United States Department of Agriculture
Pacific Southwest Research Station
General Technical Report
On SFC, the minimum summer discharge (instantaneous mean daily flow) increased an average of 38 percent or 0.25 L s-1 km-2 between 1972 and 1978. The maximum increase was 0.42Ls-1 in 1973, the final year of timber harvesting on this watershed. No increases were detected in 1977, the driest year of record. Summer discharge minimums returned to prelogging levels beginning in 1979(table2).
The NFC response to timber harvesting operations yielded increased minimum summer discharges, as well. Figure 2 compares changes in minimum summer discharge levels for SFC and NFC. Even after clearcutting only 12 percent of the watershed, a 161 percent (0.30 L s1 km-2) increase was detected in 1986. Between 1990 and 1997, minimum summer discharge averaged 148 percent (0.40 L s1 km-2) greater than preharvest predictions. The minimum increase was 75 percent in 1990 and 1996. The largest relative increase was 287 percent in 1992, but the maximum absolute increase, 0.67 L s1 km-2, occurred in 1997 (table 2).
Figure 2 Minimum summer discharge at the North Fork (NFC) and South Fork (SFC) weirs. Departures are relative to predictions using 1963-1971 calibration regression. SFC was roaded in 1967 and was logged in 1971-1973. NFC was logged in 1985 and 1989-91.
L s1 km-2
Additional indications of summer flow increases after the NFC logging were observed further up the watershed. Minimum summer stages in the tributary basins are shown in figure 3. Before logging, all of the sub-basin gaging stations were seasonally dry, with the exception of the three mainstem rated sections (ARF, FLY, and LAN) and three tributary stations (DOL, IVE, and KJE). Even with higher rainfall after the NFC logging, this pattern of no flow continued at the control gages, HEN and MUN. However, of the remaining five gaging sites known to cease flowing during the summer season before logging (BAN, CAR, EAG, GIB, and JOH), all had surface flow well into the summer period after logging. Instrumentation problems prevented analysis of the GIB and IVE data. For the remaining tributary sites, the mean postlogging minimum stage was greater than the predisturbance value according to both the ttest and the Mann-Whitney test statistics (table 3).
Figure 3 Minimum summer stages at clearcut tributary gages (BAN, CAR, EAG, KJE) and unlogged control gages (HEN, MUN). Positive stage values indicate surface flow persisted for the entire water year. Negative values indicate the minimum intergravel flow level measured at the instream well tube. Arrows indicate time of logging.
Large increases in late-July soil pipeflow from the 100 percent clearcut K2 swale were detected in 1990 to 1992, and 1994 (fig. 4). Increases in this mid-summer pipeflow variable averaged 179 percent or 0.45 L min1. The largest increase, 478 percent or 0.95 L min-1, occurred in 1991. After 1994, K201 pipeflow fell below the calibration prediction, but this was not a statistically significant decrease (table 4). Because the range of the pretreatment calibration (1987-1989) only extended to about 0.3 L min-1, observations during 1990, 1993, 1995, and 1996 required extrapolation of this calibration and the departures are suspect. However, it appears that summer pipeflow increases were greater the first and second year (1990 and 1991) after logging and returned to near pretreatment levels by 1993 or 1994.
Figure 4 Summer (late July) pipeflow measured at the pipe K201 (K2-site logged 1989) and the control pipe M106 (M1-site untreated). Calibration regression (solid line) uses years 1987-1989. Dashed line is extrapolation of regression line beyond calibration data.
|100 percent clearcut subwatersheds||Control subwatersheds|
After logging, tensiometer gages in watershed KJE indicated reduced tensions (higher soil moisture) on July 1. For example, at a depth of 150 cm at site C4, tensions on July 1 were 55 and 32 cb for the 2 years before logging. After logging on July 1 tensions did not exceed 10 cb. Similarly, maximum tensions at this site were 84 cb and 74 cb before logging, but never exceeded 60 cb in the postlogging monitoring period (fig. 5). In other words, postharvest soil moisture levels throughout the summer dry season were at or above the preharvest levels. This suggests additional soil moisture was available to support forest regrowth. Unfortunately, much of the tensiometer data was interrupted or discontinued after 1993 because of equipment problems and vandalism. It would be interesting to see if the soil moisture tensions have returned to prelogging levels as apparently the pipeflow has done.
Figure 5 Soil moisture tension at 150 cm depth at the K2-site before and after the August 1989 logging.
These data clearly indicate that streamflow was enhanced after timber harvest operations in the NFC and SFC. Significant increases were detected in both summer flows and annual water yield. Previous analyses (Keppeler and Ziemer 1990, Ziemer and others 1996) have shown similar results although the stated magnitudes and duration of postlogging flow enhancement have varied according to the definitions of the streamflow variables and calibration years, and the length of the data set.
As previously noted (Henry, these proceedings) the Caspar Creek watersheds receive annual precipitation averaging 1,200 mm (45 inches). However, not all of this precipitation is routed to the channel as streamflow. Perhaps 10 to 15 percent of this precipitation is intercepted by the forest canopy and evaporated back into the atmosphere without contacting the ground. A minor amount of precipitation may percolate through the entire soil profile to bedrock fissures and into deep groundwater supplies, thus escaping our stream gaging instrumentation. Less than 50 percent of annual rainfall is measured as streamflow at the North and South Fork weirs (Ziemer 1997). Most of the remainder infiltrates into the soil where it is eventually absorbed by plant roots and transpired by forest vegetation.
It has long been acknowledged that evapotranspiration dominates the water balance over most of the landmass. Potential evapotranspiration (PET) is the return of water to the atmosphere by vegetation unrestrained by soil water limitations. Using the Thornthwaite method, Ziemer (1997) calculates annual PET at Caspar Creek to be 660 mm or about half of annual rainfall. During the months of May through September, water demand by the vegetation (or PET) far exceeds rainfall inputs; thus stored soil moisture is depleted. As soil water is depleted by subsurface drainage and plant use, actual evapotranspiration is greatly reduced. Removal of vegetation by timber harvest, forest fire, or other means will most heavily affect the evapotranspiration component of this water balance.
In light of this, the streamflow enhancements observed at Caspar Creek are not unexpected. In reviewing the results of 11 watershed studies in the Pacific Northwest, Harr (1979) reported increased annual water yields of up to 620 mm. At Caspar Creek, the increase was far more modest, averaging less than 100 mm. This is not surprising given that absolute increases due to timber harvest are largest in high rainfall areas and Caspar Creek is at the southernmost extension of this region. If we accept that the PET of the fully forested NFC watershed was 660 mm, then logging 50 percent of the timber volume could reduce plant water use by about half, at most (a savings on the order of 300 mm), under conditions where soil moisture is unlimited. However, actual evapotranspiration from this site is certainly limited by soil moisture levels during the dry summer months, thus lesser enhancement is expected. The average postlogging increase at Caspar Creek was less than one third of the potential amount. Cutting trees reduces, but does not eliminate transpiration. Residual and new understory vegetation in the cutblocks probably responded to increased soil moisture by increasing transpiration. Evaporation from exposed hillslope surfaces continued, and possibly increased, after logging. Also, at NFC, the riparian vegetation was mostly retained. Riparian trees and vegetation use more water than the forested hillslopes.
Harr's review also reports that summer flows elsewhere have been as much as quadrupled. After the NFC logging, minimum summer discharge almost tripled during one postlogging year, but averaged about 150 percent after clearcutting 50 percent of the watershed area. The largest increases noted in Harr's analysis occurred where complete clearcutting was done. At the fully clearcut K2-site on the NFC, the magnitude of summer pipeflow increases were similar to the maximums that Harr reports. The largest relative increases in summer minimum flows occurred after Caspar Creek's two driest rainfall years on record, 1977 and 1991.
Why were the summer flow increases larger and more persistent at NFC compared to SFC after logging? The harvest prescription appears to be the important factor. In the SFC selection cut, single trees and small groups were logged, leaving a dispersed residual timber stand and understory vegetation ready to take advantage of any soil moisture surpluses occurring after harvest. In the NFC clearcuts, soil water surpluses occurred in discrete cutblocks where only peripheral contact with residual vegetation existed. New vegetation must grow, or, in the case of redwoods and some hardwood species, resprout, and develop ample foliage to make full use of stored soil water. The postharvest site treatments may also play a part in the magnitude and duration of flow enhancements. About 20 percent of the NFC watershed area was broadcast burned after harvest. One half of the cutblocks were burned approximately one year postharvest, thus setting back the regenerating vegetation. In addition, these burned units were later treated with herbicide to control competition from brush species, also setting back the recovery of the vegetation. The YZ 1985 timber sale was precommercially thinned in 1995, once again pushing back the recovery of that portion of the NFC watershed. Additional precommercial thinning is planned for other NFC cutblocks beginning with Unit K this year (1998). Although the NFC cutblocks are now green and the regeneration appears to be growing vigorously, water use by the new vegetation appears to remain far below that of the older second-growth forest in the NFC watershed. Evapotranspiration is a function of total leaf area. Leaf area in the revegetating cutblocks remains far below that of the uncut portions of the watershed.
What is the role of fog and fog precipitation in the postlogging water balance at Caspar Creek? Literature suggests that fog plays a crucial role in the ecology of the Pacific Northwest. In this region, warm, moist air contacts cool coastal waters, lowering temperatures below the dew point and forming fog. This fog layer may travel far inland depending on the strength of the onshore breeze and local topography. The Coast Range forms a partial barrier to this marine layer, preventing penetration to inland areas except where breaks in topography occur such as along river valleys. Fog dissipates when sufficient warming of the airmass occurs to raise temperatures and re-evaporate the fog droplets. When summer fog blankets the forest, relative humidity is increased while insolation and temperature are decreased, thus reducing transpiration by the vegetation. Summer fog influences the species composition of the coastal forest. Lacking stomatal control, the coast redwood is limited to a narrow belt along the California and southern Oregon coastlines, in large part because of the prevalence of summer fog.
Fog precipitation, or fog drip, occurs when fog droplets encounter an obstruction, coalesce, and fall to the ground. This phenomenon is largely limited to exposed ridges or crests during periods of cool temperatures less than 10 C (Freeman 1971). The redwood, Douglas-fir, and spruce forest canopy is particularly efficient at intercepting water droplets and inducing fog drip. Kittredge (1948) reported that 285 mm of fog drip was collected during one summer season under an 85-year-old spruce/hemlock stand in coastal Oregon. Azevedo and Morgan (1974) reported seasonal fog drip totaling as much as 425 mm in northern California's Eel River valley. More recently, Dawson (1996) concluded that 8 to 34 percent of the water used by the coast redwood and 6 to 100 percent of water used by understory vegetation originated as fog precipitation at his study site on a hillslope near the mouth of the Klamath River. Ingraham (1995) analyzed fog, rain, and groundwater samples from the Point Reyes Peninsula and found that the isotopic composition of groundwater reflects the contribution of fog water.
Fog drip has not been measured at Caspar Creek. However, field observations suggest that it does occur. The frequency of daytime fog can be discerned from solar pyranometer data collected at a Caspar Creek site since 1988. Summer fog within the experimental watershed area is far less frequent than in coastal Mendocino County because of the more inland location of our study site. Along the coast, 30 to 50 percent of days during June, July, and August have morning fog (Goodridge 1978). At Caspar Creek, solar radiation data collected between 1988 and 1994 indicate that only 10 to 35 percent of the June-through-August days have insolation reduced by more than about one-third because of fog or cloud cover.
Did logging reduce fog drip? The removal of the forest canopy, especially near the ridges, probably resulted in less fog interception and drip. The pipeflow swales are located near the ridge in the NFC headwaters. Here, one might expect fog drip to play a more prominent role in the water balance than in the watershed overall, but July pipeflow increased dramatically during the first few postlogging seasons, suggesting that this was not the case. July measurements for 1995 and 1996 at K201 were below the predicted levels, but this reduction was not statistically significant. At some point during postharvest recovery, one might expect that evapotranspiration rates will return to preharvest levels while the forest canopy is not yet functioning as an effective fog drip collector. This has yet to be documented and thus remains hypothetical. Because these finer nuances of postlogging recovery will be difficult to detect at the weirs before additional timber harvest commences, continued evaluation of the postlogging summer discharge trend at the pipeflow sites is warranted.
Perhaps the smaller postharvest streamflow increases observed on SFC relate, in part, to this loss of fog drip. It is quite plausible that SFC receives more fog than NFC because of its proximity to the coast, but this has not been documented. What has been documented is a measurable postharvest increase in streamflow at both the NFC and SFC weirs, as well as increased soil moisture and pipeflow in the NFC watershed. If fog drip were an important component of the hydrology of Caspar Creek, we would have seen a decrease in soil moisture, pipeflow, and streamflow in the cut units, not the increase reported here.
The impacts of the Caspar Creek harvest treatments on stream and riparian ecology are more difficult to discern than the physical changes. An increase in summer discharge implies that the stream is less susceptible to water temperature increases. Maximum water temperatures increased about 9 C (from 16 C to 25 C) after right-of-way clearing and road-building in the SFC riparian zone (Krammes and Burns 1973). Increased summer flows did not buffer these temperature effects. On the NFC, stream temperature changes after logging were not significant (Cafferata 1990, Nakamoto, these proceedings). The use of stream-side canopy retention zones on Class I and II channels was probably far more important in preventing increases in strea temperature than the summer streamflow enhancement.
Perhaps a more important effect of enhanced summer discharges is the increase in aquatic habitat developed in the channel. Higher discharge levels increased habitat volumes, and, as witnessed at the tributary gages, lengthened the flowing channel network along logged reaches. Nakamoto (these proceedings) concludes that the amount of slow water habitat on the NFC increased after logging, but reports no corresponding increase in biomass of stream vertebrates.
In terms of both stream temperature and habitat availability, the summer flow enhancements are of greater importance than the increases in total annual water yield because it is during the summer streamflow recession that temperature and habitat carrying capacity are most critical. However, these discharge impacts are variable and relatively short-lived. Impacts of timber harvest associated with effects other than summer flow increases appear to be of greater significance to the ecology of Caspar Creek, as the other reports in these proceedings illustrate.
The Caspar Creek watershed studies reveal increased water yields and summer flows after timber harvesting. Increases observed at this site were more modest than those documented at other sites in the Pacific Northwest. Streamflow changes due to logging were most evident during the long, dry summer season. During this prolonged recession, zones of deep perennial saturation maintain streamflow (baseflow). After logging, reduced evapotranspiration allows for additional water to be stored and routed to streams as summer streamflow. At Caspar Creek, enhanced soil moisture in the rooting zone followed timber harvest in the North Fork clearcut units. Previously intermittent stream reaches and soil pipes became perennial. The larger increases in minimum flows observed on the North Fork are probably due to wetter soils in the clearcut units where minimal vegetation exists to use this enhanced moisture. On the South Fork, older second-growth residual forest vegetation more readily exploited this additional soil moisture.
Fog plays an important role in the regional ecology by moderating evapotranspiration. However, Caspar Creek data indicate that any possible postlogging loss of fog drip does not result in a net reduction in streamflow. Moisture savings due to reduced evapotranspiration appear to override fog precipitation losses at this site.
Continued monitoring will document the duration of summer flow increases due to the most recent North Fork logging only as long as additional harvest operations are postponed in both the North and South Fork watersheds. Fortunately, the M1 and K2 pipeflow sites provide the opportunity for continued evaluation of the effects of clearcutting on the baseflow processes without the complications caused by further harvest operations in the greater watershed area. Quantification of fog drip within the Caspar Creek watershed warrants investigation in light of the documented importance of this moisture source at some sites in the Pacific Northwest.
Azevedo, J.; Morgan, D.L. 1974. Fog precipitation in coastal California forests. Ecology 55: 1135-1141.
Cafferata, Peter H. 1990. Temperature regimes of small streams along the Mendocino coast. Jackson Demonstration State Forest Newsletter, No. 39, October 1990. p. 4-7.
Dawson, Ted E. 1996. The use of fog precipitation by plants in coastal redwood forests. In: LeBlanc, John, ed. Conference on coast redwood forest ecology and management; 1996 June 18-20; Arcata, CA. Berkeley, CA: University of California; 90-93.
Freeman, Gary J. 1971. Summer fog drip in the coastal redwood forest. Arcata, CA: Humboldt State Univ.; 101 p. M.S. thesis.
Goodridge, James D. 1978. Climate summaries of California airports. Sacramento, CA: California Dept. of Water Resources.
Harr, R. Dennis. 1979. Effects of streamflow in the rain-dominated portion of the Pacific Northwest. In: Proceedings of workshop on scheduling timber harvest for hydrologic concerns; 1979 November 27-29. Portland, OR: Pacific Northwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 45 p.
Harr, R. Dennis. 1982. Fog drip in the Bull Run Municipal Watershed, Oregon. Water Resources Bulletin 18(5): 785-789.
Hibbert, A.R. 1967. Forest treatment effects on water yield. In: Sopper, William E.; Lull, Howard W., eds. Forest hydrology. Proceedings of a National Science Foundation advanced science seminar; 1965 August 29 - September 10; University Park, PA. New York: Pergamon Press; 527-543.
Ingraham, N.L. 1995. The importance of fog-drip water to vegetation: Point Reyes Peninsula, California. Journal of Hydrology 164: 269-285.
Ingwersen, J.B. 1985. Fog drip, water yield, and timber harvesting in the Bull Run municipal watershed. Water Resources Research 21(3): 469-473.
Keppeler, Elizabeth T.; Ziemer, Robert R. 1990. Logging effects on streamflow: water yields and summer low flows at Caspar Creek in northwestern California. Water Resources Research 26(7): 1669-1679.
Kittredge, Joseph. 1948. Forest influences. Dover, NY: McGraw-Hill Book Co.; 418 p.
Krammes, J.S.; Burns, D.M. 1973. Road construction on Caspar Creek watersheds a 10-year progress report. Research Paper PSW-93. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, USDA Forest Service; 10 p.
Rothacher, Jack. 1971. Regimes of streamflow and their modification by logging. In: Proceedings of the symposium on forest land use and stream environment; Corvallis, OR. Corvallis, OR: Oregon State University, 55-63.
Ziemer, Robert R.; Lewis, Jack; Keppeler, Elizabeth T. 1996. Hydrologic consequences of logging second-growth watersheds. In: LeBlanc, John, ed. Conference on coast redwood forest ecology and management; 1996 June 18-20; Arcata, CA. Berkeley, CA: University of California; 131-133.
Ziemer, Robert R. 1997. Caspar Creek Thornthwaite potential evapotranspiration, water years 1990-1995. Website http://www.fs.fed.us/psw/rsl/projects/water/Thornthwaite.html; 1 p. preventing increases in stream temperature than the summer streamflow enhancement.