The Caspar Creek watershed studies: long-term research in a temperate, rain-dominated forest

Abstract

In 1962, the U. S. Forest Service and the California Department of Forestry and Fire Protection established the 8.9-km Caspar Creek Experimental Watersheds in northwest California, United States, to study the influence of logging on flow, sediment, and aquatic habitat. Caspar Creek is located on uplifted marine terraces carved from Paleocene to Eocene sandstone and shale. Average rainfall is 1170 mm/yr, with 95% falling from October through May. Old-growth redwood forests were logged between 1860 and 1905, and by 1962 the watersheds supported mature second-growth stands dominated by coastal redwood (Sequoia sempervirens) and Douglas fir (Pseudotsuga menziesii). The first watershed-scale experiment evaluated effects of tractor-yarded logging in the 4.2-km South Fork watershed using the 4.7-km North Fork as a control. Flow, suspended sediment concentration, and pond infill were monitored at weirs constructed in 1962 in each watershed. The calibration period ended in 1967 with riparian road construction along the mainstem, followed by 4 more years of monitoring before the watershed underwent a 61% volume selection cut in 1971-1973. Lags-to-peak decreased slightly after road construction, and dry-season flows appear to have increased slightly. Other hydrologic responses were not evident; however, sediment loads increased markedly, in part from failure of an old crib dam below a road crossing. After logging, water yield increased 15 %, with 90% of the increase occurring during the wet season. Yield had not fully recovered by 1985, when the experiment was terminated by the initiation of logging within the control watershed. Dry-season flows also increased and then reattained pretreatment values in about 8 years. Initial conflicting results for peakflows were found simply to reflect differing analysis periods: large winter peaks increased an average of 13 % for 8 years after logging, and lags-to-peak decreased. A preliminary reanalysis of sediment loads suggests that sediment yield during and after logging averaged 2-3 times that expected on the basis ofpretreatment regressions. Sediment yield reattained pretreatment levels by 1980. The second experiment was designed to quantify cumulative impacts on downstream sediment and flow from cable-yarded clearcut logging in the North Fork watershed. In 1985, stream gages were constructed on 8 tributaries and at 5 additional downstream sites, allowing responses to be tracked downstream. In 1985-1986, 13 % o f the downstream watershed was logged; this was not considered part o f the treatment. About 39% o f the watershed was then logged in 1989-1992, with 3 tributaries left as controls. All gages were monitored until 1995, and monitoring continues at 8. Ancillary studies were carried out to eva1uate mechanisms for observed changes (ZIEMER 1998). Abou t 22 %o f rain was foun d to be intercepted by forest canopies even during large storms. Reduced transpiration and interception after logging are sufficient to explain the 27% increase in large winter peakflows observed in clearcut tributaries (REID & LEWIS 2007). Average water yield at the downstream weir increased 15 % over that expected for 1989-1997, and water yields and minimum flows remain higher than expected as o f 2006. Suspended sediment yield increased an average of 89% at the weir during 1989-1995, with recovery to near background levels in about a decade. Four of the 5 fully logged tributaries showed increases of 120260 %, while the fifth showed a decrease, probab1y caused by deposition associated with near-channel b1owdown. Downstream reaches also showed aggradation associated with logs blown down from buffer strips. Landsliding contributed some o f the increased sediment yield, but nested gages revealed that some new sediment originated from unlogged reaches downstream oflogged areas, probably due to acce1erated gully erosion from increased runoff.