Unveiling farmers' perceptions: a citizen science and machine learning approach to exploring drivers in the adequacy and fairness of water systems

Agricultural production sustains a large part of the population in the Great South area (sub-Saharan Africa, South America and South Asia), relying heavily on rainfed agriculture and partly on reservoir-based irrigation schemes. This study evaluates the effects of citizen science approaches in shaping the farmers' perceptions towards adequacy of the quantity of the provided water and the fairness of the irrigation distribution system in the area, using as a case study a project implemented in the Upper Blue Nile (UBN) region of Ethiopia in two irrigated communities. Harnessing the analytical power of machine learning models in extracting patterns from data, the informational content of social surveys coupled with hydrological data for the survey region from a calibrated MODFLOW-NWT groundwater (GW) model, we draw inferences on the importance of socioeconomic rather than hydrological variables as drivers in agricultural decisions about crop selection and planting period, underscoring those factors as potential criteria in drawing successful agricultural policies for crop yield optimization in the Great South area. This study extends the existing literature towards understanding the interplay between people and water under a qualitative framework with distinct policy implications.