A Catchment Based Study on Streamflow Data Assimilation
Christoph Rudiger, Jeffrey P. Walker, Jetse D. Kalma, Garry R. Willgoose, and Paul R. Houser
Soil moisture is an important variable in land surface modelling with a significant impact on climate prediction. Specifically, soil moisture content in areas with dense vegetation cover, such as the Sahel and Amazon, have been shown to have the most potential for positively influencing the predictability of precipitation. While much work has been concentrated on the assimilation of remotely sensed surface soil moisture observations to constrain land surface model predictions of soil moisture, the use of these measurements is limited to areas of low-to-moderate vegetation only. This work proposes to positively impact on soil moisture prediction in these densely forested areas through the assimilation of streamflow observations. The potential for this approach is demonstrated for a 7000km2 semi-arid catchment in a synthetic twin experiment. Observation and validation data were obtained from a “truth” simulation, using a spin-up period of 10 years to obtain the true initial conditions. An “open-loop” simulation with a degraded soil moisture initialisation provided the control experiment. Streamflow data from the truth simulation were then assimilated into the open-loop simulation in order to retrieve back the true soil moisture states. Internal catchment routing has been given careful consideration in the assimilation and subsequent retrieval of soil moisture states for each catchment. Further work will include the comparison between the individual runoff and soil moisture results of the nested subcatchments and the whole catchment as a total. This will combine the effects of time-lags due to both internal catchment routing and inter-catchment routing, showing the influence of different scales on the proposed technique.