Sediment Dynamics and Sources in a Grazed Hardwood Rangeland Watershed

From 1994 to 1998 we documented sediment transport dynamics and sources in a 137 ha grazed hardwood rangeland watershed on granitic soils at the San Joaquin Experimental Range in Madera County. Sediment transport for this watershed was determined by measuring total suspended solids, bedload and flow at an H-flume installed in 1994. Sediment movement as bedload is the primary means of sediment transport in this watershed, with minimal transport of suspended solids. This is attributed to the large sediment particle size and low stream power characteristic of this low gradient intermittent watershed. Bedload transport can be predicted by stream flow (p<0.0001, R=0.68). Upland sediment sources were surveyed using paired hillslope plots to estimate upland erosion for three slope classes. Sediment traps were installed to compare erosion on cattle trails and adjacent non-trailed areas. Ten stream cross section profiles were averaged to determine grazing treatment and year effects on instream erosion and deposition along three intermittent stream channels. Due to rapid infiltration, runoff and sediment yield from the hillslope, plots averaged only 151 mm and 36 kg/ha, respectively. Unvegetated, disturbed soil surfaces in cattle trails were a significantly greater (p<0.002) source of sediment (0=238 g/trap, n=8 traps) than adjacent well-vegetated soil surfaces (0=6 g/trap, n=8 traps). From 1994 to 1998 stream channel morphological parameters did not change in response to no grazing, winter moderate or concentrated grazing, or dry season moderate or concentrated grazing. There was a year effect on channel depth due to the dynamics of bedload transport in response to variation in peak storm events from year to year. The results of these studies suggest that sediment movement from cattle trails into stream channels is the main grazing-induced sediment source in this watershed. Introduction Most of California’s surface water flows through the State’s 6.8 million ha of annual rangeland. Sediment is the most prevalent non-point source pollutant in these surface waters (State Water Resources Control Board Staff 1999). Causes of erosion within these rangelands include natural processes and historic land use, as well as anthropogenic activities such as road construction and livestock production (Lewis and others 2001). Concerns exist throughout California’s Sierra Nevada Mountains 1 An abbreviated version of this paper was presented at the Fifth Symposium on Oak Woodlands: Oaks in California’s Changing Landscape, October 22-25, 2001, San Diego, California. 2 Range and Pasture Specialist and Rangeland Watershed Specialist, respectively, Agronomy and Range Science, One Shields Ave, Davis, CA 95616-8515 (e-mail: [email protected] and [email protected], respectively). 3 Natural Resources Advisor, Madera County UCCE, 328 Madera Ave., Madera, CA 93637-5498 (email: [email protected]). 4 Watershed/Natural Resource Advisor, San Luis Obispo County UCCE, 1734 Paso Robles Street, Paso Robles, CA 93447 (e-mail: [email protected]). USDA Forest Service Gen. Tech. Rep. PSW-GTR-184. 2002. 65 Sediment in a Grazed watershed—George, McDougald, Tate, and Larsen (Sierra Nevada Ecosystem Project 1996) that livestock grazing increases hill slope and stream channel erosion. Several research and case studies have reported livestock-induced streambank erosion leading to channel down cutting or widening (Hall and Bryant 1995, Kauffman and Krueger 1984, McDonald and others 1991, Sierra Nevada Ecosystem Project 1996). In 1994 we began documenting sediment transport dynamics and sources in a grazed hardwood rangeland watershed on granitic soils at the San Joaquin Experimental Range in Madera County. Site Description The 1,752 ha San Joaquin Experimental Range (SJER) in Madera County, Calif. has been a USDA Forest Service research facility since 1935 (Kie 1990). SJER lies in the lower central Sierra Nevada foothills in the oak savanna vegetation type. California State University at Fresno (CSUF) maintains a herd of 210 commercial beef cows at SJER. In 1994 an H-flume was placed at the bottom of a 137 ha watershed that is drained by an intermittent tributary to Cottonwood Creek. Cottonwood Creek is a fourth-order stream that drains into the San Joaquin River just below Friant Dam. The Station has a Mediterranean climate with annual precipitation ranging from 250 to 800 mm, with a mean of 480 mm, coming almost entirely between October and April. Mean monthly air temperatures range from 6oC in January to 27oC in July. Elevation ranges from 213 m to 518 m. Soils are derived from granitic rocks, and most are less than 0.76 m deep. The Ahwahnee series (coarse-loamy, mixed thermic Mollic Haploxeralf) is common, covering about 96 percent of SJER. The Visalia series soils (coarse-loamy, mixed thermic Pachic Haploxeralf) are found on alluvial or swale sites (Ulrich and Stromberg 1962). Methods Suspended Sediment, Bedload, and Discharge Bedload, total suspended solids (TSS), and discharge were determined at a 90 cm H-flume installed at the bottom of a 137 ha. watershed. During the first year TSS was determined from hand collected samples. Beginning in 1995-96, TSS was determined via vacuum filtration through a 0.45 micron filter using water samples collected hourly with an automatic water sampler. Bedload was estimated during four storms in January and February 1998 and 2000. Due to low rainfall in 1999 and 2001, bedload was not measured. Bedload samples were collected using a Helly-Smith bedload sampler (Shen and Julien 1992) to collect three 1-minute samples every hour during a runoff event. Twenty-four means of three hourly samples were plotted against instantaneous discharge to develop a linear model for predicting bedload from discharge. Upland Runoff To estimate overland sediment transport within the watershed, three sets of paired runoff plots were constructed using methods similar to those of Singer and others (1980). The three pair were on 10, 20 and 30 percent slopes respectively. The runoff plots and surrounding rangeland were moderately grazed leaving a minimum of 1000 kg/ha of residual dry matter (Clawson and others 1982). Each runoff plot is 2 USDA Forest Service Gen. Tech. Rep. PSW-GTR-184. 2002. 66 Sediment in a Grazed watershed—George, McDougald, Tate, and Larsen m wide by 21 m long (parallel to the slope). These runoff plots were established by cutting a 10 cm slit in the ground around the plot and installing 15 cm wide metal flashing. Surface water and sediment leaving each plot was collected and measured after each storm starting with the 1994-95 water year. Sediment Trap Study During September 1996, 1997, and 1998 sediment traps (Wells and Wohlgemuth. 1987) were placed in pairs near the first order intermittent stream that drains the research watershed at the San Joaquin Experimental Range. One of each pair was placed in a cattle trail near the point where the trail crosses the stream channel. The second sediment trap of each pair was placed in well-vegetated areas of similar slope and slope length adjacent to the trap in the trail. The sediment traps were emptied as needed or following large storms. Sediment samples were dried and weighed. ANOVA was used to separate treatment and year differences. Stream Channel Grazing Impacts Beginning in summer 1994 five grazing treatments were applied to five randomly selected 0.4 ha pastures established for a long-term study along each of three intermittent streams. During this study, streamflow began in early January following 270 to 360 mm of rainfall during October through December. The channels are 0.6 to 3 m wide, 0.3 to 1 m deep, and bedrock-controlled in many reaches. The study reaches are low gradient with less than 2 percent slope and are Rosgen Class B5 (Rosgen 1996). Stream channels 1, 2 and 3 are 2 to 3 kms apart and at an elevation of 274 to 411 m. The treatments were: