Abstract
Ethiopia's traditionally rain-fed agricultural sector produces about 90% food during rainy season that sometimes causes a case of food insecurity due to high seasonal precipitation variability, coupled with limited dry season irrigation facilities. Further, there is limited understanding of the water-food security aspects pertaining to different types and amounts of irrigation. Our research aims to simulate the groundwater-vadose zone interactions for water-food security at two irrigation sites in Ethiopia, namely, the large-scale Koga project in Merawi, and the small-scale Quashni project in Dangla, with different irrigation and water management (IWM) scenarios. A Modflow NWT based local groundwater (GW) model--forced with reanalysis datasets, i.e. precipitation from MSWEP v1.0 and IMERG, simulated runoff and streamflow from CREST model, and in-situ data on irrigation releases - is used in this research. The model primarily produces soil moisture (theta ), hydraulic head (H) and evapotranspiration (ET) as output. It is calibrated mainly by the hydraulic conductivity (K) and specific yield (S (sub y) ) using in-situ citizen science measurements of H and theta as reference. The GW model is simulated from Jan-2008 to July-2019, and is used to forecast theta for the current dry season (October 2019 to May 2020). The model reveals soil moisture dynamics in the vadose zone for different IWM scenarios that demonstrates a pathway for optimal water use. Moreover, the root zone theta is used as input to a calibrated crop model DSSAT to predict crop yields, in attempts to derive an understanding of future food security. The research would greatly benefit local communities in terms of decision making on optimal irrigation strategies (water amounts and suitable crop types); affirming not only the best soil conditions but also the highest crop yields, and hence contribute to water-food security improvement in the region.