Scale mismatches between vegetation productivity and precipitation in the agropastoral landscape of Narok, Kenya

Abstract:

Climate change can have devastating impacts on social-ecological sustainability of vulnerable agropastoral systems. Historical trends for Eastern Africa show shifts in rainfall patterns, increasing inter-annual variability and fluctuations in rains, while projections indicate likelihood of increased rainfall intensity and shifts in climate suitability for crops. Given the recent expansion of cropland into arid and semi-arid land and forests, it is important to understand how the changes affect ecosystem function at landscape scales, and their feedbacks with social-ecological systems. Limited studies have been done to characterize dynamics of agropastoral landscapes in the highly vulnerable sub-Saharan region. We investigate landscape dynamics of the agropastoral landscape of Narok, Kenya, focusing on the spatial-temporal relationship between precipitation and vegetation productivity. To study vegetation dynamics, we utilize 16-day Moderate Resolution Imaging Spectroradiometer (MODIS) Enhanced Vegetation Index (EVI) at 250 m spatial resolution. To understand precipitation patterns, we use the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS), available at 5 km spatial resolution. For both CHIRPS and EVI, we quantify spatial-temporal trends by calculating the rate and direction of change between 2003-2019. Then, beginning at the landscape level, we calculate spatial-temporal correlation between the two at multiple levels, to determine the localized role of, or lack thereof, of precipitation on productivity. Preliminary results indicate declining productivity in most areas and for all months apart from October, November and December, little but significant decline in January precipitation, extensive and significant increases in precipitation in June, July and October, and no significant change for the rest of the months. The relationship between productivity and precipitation was evident at local but not landscape scales, indicating a scale mismatch. Narok is typical of most tropical agropastoral systems especially those in the sub-Saharan Africa. This study will generate essential knowledge on dynamics of such landscapes that are necessary for determining contextualized adaptation strategies and enhancing social-ecological resilience.