Spatio-temporal variability and trend of water footprints in the upper Awash basin, central Ethiopia

The water resources in the upper Awash basin in central Ethiopia are intensively utilized by the densely populated urban centers and extensive agricultural activities in the wide rift plains surrounding the major cities. The green, blue, grey and total water footprints of major crops dominantly cultivated in the sub-basin (teff, maize, sorghum and sugar cane) are characterized. The water footprints were analyzed for the period 2000 to 2010 using climatic as well as soil and crop yield data from the Debrezeit, Wonji, Melkassa and Metehara stations. The temporal and spatial variations of total water footprints (TWFs) in the basin were analyzed and mapped. The temporal trends of the Water Footprints (WFs) are interpreted using Mann-Kendall and Sen's slope tests. The crop production in the sub-basin mainly depends on green water consumption. Teff and sorghum crops in the climatically similar Debrezeit and Melkassa areas had relatively higher TWFs (4205 m³/ton and 2539 m³/ton, respectively) compared to maize. Blue water requirement to supplement rainfed agriculture decreases from Metehara, Melkassa, Wonji to Debrezeit in that order. The TWF of irrigated sugarcane (117–212 m³/ton) is less than those of rainfed crops such as maize which has a minimum TWF of 1752 m³/ton. The spatial and temporal variations in WFs in the sub-basin are determined by both climatic (effective rainfall, evapotranspiration, length of wet spell and length of growing period) and non-climatic parameters (such as fertilizer consumption, soil types, crop yields and agricultural management practices). Our results suggest cultivating less water-intensive, high-yield crops can significantly decrease the TWFs in the sub-basin. The results could be used to develop well-informed regional and national freshwater resource management policies. More broadly, our novel approach of characterizing WFs at the sub-basin scale using limited amount of locally measured data sets could be a useful tool in other parts of the world where locally measured data sets are scarce.