Modelling the drivers of soil moisture in the landscape in order to apply the STAMINA model at a regional level

The STAMINA (Stability And MItigatioN of Arable systems in hilly landscapes) model is a new crop yield model than takes into account the impact of terrain on crop growth. The ability to model yield variations as a function of terrain can help policy makers plan for potential changes in climate. In its original form the STAMINA model is too complex to be run at a regional extent. The literature pointed to the key drivers of crop growth being linked to water availability in the Iandscape. The research was therefore split into two sections. The first section outlines the investigation of soil moisture in the landscape. Two field experiments were undertaken. The first measured surface soil moisture at over 100 locations, using a Delta-T ThetaProbe on eight occasions in three different fields. The results from regression models showed that up to 80% of the variation in surface soil moisture can be explained using 1:10,000 taxonomic soil units. Radiation and wetness indices combined explained a maximum of 41% of the variation. These variables also explained significant additional variation when combined with 1:25,000 soil units. The second experiment measured soil moisture at six depths from 100 mm to 1 m at 38 locations in two fields using a Delta-T profile probe. Regression models showed that 1:10,000 taxonomic soil units combined with depth explain up to 67% of the variance in soil moisture. These results highlight the importance of the drainage characteristics of the soil profile in determining soil moisture content. The second section focuses on the development of an index approach to apply the STAMINA model at a regional extent. This method and STAMINA version 1.8 were tested on three representative catchments in a 10 km2 area of Bedfordshire. They succeeded in highlighting similar areas of the landscape that are at risk from low yields. Grid size analysis suggested that a grid size of 100 m was sufficient for running the index based version of the STAMINA model. This still maintained an accurate representation ❑f the topographic features that control the modelled yield variability. Further investigation into the soil hydrology module in STAMINA version 1.8 suggested that predicted yield was very sensitive to a change in the way soil water drainage was modelled, in particular the potential for a soil to recharge over the winter months.