DINÂMICA ESPAÇO-TEMPORAL DE VERANICOS DURANTE A ESTAÇÃO CHUVOSA NO ESTADO DO MARANHÃO, BRASIL

Abstract

Dry spells cause losses in food production, especially when prolonged and coinciding with the reproductive phase of agricultural crops. They can also impact the water resources and ecology sectors. In the state of Maranhão, despite the associated risks, dry spells are poorly studied. Therefore, the objective of this study was to analyze the occurrence of dry spells in Maranhão. The first chapter presents a literature review on climatic patterns, characteristics and causes of dry spells in Maranhão, the impacts of dry spells on agriculture, among other supporting theoretical topics. The following two chapters present an analysis of the spatiotemporal dynamics of dry spells during the rainy season (December–May) using daily precipitation data from 123 meteorological stations distributed in Maranhão (93) and surrounding regions (30). In the second chapter, dry spells were defined as ≥3 consecutive days with precipitation <1 mm (1994–2023), categorized into classes A (3–6 days), B (7–10 days), C (11–14 days), and D (≥15 days), and spatially represented using the Inverse Distance Weighting method. Principal component analysis (PCA) was applied to classes A (>frequency) and D (>duration) to identify patterns of variability in their frequencies related to geographic and atmospheric variables, and under which climatic characteristics they occur. Class A represented 62.29% of the dry spells, followed by classes B (21.09%), C (8.55%), and D (8.07%). These events were more frequent at the beginning and end of the rainy season in the south-central region of Maranhão. In PCA, most of the variation was explained by the combination of higher latitudes and altitudes (south-central), greater (lesser) cloud cover in December (March to May), with: a higher frequency of class A dry spells from December to April; and a lower (greater) frequency of class D dry spells in December (April and May). In both cases, these relationships occurred in colder and drier environments. In the third chapter, the time series (between 1963 and 2023) of extreme dry spells were subjected to regional frequency analysis using the L-moment method. Based on geographic attributes and the average duration of extreme dry spells, five homogeneous sub-regions were identified in Maranhão. The goodness-of-fit measure (Z DIST) and L-moment diagrams indicated that, for sub-regions 1 (east) and 3 (west-central), the generalized extreme value distribution is the best fit, while for sub-regions 2 (south), 4 (south-southwest), and 5 (north), the generalized logistic, Pareto type III, and generalized normal distributions are the best fit, respectively. All distributions showed satisfactory accuracy in the estimates of regional quantiles and in each station for different return periods. North (south) of Maranhão, subregion 5 (2) shows greater susceptibility to the occurrence of extreme dry spells, largely due to the inconsistency of rainfall at the beginning (end) of the rainy season. Therefore, south-central Maranhão is more prone to dry spells, and the scarcity of rain at the beginning (end) of the rainy season contributes to a higher frequency of extreme dry spells in northern (southern) Maranhão.