Differences in the Spatio-temporal Distribution of Soil Moisture from Near Surface Hydrology in a Land Surface Scheme

Abstract:

Hydrological land-surface schemes (H-LSS) are important tools for numerical weather prediction and water resources modelling. Because they are applied at low resolution, lateral water flows within and on top of the surface are often simple, ignoring catchment scale runoff mechanisms (Bierkens, 2015). This limitation has effects on the spatial and temporal distribution of soil moisture, which is known to have a strong impact on land-atmosphere coupling (Koster et al. 2004), and may limit the effectiveness of using the runoff from these models for applications such as flood forecasting. In an effort to better partition the water balance, Soulis et al. (2000) introduced interflow and overland flow mechanisms into the Canadian Land Surface Scheme (CLASS) and coupled it to a routing model. This enhanced H-LSS, known as MESH, has been demonstrated to simulate more realistic runoff than the traditional version of CLASS (Pietroniro et al. 2007) and is one of few large scale models to include interflow processes. In a comparative study using Canadian basins we explore the impact of enhancing the hydrological processes in MESH on the spatio-temporal distribution of soil moisture by variance decomposition (Mittelbach and Seneviratne, 2012) and wavelet analysis. Compared to CLASS, the soil moisture in MESH tends to dry faster following rainfall and during snow melt as water is distributed to streams. The difference is most pronounced during wet seasons (spring) while the differences are minor during dry periods, even following rain. As a result, the influence of the time variant and time in-variant controls on soil moisture variance were 23% higher and 21% lower respectively for MESH compared to CLASS. Significant differences were most prominent at short time scales, but were found to persist for periods up to two weeks during wet periods. These findings have implications for the timing and strength of land-atmosphere coupling and the timing and strength of runoff.