Soil moisture dynamics in a disturbance mediated forest, savanna and grassland ecotone

Abstract

The Pushmataha Forest Habitat Research Area in southeastern Oklahoma provides a unique opportunity to improve the knowledge of the relationships between soil moisture dynamics, plant community structure and aboveground net primary production (ANPP) in tree-grass systems. This long-term experiment contains treatments reflecting a full range of vegetation structure that can be found across the forest-grassland transition zone due the interactions of initial harvesting practices with different prescribed fire intervals. To compare vegetation structural influences on soil moisture dynamics, measurements of soil moisture, understory ANPP, tree basal area, woody species canopy openness, and soil structure were assessed. Periodic mean volumetric water content (VWC) ranged between 4.9 and 6.6 %. The highest individual VWC value was 41.9 % and the lowest VWC was 0.7%. The VWC0 (water content 24 hours following the end of the rainfall event) was significantly related to point-level structural covariates. The final model explained 72% of the residual error variance compared to a null model. Tree basal area was negatively related to VWC0, although conifer BA was not. The potential evapotranspiration (PET) ranged from 0.25 to 12.2 mm per day; reaching the peak in the summer and decreasing with the end of growing season. A simple model using only meteorological covariates was able to predict growing season soil moisture loss between rainfall events and explained 43% of the residual variance compared to the null model. Including vegetation structural and plot-level environmental covariates in the soil moisture decay model, decreased residual variance by only 2.5%. The index of stoniness was significant (p<0.0001) and positively related with soil moisture loss. Other vegetation structural covariates were non-significant and were dropped from the final model. Heavily forested treatments with thicker litterfall layer had the lowest upper soil moisture content. However, treatments with conifers dominance acted more similarly to open canopy grass dominated treatments. The final model used to predict soil moisture decay showed that steepness on the slope increased with basal area.