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

Home Browse Search

but significant water repellency was found to a depth of just over two inches. The strength of the water repellent layer generally increased with increasing percent sand, and water repellency generally was reduced when gravi- metric soil moisture exceeded 15 -25% (Huffman et al., 2001; MacDonald and Huffman, in review). Repeated measurements on the Bobcat Fire southwest of Fort Col- lins have shown a progressive weakening of soil water repellency 3 and 12 months after burning, and this is consistent with the weak water repellency found in the Crosier Mountain fire 22 months after burning (Huffman et al., 2001; MacDonald and Huffman, in preparation). The tremendous spatial variability in soil water repel- lency after fires makes it difficult to accurately determine the magnitude and distribution of soil water repellency, or to predict the likely effects on runoff at the watershed scale. This work and other studies indicate that summer con- vective storms in the first two, or possibly three, years after burning pose the greatest risk for post -fire flooding and erosion (Moody and Martin, 2001; MacDonald et al., 2001b). The combination of high precipitation inten- sities, high percent bare soil, and a strong water repellent layer can increase runoff rates from severely -burned areas by one or two orders of magnitude (i.e., 10 -100 times) relative to unburned areas. Such increases in sum- mer storm runoff have been documented for the 1994 South Canyon fire on Storm King Mountain in western Colorado (Cannon et al., 1998), the 1996 Buffalo Creek fire southwest of Denver (Jarrett and Browning, draft manuscript), and the Bobcat fire southwest of Fort Col- lins (Lange, 2001; Kunze, 2003). Other wildfires in Colorado and the Rocky Mountains have not resulted in as large a change in runoff and ero- sion rates. For example, the 1994 Hourglass fire near Pingree Park did not appear to have the same effect on runoff as the Bobcat or Buffalo Creek fires. Earlier stud- ies in the same area noted little or no changes in runoff after the Comanche wildfire (Delp, 1968; Meyers, 1968). Post -fire runoff from a basin burned in the 1988 Yellow- stone fire was consistent with the change expected from forest harvest (Troendle and Bevenger, 1996) rather than the sharp increase in peak flows that were observed after the South Canyon, Buffalo Creek, and Bobcat fires. The absence of detailed data from most of these sites makes it difficult to determine why there have been such differences in response, but one possible explanation is the relative likelihood of high- intensity rain storms. At lower elevations in the Front Range most of the recent large fires have been followed by relatively extreme precipitation events, while comparable events appar- ently did not occur after the Hourglass or Comanche fires. Some atmospheric scientists have suggested that large burned areas affect local atmospheric conditions and thereby facilitate the development of high- intensity convective storms. At higher elevations there is a lower likelihood of high- intensity rainstorms. The possible increase in the frequency and magnitude of convective rainstorms after large fires is a topic that should be inves- tigated further, as the empirical evidence indicates that relatively large storm events have occurred after each of the recent large fires in the Front Range except for the first year after Hayman fire, when there was an unusually severe drought. It is important to recognize that because forest soils have high infiltration rates, a post -fire water repellent layer does not have to be completely eliminated before it becomes hydrologically ineffective. Prescribed fires may have patches burned at moderate or high severity that are strongly water repellent, but these patches are usually limited in size and spatially discontinuous (Huffman et al., 2001). Much or all of the runoff from these water repellent areas is likely to infiltrate further downslope, so the increase in runoff at the watershed scale should be much less for prescribed fires than large wildfires. 18 In summary, high- severity fires in forested areas can greatly increase runoff rates. The effect of low- severity fires on runoff generally is consistent with the effects of forest harvest on runoff, as low- severity fires do not substantially alter the runoff processes and pathways. Detailed, process -based studies need to be implemented immediately after severe fires in order to better docu- ment the magnitude and causes of changes in runoff, and better predict the risk to downstream areas. The observed increase in peak flows after the South Canyon, Buffalo Creek, and Bobcat fires confirm that high- severity wild- fires can greatly alter the basic rainfall -runoff response, and there is a need to determine whether similar changes can be expected in other areas of Colorado. 2.5. Threshold of Response, Spatial Scaling, and the Detectabili1y of a Change in Flows Most of the information in the previous sections was derived from plot or small watershed studies. From a management perspective, however, the question is whether these results can be extrapolated to larger -scale basins. There are effectively four components to this is- sue, and these are: (1) How much of a forested basin has to be treated in order to generate a detectable change in runoff? (2) Will a given change in runoff be translated