Evaluation of the Performance of the Water Cloud Model and Modified Water Cloud Model in Estimating Soil Moisture. A case study of Kiruuli Village
Abstract
Periodic droughts and reliance on rain-fed agriculture in Uganda have limited the crop yield obtained. Irrigation farming as a straightforward measure has been adopted by the government through rehabilitation of old schemes and helping farmers set up micro-irrigation farms. The maximization of crop yield through irrigation, however, necessitates soil moisture information for irrigation scheduling. The Water Cloud Model (WCM) has been used in past studies to estimate soil moisture over vegetated areas. However, its assumption that vegetation is a homogeneous scatterer limits its accuracy. To resolve this, the Modified Water Cloud Model (MWCM) introduces the vegetation fraction parameter to account for the uneven distribution of vegetation. On a global scale, assessment of both models has been carried out using Radarsat-2 and Landsat-8 imagery over wheat fields.  This study further evaluates both models using Sentinel imagery which has a higher temporal resolution over a coffee(one of the country’s top exports) growing area. The model-estimated soil moisture was compared with in situ data collected in the topmost layer (5cm depth). The MWCM had better performance than WCM with RMSE of 3.3346 and R2 0.6175 in comparison to RMSE of 3.7482 and R2 of 0.6093 of the latter. The marginal difference was attributed to the relatively high vegetation fraction at that time. The use of more penetrating L-band SAR data, incorporation of surface roughness parameters and usage of the SAFY that is able to estimate vegetation characteristics should be explored to improve on the accuracy of the results. In addition, the setting up of in situ soil moisture monitoring stations is recommended to aid Spatio-temporal evaluation of the models.
Key Words: Soil moisture, remote sensing, modified water cloud model