EIS Arapahoe & Roosevelt National Forest, Pawnee National Grassland

Home Browse Search

used GIS to deteimine the area ot each vegetation type within each precipitation zone (e.g. 16-20 ? inches. 20-25 inches) and deteimine a mean weighted annual precipitation as described above tor aspect. The model requires monthly precipitation. To determine this, I gathered monthly precipitation data from SNOTEL sites wrthin or adjacent to the Forest, converced each monthly value to a percentage of annual precipitation and then avei•aged the values. I then used the monthly distributions and the weighted mean annual precipitation for each of the vegetation types as precipitation input for the model. At least two other methods may be available for estimating precipitation. Soil sutveys provide precipitation ranges for each ecological land unit (ELU). ELU's could be overlayed with vegetation maps to determine the area of each vegetation type within each precipitation zone. Some Forests may have developed precipitation vs. elevation curves for vaiious areas on the Forest. Foi• this case, elevation zones could be overlayed with vegetation maps to estimate precipitation for each vegetation type. Chuck Troendle suggested that rather than calculating a mean weighted annual precip for each timber type, WRENSS should be run for every precip reeime for a given timber type and the resulting water yield should be weighted by area. He suggested calculating water yields using this method for one or two timber types and comparing the results to those provided by the alternate method. When I did this for west slope and east slope lodgepole pine. I found that predicted water yield increased by 2% for the west side and decreased by 8% for the east side. concluded that any changes would be largely canceled out for Forest and that the potentially greater precision did not wai7•ant the significantly greater data input which would be required. The WRENSS procedure adjusts precipitation and evapotranspiration based on the windward width of the harvested opening in relationship to the surrounding forest canopy. Changes in aerodynamics cause increased snow deposition into the harvested areas. The model predicts that deposition will increase with increased opening size until the windward width is equal to approximately five tree heights. Deposition then begins to decrease until it retui•ns to pretreat- ment conditions at approximately 13-14 tree heights in width. For openings larger than 24 tree heights, WRENSS predicts a reduction in snow retention as snow scour increases (see WRENSS Fig. III-6). This adjustment can be made in the model by adjusting the "Windward Width" variable. Troendle recommended another approach. He said that more recent research in both Canada and the U.S. indicated that if sufficient roughness were left in harvest units to retain snow, harvest units of any size would accumulate additional snowpack, and that the amount of increased snow retention was primarily a function of aspect. He suggested manually increasing the post-harvest precipitation by 20% for south aspects. 30% for east-west aspects, and 40°1o for north aspects. Because the model does not allow for the adjustment of post-harvest precipitation (except through the indirect adjustment of windward width), this adjustment requires that the model be run twice for each species-location regime. The model is tirst run with unadjusted precipitation to estimate pre-treatment water yield, then the precipitation is increased and the model is run again to estimate post-treatment water yield. "Windward Width" is set to zero for both runs to effectively turn off the models adjustment of precipitation. Water yield increase for the regime is then the difference between pre-treatment and post-[reatment water yield. 6