Supplementary material to "From insufficient rainfall to livelihoods: understanding the cascade of drought impacts and policy implications"

implications'S. 1 -Comprehensive data colletion and analysis Drought impacts monitoring dataObservers collected the first dataset as part of their job routine.In addition to their various tasks, they regularly complete monthly questionnaires for each municipality, providing information on drought impacts and other relevant information.Employed by Ematerce, these observers are based across the state, with most offices overseeing two or three municipalities, covering 184 municipalities.The questionnaire consists of four multi-choice questions addressing drought conditions, rainfall occurrence, agriculture, and water accessibility.Additionally, there is one open question asking for information on any impacts.The observers in each state may add additional locally relevant questions.In the final question, observers are free to express any pertinent information they consider important at the time.This means the reported impacts extend beyond just those related to drought.Especially considering this open question, it stands out globally as a rare and valuable example of monitoring drought impacts through the perspectives of people "on-the-ground" who directly experience the impacts.Therefore, in this research, the analysis focuses primarily on the responses to this open-ended question.We refer readers to Table 1 for the complete questionnaire used by observers.Further details on the data collection and analysis of this dataset can be found in Walker et al. (2024). Analysis of drought impacts monitoring dataWe analyzed the open question on the reporting of impacts (question 5) using inductive reasoning and thematic analysis as a method to identify and analyze recurring patterns or themes within the dataset.This type of analysis is particularly suitable for areas lacking empirical research and provides a rich description of predominant themes across the dataset (Braun & Clarke, 2006).The analysis consisted of 3 main steps.Step 1 Coding responses: Question 5 responses were manually coded through inductive reasoning to identify within the qualitative data.We leveraged our expertise in the field to search for patterns in relation to drought impacts occurrence.To conduct this analysis, we employed the qualitative analysis program Atlas.ti (version 22), where 3641 reports completed by observers were uploaded.We created distinct groups to organize our analysis, by Ematerce offices, periods and questions.The "codes" function within Atlas.ti was used to label identified drought impacts within the reports.This process resulted in the identification of 204 different codes (Table 2).