A hydropedological approach to describing catchment spatial organization: linkages between soil development, groundwater regimes, and solute patterns in a headwater catchment

Abstract:

Catchments exhibit heterogeneity in biophysical properties at different scales, making prediction and understanding catchment function a challenge. One source of this complexity stems from the interactions between the spatiotemporal variability of hydrological states and fluxes and the difficulty in quantifying spatial structure and organization of soils. At the Hubbard Brook Experimental Forest in New Hampshire, we have embraced the structure and organization of soils using soil functional groups that reflect dominant pedogenic and hydrologic process histories to understand the spatial relations between runoff sources and solute export and retention. We found striking spatial patterns of soil development that reflect the influence of transient water tables within the solum in nearly all landscape positions of the catchment. Shallow bedrock and variably low hydraulic conductivity in the subsoil promoted lateral flow and the development of soil horizons along hillslope flowpaths rather than in vertical profiles. Distinct groundwater dynamics and unsaturated flow direction among soil groups produced laterally and non-laterally developed podzols in a spatially organized manner across the catchment.

Our approach was based on a hydropedological conceptual model developed from over 170 soil characterization pits in podzolized soils, 25 continuously recording groundwater wells, a matric potential sensor network, and both soil and soil solution chemistry in a 42 ha glaciated, forested catchment. These podzolized soils were classified into five functional groups called hydropedologic units (HPUs), which were established according to soil morphology, but had distinct groundwater dynamics and unsaturated flow direction.

We show that HPU spatial patterns were predicted from a multinomial logistic regression model with 80% overall accuracy using topographic and bedrock-related metrics determined from a LiDAR-derived DEM and field surveys of bedrock outcroppings. The most important metrics for predicting soil groups were Euclidean distance from bedrock outcrop, the topographic wetness index, bedrock-weighted upslope accumulated area, and topographic position index.

Threshold responses were identified in storage-discharge relationships of subsurface flow, with thresholds varying among HPUs suggesting variably connected/disconnected active areas that promote streamflow generation. The spatial distribution of HPUs appeared to be highly dependent on local drainability and frequency and duration of transient saturation within the solum. These results suggest that soil horizonation within HPUs was indicative of distinct groundwater flow regimes. The hydrologic connection of shallow upland soils of the catchment where strongly lateral podzols developed also appeared to be a source for dissolved organic carbon (DOC) in the stream network. Spatial patterns of DOC concentrations were associated with the distribution of HPUs and their thresholds to stormflow generation.

Overall, the hydropedological approach we developed that explicitly recognized the spatial organization of soils in a catchment provides insight into the spatial and temporal contributions to streamflow generation and solute export from headwater catchments. This approach may help structure spatial models of catchment hydrology and link spatial patterns of soils to elusive subsurface flow regimes and the variability of chemical fluxes that occur throughout stream networks.