Developing diagnostic signatures from in-situ soil moisture networks under different land-uses

Abstract:

Recent development of in-situ soil moisture networks (catchment- to national- scale clusters of sensors) has enabled us to capture soil moisture dynamics at unprecedently high spatial and temporal resolutions. Yet, soil moisture models are often evaluated based on the spatial mean or variability over catchments. Here, soil moisture values may not fully characterize the watershed behavior, or enable us to test whether models accurately represent processes such as deep-drying of soil layers or anomalies in seasonal cycles for drought prediction.

In this study, we applied the emerging concept of hydrologic signatures to in-situ soil moisture network data to quantify the soil moisture dynamics and derive the process implications at a catchment scale. Hydrologic signatures are indices that characterize multiple aspects of catchment functionality. Model parameters and structures can be evaluated based on the models’ ability to reproduce the signatures derived from observation data.

We selected eight soil moisture network sites under different land-uses, from shrublands to forests, snowy mountains to wetlands. From the time series of soil moisture data, we extracted six types of signatures (see Figure, modified from Branger and McMillan, 2019): field capacity, number of peaks in the probability distribution function, event response magnitude and amplitude, and seasonal transition duration and dates. Signatures extracted at different sensor depths were compared using a one-way analysis of variance, and the results were used to understand the surface-subsurface dynamics. Lastly, to link the signatures and processes, we built hypotheses about expected signatures by reviewing literature and tested the hypotheses with the derived signatures and relevant hydro-climatic variables.

Our results present surface-subsurface soil moisture dynamics and underlying processes at a catchment scale under different hydro-climatic environments. The results advance our ability to create standardized classifications of catchments and diagnostic tools for soil moisture models.