Assessing Climate Change Impacts on Large-Scale Soil Moisture, its Temporal Variability and Associated Drought-Flood Risks

Abstract:

Soil moisture is a dynamic variable of great importance for water cycling and climate, as well as for ecosystems and societal sectors such as agriculture. Model representation of soil moisture and its temporal variability is, for instance, central for assessing the impacts of hydro-climatic change on drought and flood risks. However, our ability to assess such impacts and guide appropriate mitigation and adaptation measures is challenged by the need to link data and modeling across a range of spatiotemporal scales of relevance for the variability and change of soil moisture in long-term time series. This paper synthesizes recent advances for meeting this challenge by a relatively simple, analytical, data-driven approach to modeling the variability and change of large-scale soil moisture under long-term hydro-climatic change. Model application to two major Swedish drainage basins, and model-data comparison for ten study catchments across the United States shows the model ability to reproduce variability dynamics in long-term data series of the key soil-moisture variables: unsaturated water content and groundwater table position. The Swedish application shows that human-driven hydro-climatic shifts may imply increased risk for hydrological drought (runoff-related) and agricultural drought (soil moisture-related), even though meteorological drought risk (precipitation-related) is unchanged or lowered. The direct model-data comparison for ten U.S. catchments further shows good model representation of seasonal and longer-term fluctuation timings and frequencies for water content and groundwater level, along with physically reasonable model tendency to underestimate the local fluctuation magnitudes. Overall, the tested modeling approach can fulfill its main aim of screening long-term time series of large-scale hydro-climatic data (historic or projected for the future by climate modeling) for relatively simple, unexaggerated assessment of variability and change in key statistics of soil moisture over large spatiotemporal scales.