Impact of Forest Service Activities on Stream Flow

Home Browse Search

Troendle/Nankervis/Porth Page31 5/22/2003 interception savings and thus allow a net increase in snow pack water equivalent to occur under harvested conditions (See figures 13 - 15 in Troendle and Nankervis 2000). The Rho, or redistribution, function used in the original version of the WRENSS hydrologic model was disabled. The snow pack accumulation in harvested areas was increased in direct proportion to the percentage of basal area removed below the "fully forested" state (BA..). Since WRENSS procedure was developed to assess "change" in water available for stream flow following silvicultural activity this adjustment was procedurally consistent with the original model although it resulted increasing net precipitation. In the original WRENSS hydrologic procedure (Troendle and Leaf 1980), the Swanson and Bernier (1986) electronic version, and the Troendle and Nankervis (2000) modified application of the WRENSS Hydrologic model, the estimate of precipitation input to the model was consistently increased in harvested areas, relative to the surrounding forest. However, depending on whether or not the estimate of precipitation input to the model represents gross precipitation (entering the top of the canopy) or net precipitation (through fall) increasing the estimate of precipitation may have been in error. In the present North Platte analysis, the precipitation estixnate used for each stand polygon was obtained by intersecting that stand polygon with the Oregon State Climate (precipitation) map for Colorado using a GIS. Because of the way in which it was derived, the mapped estimate of precipitation indexes the precipitation that enters the top of the canopy (gross precipitation) and as such it represents an index to precipitation not previously subjected to interception losses. Therefore we assume this estimate of precipitation reflects what would actually fall into a forest opening and which would be subject to interception losses as it enters a canopy (see Sturgis , Schmidt and Troendle 1989, Troendle, et al 1994). To accommodate this interpretation, we changed the functions in the hydrologic model. The WRENSS hydrologic model as used in this effort inverts the interception functions presented earlier (Troendle and Nankervis 2000) and assumes the estimate of precipitation input to the model reflects gross precipitation and equates to what would be measured in a forest openttlg, not subject to wind scour. Deposition, or net input, under any other forest canopy reflects an abstraction due to interception reduction that is in proportion to the increase in stand density from completely open to 90 percent of the fully forested situation (see figure 1). The evaporative (interception) loss that occurs under fully forested conifers during the winter and spring seasons represent 32, 26, and 17 percent of the gross precipitation respectively for north, East and West, and South facing aspects. The average winter spring interception loss for all aspects is 25 percent of the gross precipitation for conifers and 11 percent for aspen. The average loss of 25 percent for conifers compares well with other published estimates of interception loss (Wilm and Dunford 1948, Kittredge 1948, Helvey 1967, Schmidt and Troendle 1992, and Troendle and Reuss 1997). This approach does cause the estimate of net precipitation, and ultimately the estimate of water available for stream flow to decrease. However, the resulting stream flow estimates generated using the current version of the WRENSS hydrologic model appear more reasona.ble when compazed to observed stream flow measurements at Fool Creek, Deadhorse Creek, and Coon Creek. Regardless, it must be remembered that WRENSS hydrologic procedure was developed to focus on the prediction of change. The user of the WENSS analysis was encouraged to calibrate the level and distribution of flow using representative stream flow measurements while relying on the WRENSS hydrologic model to simulate expected change (Troendle and Leaf 1980). In the original calibration process in developing the WRENSS nomographs (Troendle and Leaf 1980) and in the re-calibration process tha.t developed the new ET Modifier coefficients (Spearnak 1990), the