RESULTS OF ANALYTICAL AND 

 FIELD STUDIES 



Piezometric Responses of Slopes 

 to Vegetation Removal 



A number of hydrologic processes are regulated by vegeta- 

 tion on a forested slope. For the present study area, important 

 processes are snow accumulation and melt and factors affect- 

 ing total evaporation losses including interception, transpira- 

 tion, and direct evaporation from the soil. These processes in 

 turn influence slope stability by helping to regulate the water 

 content of the unsaturated soil moisture zone and the depth of 

 any underlying piezometric surface. The net effect of vegetation 

 removal is to reduce evapotranspiration losses and to increase 

 snow accumulation and snowmelt rates thereby creating a 

 tendency to raise the piezometric surface. 



Increased soil moisture contents at the end of the growing 

 season document the fact that evapotranspiration losses were 

 reduced following clearcutting and fire on the study watersheds. 

 Also snow survey and snowmelt lysimeter data verify increased 

 springtime snow accumulation and snowmelt rates following 

 timber removal (Megahan, unpublished data). The combined 

 effect of an increased snowpack melting faster in conjunction 

 with a greater carryover of fall soil moisture storage causes an 

 increase in peak springtime piezometric depths compared to 

 the forested condition (fig. 22). The piezometric data shown are 

 the annual peak values for the piezometer on each watershed 

 that registered the greatest water depth throughout the entire 

 study period. The snowpack data are from the nearby Silver 

 Creek study area that remained undisturbed throughout the 

 course of the present study. 



X Watershed no. 1 (piezometer no. 51 

 • Watershed no. 4 (piezometer no. 3) 

 o April 1st Water equivalent (in) _ 



YEAR 



Figure 22. — Annual peak piezometric response for peak sample 

 points on Watersheds No. 1 and No. 4 and corresponding snow 

 accumulations. 



The consistently greater piezometric responses on Water- 

 shed No. 1 as compared on Watershed No. 4 reflect the greater 

 potential drainage area for piezometers on No. 1 relative to 

 No. 4. Water depths on both drainages decreased from 1 971 to 

 1972 in response to a smaller snowpack in 1972. The 1973 

 snowpack was unusually low (there was a 7 percent chance of 

 having a snow water content equal to or less than the recorded 

 value) and the piezometric levels on the undisturbed Watershed 

 No. 4 dropped accordingly. 



A similar trend in piezometer levels did not occur, however, 

 on Watershed No. 1 , which had been clearcut logged the pre- 

 vious November. Rather, the maximum piezometer level 

 actually exceeded those recorded in 1972 when snow water 

 contents were much greater (there was an 81 percent chance of 

 having a snow water content equal to or less than the recorded 

 value in 1972). The late fall cutting date for Watershed No. 1 

 prevented the development of differences in growing season 

 soil moisture storages on the two study watersheds; therefore, 

 the increased piezometric levels on Watershed No. 1 are mainly 

 the result of differences in snow accumulation and melt during 

 the previous winter and spring. 



The following summer, the total soil moisture storage on 

 Watershed No. 4 decreased more rapidly than on Watershed 

 No. 1 because of reduced evapotranspiration losses caused by 

 the timber removal on Watershed No. 1 . By mid-August of 1 973, 

 49 percent more water was stored in the 12- to 60-inch (30- to 

 152-cm) depth on Watershed No. 1 (a total of 5.8 inches [14.8 

 cm] on Watershed No. 1 and 3.9 inches [9.9 cm] on Watershed 

 No. 4). On August 20, 1973, both watersheds were burned in a 

 wildfire that killed all nonsprouting vegetation. In the spring of 

 1974, piezometric levels were greatly increased on Watershed 

 No. 1 relative to predisturbance levels even though the April 1 

 snow accumulation was less than the two prelogging years 

 studied. The difference reflects both changes in evapotran- 

 spiration losses and in snow accumulation and melt rates. 

 Piezometric levels also increased on Watershed No. 4, but prob- 

 ably not to the maximum because of the development of a 

 partial soil moisture deficit prior to the fire. April 1 snow depths 

 increased slightly in 1975, relative to 1974, but still were below 

 pretreatment amounts. Large increases in piezometric levels 

 occurred on both study watersheds in response to combined 

 effects of decreased evapotranspiration and snow accumula- 

 tion and melt. 



These data do not lend themselves to rigorous statistical 

 analysis because of limited sample sizes. A simple graphical 

 analysis relating the annual peak piezometer depth to the cor- 

 responding April 1 snow water content is informative, however, 

 because it suggests the relative effects of vegetation removal 

 on piezometric levels (fig. 23). A simple linear regression was 

 fitted to the three predisturbance data points for Watershed No. 

 4. In spite of the limited data points, the regression had an r 

 square value of 0.85 and a standard error of 0.27 feet (0.09 m). 

 The regression coefficient was significantly different from zero 

 at the 85 percent confidence level. By using the snow water 

 contents for 1 974 and 1 975 to predict the piezometric levels for 

 these years and by comparing predicted to measured 

 piezometric values, we can estimate increases of about 65 

 percent and 1 00 percent in soil water levels for 1 974 and 1 975, 

 respectively. A similar analysis is not possible for Watershed 

 No. 1 because there are only two pretreatment data points. If the 

 assumption is made, however, that the regression slope is the 

 same as for Watershed No. 4 and that the regression line 



14 



