Forests and Water: A State of the Art Review for Colorado

Home Browse Search

and freezing at the surface of the snowpack, and possibly an increased persistence of the snowpack as a result of the increased peak SWE. Figure 2.2 showed the effect of forest harvest on the av- erage annual hydrograph for the Fool Creek catchment in the FEE On average, the annual daily maximum peak flow increased by 23% as a result of removing 50% of the forest cover (Troendle and King, 1985). The average number of days with flow at or above bankfull increased from 3.5 to 7 days per year (Troendle and Olsen, 1994). A reanalysis of data from the Wagon Wheel Gap study showed that the annual daily maximum flow increased by 50% (Van Haveren, 1988), and a similar increase in the size of the annual snowmelt peak occurred as a result of the shelterwood cut on the North Fork of Deadhorse Creek (Troendle and King, 1987). These and other stud- ies indicate that complete removal of the forest canopy in small watersheds in the sub - alpine zone will increase the size of the annual daily maximum flow by about 40 %. If only part of the forest canopy is removed, the increase in the size of the annual daily maximum flow is directly proportional to the amount of basal area removed. Increases in the instantaneous annual maximum flow are more variable (e.g., King, 1989). In general, harvest -in- duced increases in the instantaneous annual maximum flow should be similar to, or slightly smaller than, the change in the annual daily maximum peak flow because forest harvest will have a progressively smaller effect on the highest peak flow (i.e., peak snowmelt is limited by the amount of energy that can be delivered to the snow - pack). For example, forest harvest had no detectable ef- fect on the three largest instantaneous peak flows on Fool Creek despite the significant increase in the annual daily maximum peak flow (Troendle and Olsen, 1994). Harvesting 24% of the 4,100 -acre Coon Creek water- shed increased both the annual daily maximum and the instantaneous peak flows resulted by about 8 %, but these increases were not statistically significant (Troendle et al., 2001). However, there was a significant increase in the 701 to 901 flow quantiles, where the value nearest zero represents the lowest flow on record and the value closest to 100 represents the highest flow on record (Troendle et al., 2001). The absence of a statistically significant increase in the annual peak flows in the Coon Creek study as compared to Fool Creek is partly due to the much smaller number of pre- and post - treatment data points, the variability in the relationship between the treated and the control basins, and the fact that only 24% of the Coon Creek basin was harvested. Peak flow in- creases also may be smaller in larger basins with a wide 14 range of elevations and aspects, as the peak flows from different parts of the basin may not be synchronized. The effect of forest harvest on the size of peak flows is much more complicated and controversial in areas where the largest peak flows are caused either by mid- winter rain -on -snow (e.g., Harr, 1986) or rain-on- spring -snow- melt (MacDonald and Hoffman, 1995). Since rain -on- snow events are rarely encountered in Colorado and don't appear to be a primary cause of the peak flows above 7500 ft (Jarrett, 1993) or the largest runoff events below 7500 ft (N. Doesken, Colorado State University, pers. comm., 2000), the effect of forest management on rain -on -snow events will not be considered further. Some forest hydrologists have suggested that watershed - scale increases in the size of peak flows can be exac- erbated or ameliorated by harvesting different portions of the basin to alter the timing of runoff from different tributaries (e.g., Harr, 1981). The argument is that for- est harvest close to the mouth of a basin may accelerate runoff and help desynchronize the runoff from the lower portion of a basin relative to the upper portion. Alterna- tively, forest harvest in the upper portions might acceler- ate the delivery of water to downstream areas and further increase peak flows by synchronizing the peak flows from the upper and lower portions of a basin. Research in Colorado and other snowmelt- dominated areas indicate that forest harvest usually has little or no effect on the timing of the peak flows. Forest har- vest caused no significant change in the timing of the snowmelt peak at Wagon Wheel Gap, Deadhorse North, Deadhorse South (Troendle and King, 1987; Troendle, 1987a), or Coon Creek (Troendle et al., 2001). In con- trast, the annual maximum peak flow on Fool Creek occurred an average of 7.5 days earlier (Troendle and King, 1985). The earlier peak at Fool Creek is attributed to the faster and earlier melt in the lower parts of the catchment, and this superimposed an earlier and sharper peak on a formerly flat or bimodal snowmelt peak (Fig- ure 2.3). Overall, there is little evidence to suggest that forest harvest can consistently and significantly affect the timing of peak flows in Colorado. Hence the issue of synchronization appears to be of little practical signifi- cance, and will not be considered further. 2.2.6 Effects of vegetation change on low flows. In contrast to changes in the size of peak flows, an in- crease in the size of low flows is generally considered beneficial, as the additional water may be useful for water supply purposes, increasing instream habitat, or improving water quality (MacDonald et al., 1991). Stud-