Evaluating the Impacts of Unexpected Forest Disturbances on Paired Catchment Calibrations of Sediment Yield and Turbidity

Abstract:

The paired catchment approach of studying the impacts of disturbance on catchment hydrology remains as perhaps the most powerful approach for direct verification of catchment scale impacts from disturbance. However, paired catchment studies are also dependent on the stability of the relationships between treated and reference catchments during calibration and evaluation periods.

A long-term paired catchment study of forest harvest impacts on sediment yield and turbidity in the Rocky Mountains of southwestern Alberta, Canada has a robust 11-year pre-treatment data record. The study intends to evaluate three alternative logging practices: clear-cutting, strip-shelterwood, and partial cutting. 3 sub-catchments in Star Creek (1035 ha) underwent harvest treatments while North York Creek (865 Ha) serves as the reference. The objective of this particular study was to explore the potential effects of unplanned and unanticipated watershed changes in two watersheds during an 11-year calibration. Sediment yield (kg ha^-1 d^-1) and turbidity (NTU) were monitored throughout the calibration period (2004-2014) prior to the 2015 harvest in Star Creek. Two unanticipated disturbances including backcountry trail rehabilitation in North York (2010) followed by a >100 year storm event in both watersheds in June 2013 may have affected the sediment yield and turbidity calibration relationships. Analysis of covariance (ANCOVA) was used to evaluate the effects of this trail rehabilitation and flooding by comparing the calibration relationships before and after these disturbances.

Despite qualitative field observations of periodically affected sediment regimes, no impact on pre- or post- calibration relationships was observed. Backcountry trail rehabilitation in North York (p=0.904 and 0.416 for sediment yield and turbidity, respectively) or flooding in both watersheds (p=0.364 and 0.204 for sediment yield and turbidity, respectively) did not produce significant changes to the calibrations. Although calibrated watershed studies bear meaningful risks from unplanned disturbances, in contrast to both scaling up process-based research and modelling studies, paired catchment studies remain as the demonstrable benchmark for powerful inference on the impacts of abrupt landscape change on watershed functions.