Introduction

Shelterwood cutting and small patch clearcutting have been proposed to overcome problems of reforestation in the mixed-conifer zone of southwestern Oregon. In 1962, a study was begun to determine changes in quantity and quality of streamflow following these two harvest methods and to compare them with changes caused by complete clearcutting. Specifically, the objectives of this study are to determine increases in size of annual yield, seasonal yield, and peak streamflow after timber harvest. This report describes changes in streamflow following the first of three phases of timber harvest. Water quality changes will be described in a subsequent report.

That cutting forest vegetation increases streamflow has been known for some time, although the size and duration of increases are known for only a few local areas. Taken collectively, however, water yield improvement studies indicate the range in size of increases that can be expected in various regions (Hibbert 1967). In the maritime climate of western Oregon, two such studies have shown that clearcut logging of entire small watersheds in mountainous topography of the Coast and western Cascade Ranges can cause absolute increases in annual yield that are among the largest in the world (Rothacher 1970, Harris 1973, Harr 1976). Increases in size of peak flows have also been related to timber harvest activities in western Oregon (Rothacher 1973, Harris 1973, Harr et al., 1975). Changes in streamflow have been due to both reduced evapotranspiration which makes more water available for streamflow, increased overland flow caused by soil compaction during roadbuilding, logging, and slash disposal, and interception of subsurface flow by roadcuts and ditches.

Mixed conifer forests cover approximately half of the five-county area making up southwestern Oregon and also extend well into northern California. Old-growth forests contain valuable timber which supports the most economically important industry in the region. These watersheds also yield a major portion of irrigation water for agriculture, the region's number two industry. Streams are also important as spawning and rearing areas for anadromous and resident fish. Natural, low summer streamflow, the demand for irrigation water, and runoff-erosion relationships have made forest-streamflow relations of prime interest and concern to land managers and water resource planners in this region (Hayes 1959, Hayes and Herring 1960).

Study Area

WATERSHED CHARACTERISTICS

The four Coyote Creek Experimental Watersheds are located about 55 km southeast of Roseburg, Oregon, at the head of Coyote Creek, a small tributary of Buckeye Creek which flows into the South Umpqua River. Watersheds range in size from 48.6 to 69.2 ha and generally have welldefined boundaries (fig. 1). Watershed aspect ranges from east-northeast for CC-1 (Watershed 1) to north for CC-3. Elevation of the watersheds ranges from 730 to 1 065 m above mean sea level.

The watersheds are underlain by the Little Butte Formation, rhyodacitic pyroclastic rocks consisting of welded and nonwelded ash-flow tuffs with andesite and basalt common on ridges (Kays 1970). Although many slopes are relatively smooth, some have poorly developed external drainage patterns that attest to past and present mass erosion processes (Swanson and Swanston 1977). Ranging from 20 to 80 percent, slope gradients of the experimental watersheds are typical of much of the surrounding region.

Dumont

Two soil series occupy most of the study area (Richlen 1973). soils are moderately permeable, well-drained gravelly loams at least 150 cm thick, and derived from reddish breccia parent material. Straight soils are similar to Dumont soils but are only 50-100 cm thick. Surface soils of both series have relatively high permeabilities but may be underlain by denser soil layers that impede vertical movement of soil water.

The study area is in the mixed conifer zone (Franklin and Dyrness 1973). Here, Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) of the more mesic regions to the north and west is intermingled with ponderosa pine (Pinus ponderosa Laws.), sugar pine (Pinus lambertiana Dougl.), and

incense-cedar (Libocedrus decurrens Torr.) characteristic of warmer, drier
sites. Within the watersheds, more mesic habitats contain western hemlock
(Tsuga heterophylla (Raf.) Sarg.), grand fir (Abies grandis (Dougl.) Lindl.),
and big leaf maple (Acer macrophyllum Pursh). Before logging, both age
class and overstory density varied considerably within a watershed as well
as among watersheds.

CLIMATE AND STREAMFLOW

Table 1--Annual precipitation at the

The climate of the study area is influenced by the Pacific Ocean 150 km to the west. Annual precipitation at the climatic station adjacent to CC-2 has averaged 123 cm, ranging from 88 cm to 156 cm (table 1). Winters are cool and wet, and summers are warm and dry. Approximately 89 percent of annual precipitation occurs in the October-March period during long-duration, low intensity frontal storms associated with cyclones which originate over the Pacific Ocean. Locally, precipitation caused by frontal activity is augmented by orographic precipitation when moist, unstable air masses move inland. Average daytime temperatures range from -0.5° to 1-5°C in winter and 15° -20°C in summer (fig. 2).

Year

climatic station adjacent to
CC-2 and streamflow at control
watershed CC-4.

Precipi- Streamflow
tation
at CC-4

Cm

Although most precipitation occurs as rain, snow is common, particularly at higher elevations. Occasionally a snowpack may remain for 1-3 months; but in most years, it usually melts within 1-2 weeks. The highest runoff in this region, as in the rest of the Pacific Northwest, has resulted from rapid snowmelt during prolonged heavy rainfall (Waananen et al. 1971, U.S. Army Corps of Engineers 1975).

Streamflow has been measured continuously with 120-degree sharpcrested V-notch weirs (fig. 3) since December 1963. (Streamflow measurements began at CC-1 and CC-2 in 1962, but data prior to December 1963 have been excluded from this study.) Because weir ponds were filled with sediment at CC-1, CC-3, and CC-4 between December 21, 1964 and March 19, 1965, five peak streamflows could not be measured accurately during this period. Data for this period has been excluded from all analyses. An October 1-September 30 water year has been used throughout.

During the 1966-1976 period, annual streamflow at CC-4, the control watershed, averaged 63 cm, ranging from 18 to 107 cm. Annual streamflow at this watershed has varied from 21 percent of annual precipitation during 1968, which had the least precipitation of the 11-year period, to about 70 percent during 1972 and 1974, two of the wettest years of the period. Apparent annual evapotranspiration, as estimated from annual precipitation minus annual streamflow adjusted to a unit area basis, has averaged about 61 cm, ranging from 50 cm in 1974 to over 70 cm in 1966. Maximum instantaneous streamflow has been over 4,000 times greater than minimum streamflow.

Watershed Treatments

In this study we used the paired watershed technique. With this technique, a hydrologic variable in one watershed is compared with that in another watershed during the calibration or pretreatment period to establish a relationship between watersheds over a range of climatic conditions. Then one watershed of a pair is treated or altered in some way while the other is left as an undisturbed control. Post-treatment measurements of