The spatio-temporal soil moisture variation along the major tributaries of Zambezi River in the Mbire District, Zimbabwe

The livelihoods and subsistence of the largest number of people in Mbire communal areas depends on soil moisture evolving from seasonal floods that is tapped for agriculture. The communities have not realised the full potential of floodplains soil moisture in increasing crop yields, vital for hunger and poverty alleviation. Proper quantification of soil moisture levels is very vital for improvement and management of agricultural activities in these floodplains. Soil moisture is typically measured in the field at point scale, which is expensive, tedious, time consuming and does not account for regional spatial variability. Methods based on Remote sensing (RS) and hydrological modelling provide alternative tools to obtain estimates of spatial and temporal variation of soil moisture which is vital in water resources management in ungauged and remote areas. In this study integration of the Surface Energy Balance System (SEBS) and the Topographic driven MODEL (TOPMODEL) was carried out in estimating soil moisture for Mbire district in Zimbabwe and comparing them against ground measurements. Five atmospherically corrected MODIS images were processed and compared from ground data collected once a month on fifty-four soil moisture sampling sites. The proportional relation of the relative soil moisture with the relative evapotranspiration was used for the estimation of soil moisture from remote sensing as SEBS was primarily developed for the estimation of surface turbulent fluxes. The rainfall runoff model (TOPMODEL) whose land surface inputs were obtained from Remote Sensing was calibrated with runoff data was used to simulate soil moisture and compared from ground information collected on fifty-two soil moisture sampling sites. An upscaling procedure to improve the comparison between point measurements, remote sensing and TOPMODEL outputs was accomplished by the use of geostatistical tools. Land suitability analysis for flood recession farming was done using distance from stream network, vertical channel distance, and land use/cover datasets. The performance indicators for TOPMODEL simulations showed an acceptable match with measured discharge, indicating a satisfactory Nash Sutcliffe model efficiency (NSE= 0.77 and 0.81) and good percent bias (PBIAS= -6.04% and -10.5%) for Manyame and Angwa sub catchments respectively. The study revealed that there is a strong relationship (R2= 0.80) between upscaled ground soil moisture measurements and remote sensing methods (SEBS) for the period of March to July 2013. This allowed this methodology for modelling initialisation to be adopted. The study revealed that there is a fair relationship (R2= 0.60) between upscaled ground soil moisture measurements and hydrological modelling (TOPMODEL) for the period of March to July 2013. Results also show that 22800 hectares of Mbire district is suitable to moderately suitable for flood recession farming. The study concludes that use of remote sensing and hydrologic models coupled with Geographic Information System (GIS) are effective in soil moisture monitoring which is vital for planning and management of available water resources for sustainable development Mbire community.,WaterNe