Inferring soil water movement and streamflow response in Canadian Prairie riparian areas using hydrologic state variables

Abstract

AbstractFlat terrain and soils with variable permeability make it difficult to assess the relative importance of surface and subsurface runoff in the Canadian Prairies, especially in riparian areas that are critical for water transmission and solute transport. The main objective of this study is therefore to determine whether patterns of hydrologic state variables, namely, near‐surface soil moisture (SM), soil electrical conductivity (SEC), and soil temperature (ST), can help infer riparian‐to‐stream soil water movement in Prairie landscapes. Focus is on the near‐level Catfish Creek Watershed (south‐eastern Manitoba, Canada) where 3 riparian sites were monitored: a naturally vegetated grassland site, a headwater forested site, and a highly impacted grassed site adjacent to an engineered drainage dyke and a man‐made drainage channel. Data from 9 to 12 SM, SEC, and ST surveys completed at each site in 2015 using a 75‐point grid are matched with riparian water table data, surface water level data from adjacent drainage channels, and indicators of antecedent moisture conditions. Pattern characteristics, in the form of descriptive statistics and variogram parameters, are estimated for each state variable and then correlated to indicators of antecedent moisture conditions, stream, and subsurface water level data to infer soil water movement. Results show that potential evapotranspiration, depth to water table, and antecedent precipitation have a significant yet variable impact on SM, SEC, and ST patterns. A switching behaviour, between dry and wet conditions, is present in riparian areas characterized by grassland vegetation and well‐drained soils. The occurrence of shallow subsurface flow is inferred during the wettest conditions. Although riparian SM conditions are useful for predicting streamflow response in adjacent channels, such is not the case for riparian SEC and ST. Further investigations are however necessary to confirm the usefulness of SM spatial patterns for predicting streamflow response in other landscapes across the Canadian Prairies.