Quantifying the Rate of Surface Soil Drying Following Precipitation Events Using PBO H2o Soil Moisture Time Series

Abstract:

Surface soil moisture affects latent and sensible heat fluxes, as well as setting the top boundary condition for water redistribution within the soil column. The fluctuations in surface soil moisture have been described in numerous modeling studies, but characterization based on measurements is lacking. We use a new soil moisture dataset based on reflected GPS signals to provide some constraints on rates of surface soil drying after a rain event.

The soil moisture time series used in this study are derived from GPS data collected at NSF’s EarthScope Plate Boundary Observatory (PBO) sites. The University of Colorado Boulder’s PBO H₂O project estimates daily near-surface soil moisture (approximately 0-5 cm) from the interference pattern between the direct and ground-reflected GPS signals. The sensing footprint is ~1000 m², and thus intermediate in scale between in situ and remotely sensed observations. Twelve sites from this network of more than 100 were used in this study.

To characterize the rate of soil drying, we fit exponential curves to daily soil moisture observations following ten isolated rainfall events at each site. Event sizes varied from 5 to 40 mm and were followed by 17 days without rain. The decay model fits the data quite well, with r² values exceeding 0.85 in nearly all cases. For 95% of the events studied, the exponential decay constant (e-folding time) fell between 2 and 6 days. Precipitation amount is not correlated with drydown rates. Instead, the rate of soil drying is well-correlated with air temperature: the exponential constant decreases by 0.1 days per degree Celsius. We are currently investigating how other factors, such as soil type and vegetation, influence soil drying. This study highlights the utility of the PBO H₂O soil moisture product. Surface soil moisture changes rapidly, and thus the dynamics of surface soil moisture cannot be accurately characterized using datasets based on less than daily measurements.