Dynamics of soil salinity in the Canadian prairies: Application of singular spectrum analysis

Abstract

We studied hidden periodicities and trends in the temporal dynamics of soil salinity in the Canadian prairies at a field scale. Singular spectrum analysis (SSA) was applied to time series of the groundwater depth (Dgw), total precipitation, groundwater electrical conductivity (ECgw), and soil electrical conductivity (ECe). We found that the temporal variability of soil salinity is determined by its seasonal and quasi-3-yr oscillations controlled by similar cycles of Dgw and ECgw, which are controlled by the precipitation. In the seasonal cycle, the system “precipitation – groundwater depth – groundwater salinity – soil salinity” proceeds as follows: The maximal level of the precipitation is reached in late May. The water table peaks 2 months after the maximal level of precipitation was reached. The seasonal oscillation of ECgw is in phase with that of Dgw. ECe peaks 1.5 months after the maximal level of ECgw was reached. In the quasi-3-yr cycle, ECe peaks about 14 months after the maximal rise of the groundwater. Most likely, this cycle of Dgw is a response to the 3-yr periodicity of the precipitation observed in the region. SSA is the effective tool to detect hidden oscillations in soil processes handling relatively short time series with missing records.