Dissertation
The research I have undertaken for my PhD dissertation at the University of Wisconsin – Madison has concentrated on evaluating – through field measurements, remote sensing, and numerical simulation – the effectiveness of restoring a stream-floodplain ecosystem in the Driftless Area of southwestern Wisconsin (see http://hydroecology.cee.wisc.edu/EBP/ for more information).
The key research questions of the project are:
-
1)what are the hydrologic controls on root water uptake and
-
2)what aspects of the soil water regime control vegetation composition and patterning.
An extensive monitoring network consisting of hydrologic (piezometers, stream gaging, surface soil moisture transects, etc.), pedologic (grain-size analysis), and vegetative (sampling quadrats and leaf porometry) methods combined with subsurface geophysical imaging (ground-penetrating radar) has revealed the important influence of fine-grained alluvial deposits (acting as confining layers) on restricting groundwater upwelling across certain parts of the floodplain. Predicting the soil water regime across the floodplain would be very challenging without a proper understanding of this hydrostratigraphic influence. To further explore this phenomenon, I have created a numerical variably-saturated groundwater flow model to simulate a vertical soil column with vegetation. This model has revealed the importance of effectively characterizing both the variability in hydraulic conductivity of floodplain alluvium with depth and how various plant species limit transpiration during periods of water and oxygen stress on the soil water regime. Currently, this model is being combined with a 2-D groundwater flow model to simulate the soil water regime across the entire floodplain site and create a more transferable methodology that can be used at other locations.
Vegetation sampling across the floodplain site has also reinforced the spatial variation in soil water regimes as reflected by variation in the presence/absence of obligate and facultative wetland plant species. Relating the water regime to plant composition was performed using a statistical model that predicts the probability of presence of certain dominant plant species based on soil moisture metrics (e.g., mean over some time period). Initially, the soil moisture predictors were determined based on field observations but linking the predictive vegetation model with the hydrologic model will allow for a spatially-extensive prediction of plant composition under potential future scenarios such as land-use change, climate change, and management activities.
I have also collected high-resolution remote sensing data from a small aircraft to support this research. Visible, near-infrared, and thermal imagery collected by cameras owned by our research group is being used to support ground-based observations of vegetation and evapotranspiration dynamics. It is especially important to have this dataset to capture the extensive changes in vegetation and hydrology in the years following the restoration activity.
