Du T, Zhang P, Ma Z, Chen X, Cao J, Li H, Li T, Zhu Y, Xu F, Hu Q, Liu G, Li G, Wei H (2024) Impacts of High Temperatures on the Growth and Development of Rice and Measures for Heat Tolerance Regulation: A Review. Agronomy 14:1\u0026ndash;24. https://doi.org/10.3390/agronomy14122811\u003c/li\u003e\n\u003cli\u003eZheng G, Zhang Y, Yue X, Li K (2023) Interpretable prediction of thermal sensation for elderly people based on data sampling, machine learning and SHapley Additive exPlanations (SHAP). Build Environ 242:1\u0026ndash;15. https://doi.org/10.1016/j.buildenv.2023.110602\u003c/li\u003e\n\u003cli\u003eZhou Q, Li Y, Wang X, Yan C, Ma C, Liu J, Dong S (2022) Effects of Different Drought Degrees on Physiological Characteristics and Endogenous Hormones of Soybean. Plants 11:1\u0026ndash;17. https://doi.org/10.3390/plants11172282\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"theoretical-and-applied-climatology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"taac","sideBox":"Learn more about [Theoretical and Applied Climatology](https://www.springer.com/journal/704)","snPcode":"704","submissionUrl":"https://submission.nature.com/new-submission/704/3","title":"Theoretical and Applied Climatology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Random Forest, Climate variability, Predictive modeling, Functional data analysis","lastPublishedDoi":"10.21203/rs.3.rs-9314260/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9314260/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe El Ni\u0026ntilde;o-Southern Oscillation (ENSO) is a major driver of climate and agricultural productivity variability, as those events modulate significant fraction of rainfall and air temperature interannual variation, key factors directly affecting crop performance. This study aimed to evaluate the impacts of ENSO on rice, common bean, and soybean yields in two periods (1961\u0026ndash;1990 and 1991\u0026ndash;2019) using a \u003cem\u003eRandom Forest\u003c/em\u003e modeling approach. The analysis focused on the states of Rio Grande do Sul (RS) and Goi\u0026aacute;s (GO), Brazil, which represent contrasting climatic regions within the country. Daily data on rainfall and maximum and minimum air temperature (1961\u0026ndash;2019) were obtained from 69 weather stations. ENSO phases were classified using the Oceanic Ni\u0026ntilde;o Index (ONI), in the Ni\u0026ntilde;o 3.4 region, and analyzed through Functional Data Analysis (FDA) and Functional Analysis of Variance (FANOVA) across two periods (1961\u0026ndash;1990 and 1991\u0026ndash;2019). Observed crop yield data were incorporated into \u003cem\u003eRandom Forest\u003c/em\u003e models to estimate the relative importance of climatic variables (air temperature and rainfall) and spatial factors (municipality) in determining agricultural productivity. The results demonstrate a clear influence of ENSO on climate variability and crop yields, particularly in Rio Grande do Sul, whereas its effects were less pronounced in Goi\u0026aacute;s. The El Ni\u0026ntilde;o phase generally favored yield increases across all evaluated crops (irrigated rice, common bean, and soybean), while the La Ni\u0026ntilde;a phase was frequently associated with yield reductions. During Neutral years, the influence of ENSO was comparatively weaker. A reduction in the relative importance of rainfall was observed in the more recent period (1991\u0026ndash;2019), indicating increased rainfall homogeneity over time. In contrast, local factors became more influential in determining crop yield. Nevertheless, ENSO remains a critical factor for explaining yield variability and for supporting agricultural management and decision-making strategies.\u003c/p\u003e","manuscriptTitle":"ENSO-driven climate variability and crop yield responses in contrasting Brazilian regions: insights from functional data analysis and machine learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-20 18:39:56","doi":"10.21203/rs.3.rs-9314260/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-05-16T15:22:25+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-14T00:48:08+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-04T16:40:57+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"215823898763992812364422354344148106328","date":"2026-04-15T22:18:00+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"242629865208143479034590715828277748074","date":"2026-04-10T13:45:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"176572199398134646246118159243724181878","date":"2026-04-10T11:36:33+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-10T09:48:45+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-07T06:15:22+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-07T06:15:04+00:00","index":"","fulltext":""},{"type":"submitted","content":"Theoretical and Applied Climatology","date":"2026-04-03T14:54:30+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"theoretical-and-applied-climatology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"taac","sideBox":"Learn more about [Theoretical and Applied Climatology](https://www.springer.com/journal/704)","snPcode":"704","submissionUrl":"https://submission.nature.com/new-submission/704/3","title":"Theoretical and Applied Climatology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"b5f9a210-5f00-417e-9fbd-1efa4a9b28a9","owner":[],"postedDate":"April 20th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Revision requested","date":"2026-05-16T15:22:25+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-14T00:48:08+00:00","index":39,"fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-04T16:40:57+00:00","index":38,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[],"tags":[],"updatedAt":"2026-05-16T15:38:31+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-20 18:39:56","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9314260","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9314260","identity":"rs-9314260","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}