Year of Award

2007

Document Type

Dissertation

Degree Type

Doctor of Philosophy (PhD)

Degree Name

Forestry

Department or School/College

College of Forestry and Conservation

Committee Chair

Hans R. Zuuring

Commitee Members

Kelsey S. Milner, Donald F. Potts, Scott S. Woods, William W. Woessner

Keywords

calibration, evaluation, snow, streamflow, SWAT, validation

Abstract

Changes in forest structure resulting from natural disturbances, or managed treatments, can have negative and long lasting impacts on water resources. To facilitate integrated management of forest and water resources, a System for Long-Term Integrated Management Modeling (SLIMM) was developed.

By combining two spatially explicit, continuous time models, vegetation patterns can be simulated forward in time based on management criteria. Output from the SIMPPLLE vegetation simulator are converted into landcover maps at every time-step and used to predict hydrologic watershed responses to time-series landcover change with the SWAT model. Long-term watershed responses to vegetation management scenarios can therefore be evaluated from both terrestrial and hydrologic perspectives.

Watersheds are common landscape analysis units, but vegetation dynamics within them do not function in isolation. Repeated century spanning SIMPPLLE simulations produced succession patterns that were significantly different in 84% of analysis watersheds when each was considered in isolation and within their landscape context. Watersheds with >30% internal forest cover, and <10% barren ground along their perimeters were more connected to landscape processes than those with more barren boundaries, and less forest cover within them.

Calibration of SWAT was based on four years of streamflow and climate data recorded within the Tenderfoot Creek Experimental Forest research watershed. Validation with an additional four years used both traditional and objective regression-based hypothesis testing procedures. Adjustment of snow process, surface runoff lag, and groundwater recession parameters contributed most significantly to model calibration. Results confirm that when calibrated in a forested mountain watershed having snow-dominated hydrology, SWAT can predict annual, monthly and daily streamflow with high levels of accuracy and efficiency.

For demonstration, SLIMM was used to evaluate natural and fire-suppressed forest management alternatives over a 300-year period. Compared to natural development, fire suppression created larger stand sizes, greater levels of aggregation, and increased the likelihood of process propagation across the landscape. Averaged over all simulations, fire suppression reduced annual water yield by up to 3%, streamflow variability by a factor of four, and the magnitude of annual peak flows by 15%. Literature supported results highlight the applicability of SLIMM as a management tool.

Download

Share

COinS
 

© Copyright 2007 Robert Steven Ahl