Study Area

The study area is in northwestern California in the upper Clear Creek watershed, primarily within Whiskeytown National Recreation Area (fig. 2). Whiskeytown National Recreation Area is at the convergence of three physiographic provinces: (1) the Cascade Range, (2) the Coast Range, and (3) the Sacramento Valley, which is considered the northern part of the Central Valley (California Geological Survey, 2002). The study area has a Mediterranean climate with hot dry summers and cool winters with moderate rainfall. Mean annual rainfall in the area is approximately 1,500 millimeters (mm). The study area lies within the Klamath-Siskiyou ecoregion (not shown in figures), an area of substantial biodiversity, especially for terrestrial plants (DellaSala and others, 1999). In the context of fish zoogeography, the study area lies within the Central Valley subprovince of the larger Sacramento–San Joaquin province (not shown in figures; Moyle, 2002). The Central Valley fish fauna includes 28 native species, many of them endemic to California, and 40 non-native species (Moyle, 2002). The region’s moderately rich herpetofauna of 38 species is attributed to the overlap between northern and southern taxa and the presence of appropriate, complex habitat (Bury and Pearl, 1999).

Purpose and Scope

The objective of this report is to describe assessments of water quality, sediment quality, benthic macroinvertebrate assemblages, fish and amphibian communities, and stream-habitat characterization of tributaries to Whiskeytown Lake after the Carr Fire. The post-Carr Fire data will be compared to available pre-Carr Fire data. Pre-Carr Fire samples were collected in spring and fall 2004–06, and post-Carr Fire samples were collected in August 2020. Constituents analyzed include metals of known biological concern such as arsenic, cadmium, chromium, copper, nickel, selenium, and zinc that were collected from water, sediment, and invertebrate and fish tissues. This information can help determine if there has been wildfire-induced changes in the aquatic communities within Whiskeytown National Recreation Area.

Water Sample Collection and Analysis

Water-quality samples were collected and processed following standard USGS protocols described in the National Field Manual (U.S. Geological Survey, variously dated). Before collecting samples from the field, all field equipment was cleaned according to USGS protocols and then rinsed with native water immediately before samples were collected. One water sample was collected at each site using a mid-stream dip method and filtered and preserved as required for the analysis of the constituent as described in the constituent referenced methodology (Shelton, 1994; McCormick and others, 2006). Sampling included one field blank and one replicate sample at two sites for quality assurance purposes. The number of blank and replicate samples taken was 20 percent of water-quality samples. Field blanks were collected according to USGS protocols as described in U.S. Geological Survey (variously dated). Samples were analyzed for nutrients, major ions, and trace metals at the USGS National Water-Quality Laboratory in Denver, Colorado. Nutrients were determined from 0.45-micrometer (μm) filtered and unfiltered samples using methods of Fishman (1993) and Patton and Kryskalla (2003). Water collected for major ions was acidified in the field and determined from filtered water samples in the laboratory using the methods of Fishman and Friedman (1989), Hoffman and others (1996), and Garbarino and Struzeski (1998). Concentrations of dissolved arsenic, boron, lithium, selenium, strontium, thallium, and vanadium were determined using methods in Garbarino (1999). Concentrations of all other trace metals were determined from acidified, filtered samples using the methods of Fishman and Friedman (1989) and Garbarino and others (2006).

Bed Sediment Sample Collection and Chemical Analyses

Bed sediment samples were collected using protocols described by Axtmann and others (1997). Generally, one composite sample was collected per site. At two sites, composite samples were collected in triplicate to compute variability of chemical concentrations for quality assurance. Sediment samples were collected by scooping material with an acid-washed polypropylene scoop from the submerged surfaces of sediment deposits reported along pools or low-velocity areas near the edge of the stream. Whenever possible, samples were collected from both sides of the stream. Individual samples of bed sediment were collected by scooping material from the surfaces of 5–10 deposits along pools or low-velocity areas within the study reach. The 5–10 individual samples were combined to form a single composite sample. Each composite sample was wet sieved onsite through a 63-μm mesh nylon-sieve cloth directly into a sample jar using ambient stream water. The fraction of bed sediment in each composite sample that was finer than 0.063 mm was collected in an acid-washed 500-milliliter (mL) polyethylene bottle.

Samples were placed on dry ice and frozen before shipping to the USGS Analytical Trace Element Chemistry Laboratory in Boulder, Colorado, for chemical analyses following analytical methods including the quality assurance and quality control (QA/QC) approach described by Roth and others (2022). Inorganic sediments were digested in a high temperature and pressure microwave environment using an Ethos EZ Milestone digestion system. Samples sets of 10 samples include 1 blank and at least 1 National Institute of Standards and Technology (NIST) standard reference material (SRM) sample to verify digestion concentrations. Organic samples are digested similarly except only nitric acid (HNO₃) was used, and the acid was added several hours ahead of time to rid the sample of excess carbon dioxide (CO₂). Sample collection and processing equipment were cleaned in the field between each sampling site using methods described by Shelton and Capel (1994).

Macroinvertebrate Sampling and Analysis

Benthic macroinvertebrates (greater than 0.5 mm) were collected using a reach-wide quantitative multihabitat sample, following methods outlined in Wulff and others (2012), similar to the USGS California Stream Quality Assessment methodology (May and others, 2020), with the exception of sieving samples only once through a 0.5-mm sieve. A D-frame net was used to collect 11 kick-net samples with 500-μm mesh openings. Samples were collected 1 m downstream from each of the 11 cross-section habitat transects, alternating in a left, center, right pattern. All kick-net samples collected at transects were combined into a single composite sample and sieved through a 0.5-mm sieve before being preserved with a 95-percent ethanol solution. Macroinvertebrates were analyzed by Benthic Aquatic Research Services in Lawrence, Kansas. Samples were processed according to a standardized protocol consistent with Woodard and others (2012). A minimum of 600 organisms were counted per sample, except when the sample contained fewer than 600 organisms. Taxonomy conformed to the Southwest Association of Freshwater Invertebrate Taxonomists Level 2 taxonomic effort (Richards and Rodgers, 2011, appendix I). Organisms were classified to the lowest practical taxon, usually genus for well-known groups or some higher taxonomic level for some of the lesser known or difficult-to-identify taxa. Macroinvertebrate data were summarized as percentage abundance at the appropriate taxa level. The percentage abundance format standardizes counts to a common scale (percentage), provides an easily understood measure of relative abundance for comparison within and among stations, and is a format commonly used for statistical analyses (Moulton and others, 2000; Rehn and others, 2005).

Invertebrates (Hemiptera, family Gerridae) were collected for tissue-metals analysis following procedures described in Hothem and others (2008). Samples were sorted and cleaned before being composited by site, with the goal of obtaining a minimum of 1-gram (g) wet biomass. Each sample consisted of 15–27 individuals. Samples were weighed on an electronic balance, placed into chemically cleaned glass jars, and stored frozen until they could be shipped to the laboratory for analysis following methods described by Roth and others (2022).

Fish and Amphibian Sampling

Fish communities were assessed using standard electrofishing techniques using a Smith-Root Type LR-24 electrofisher. Amphibians were not targeted in sampling but were measured and released if inadvertently captured while collecting fish. Two passes were done to document species richness and relative abundance of fishes and to collect riffle sculpin (Cottus gulosus) for tissue analysis. Total length and standard length in millimeters were measured for fish. Snout-vent length (in millimeters) was measured on Pacific Giant Salamander (Dicamptodon tenebrosus) larvae. A minimum of the first 30 individuals of each species were measured to the nearest 0.1 g on an electronic balance.

Riffle Sculpin was the target species for tissue metals analysis because it is widely distributed in the sampled streams. Tissue samples consist of whole-body composite samples with a target of at least five similarly sized individuals per sample. Samples were placed in individual glass sample jars and stored frozen for no more than 30 days before they were sent to the USGS Analytical Trace Element Chemistry Laboratory for analysis following methods described by Roth and others (2022).

Stream-Habitat Characterization

Instream habitat measurements were made at each site according to the NAWQA stream habitat protocol (Fitzpatrick and others, 1998; Moulton and others, 2002) and included reach length, elevation, discharge, percentage riffle/run/pool, mean wetted-channel width, mean open canopy, mean riparian canopy closure, mean depth, mean water velocity, and dominant substrate. Stream sampling reach lengths ranged from 100 to 400 m and were determined as 20 times the wetted-channel width. Habitat variables were measured at 11 transects per sample site, with transects placed at equally spaced intervals within a reach. Stream width (wetted channel) was measured directly from a transect tape, and open canopy angle was measured from midstream with a clinometer. A densiometer was used to calculate riparian canopy closure percentage. Percentage area of habitat features such as woody debris and overhanging vegetation were visually estimated from mid-channel within a section of the stream extending 2 m upstream and downstream from the transect tape. Depth, velocity, and substrate were measured at a minimum of three points on each transect. Depth was measured with a wading rod and velocity with an electronic meter (Marsh-McBirney) at 60 percent of total depth measured from the bottom. The dominant substrate was classified at each transect point. Habitat variables were analyzed as the geometric mean of the 11 transect values or the geometric mean of the 33 (or more) point values.

In addition, a more detailed characterization of sediment size classes was done at each site by quantitatively measuring channel substrate particle size using pebble counts (Wolman, 1954), as outlined in the NAWQA stream habitat protocol (Fitzpatrick and others, 1998). A minimum of 20 substrate-size measurements were collected at each transect, resulting in approximately 200 individual substrate-size measurements per survey site. The substrate-size data were collected to allow assessment of changes in fine sediment associated with future runoff events.

Water Chemistry

The highest concentration of arsenic, 3.1 micrograms per liter (µg/L), was measured at CCTMR. The Willow Creek near French Gulch sample site (WLCR2) had the highest concentrations of copper (13.5 µg/L) and zinc (27.1 µg/L). The Clear Creek at Peltier Valley Road near Shasta sample site (CLCR2) had the highest concentration of nickel (2.1 µg/L), and the Whiskey Creek near Whiskeytown sample site (WKCR1) had the highest concentration of selenium (0.5 µg/L). The range of concentrations across all sites for metals is reported as follows: arsenic (less than [<] 0.1–3.1 μg/L); copper (<0.4–13.5 μg/L); nickel (<0.2–2.1 μg/L); selenium (<0.05–0.5 μg/L); and zinc (<2.0–27.1 μg/L; table 2).

Concentrations of metals of biological concern in water were not consistently higher or lower in 2020 than in previous years (fig. 3). Pre- and post-Carr Fire concentrations of arsenic, nickel, selenium, and zinc did not exceed U.S. Environmental Protection Agency (EPA) criteria for drinking water or aquatic life (U.S. Environmental Protection Agency, 2022); however, copper concentrations at WLCR2 exceeded EPA criteria (U.S. Environmental Protection Agency, 2022). Sample site WLCR2 had the highest concentrations of copper for all sample years, and the 2020 sample was the highest concentration of copper of all available data (13.5 micrograms per gram [µg/g] copper; water hardness equals 102 milligrams per liter [mg/L]; Wulff and others, 2023); this concentration of copper exceeds the chronic species toxicity concentration (13 µg/g) for mottled sculpin (Cottus bairdii) determined by Besser and others (2007) and calculated for a water hardness of 100 mg/L (shown as a dashed line on fig. 3). Willow Creek is known to be affected by high metal concentrations attributed to its proximity to Greenhorn Mine, an abandoned copper mine operated in the early 1900s (Tetra Tech, Inc., 1998; Hothem and others, 2015). Post-Carr Fire concentrations of copper measured at sample sites Grizzly Gulch near Whiskeytown (GZGL) and Mill Creek near French Gulch (MLCR1) were also the highest of the available data.

With the exception of arsenic, concentrations of metals of concern were lowest in BRAN1 in 2020, post-Carr Fire. Concentrations of arsenic measured in BRAN1 were higher in 2020 than previously measured; however, concentrations of arsenic at BRAN1 were lower than concentrations measured at the other nine sites, and all samples were below EPA criteria (U.S. Environmental Protection Agency, 2022). Most data for metals of biological concern that were available for comparison with pre-Carr Fire data indicate variation among sites and years with no obvious patterns (fig. 3). Additional statistical analyses of available data could be completed to further explore the existence of potential patterns in the data that are not visually apparent.

Bed Sediment Chemistry

Of all sites sampled for bed sediment chemistry, WLCR2 had the highest average concentration of arsenic (28 μg/g), cadmium, (6.8 μg/g), copper (584 μg/g), and zinc (1,660 μg/g). Sample site CLCR2 had the highest concentrations of chromium (122 μg/g) and nickel (124 μg/g). The Crystal Creek near mouth near French Gulch sample site (CYCR1) had the highest concentration of lead (30.6 μg/g). The ranges of concentrations across all sites for metals are (1) arsenic (2.7–28 μg/g); (2) cadmium (0.3–6.8 μg/g); (3) chromium (58–122 μg/g); (4) copper (27–584 μg/g); (5) nickel (21–124 μg/g); (6) lead (16–30.6 μg/g); and (7) zinc (96–1,660 μg/g; table 3; Wulff and others, 2023).

Benthic sediment results for metals of biological concern were compared with the consensus-based probable effect concentrations (PEC) from sediment-quality guidelines in MacDonald and others (2000; fig. 4). Measured concentrations from 2020 sampling indicated exceedances of these guidelines at one site for chromium (CLCR2), copper (WLCR2), and zinc (CLCR2), and three sites for nickel (CLCR2, WKCR1, WLCR2). Data collected in 2004–05 indicated exceedances for arsenic, cadmium, copper, and zinc (fig. 4). Arsenic, cadmium, copper, lead, and zinc concentrations varied among sites and years, with no obvious pre- versus post-fire patterns; however, concentrations of chromium measured in 2020 were higher than previous years at all sites, and concentrations of nickel measured in 2020 was higher in seven out of eight sites. The higher concentrations of chromium and nickel in bed sediment samples could be the result of post-fire sediment mobilization, and this finding could be especially relevant for watersheds with historical mining activity (Abraham and others, 2017).

Macroinvertebrate Samples

Benthic macroinvertebrate metrics that are part of the Northern California Index of Biotic Integrity (Rehn and others, 2005) are in table 4. Percentage of Ephemeroptera (mayflies), Plecoptera (stoneflies), and Trichoptera (caddisflies), known as EPT taxa, was highest (69 percent) at WLCR2. Percentage of Coleoptera (beetles) taxa was highest (19 percent) at CCTMR. Sample site CCAR had the highest percentage of Diptera taxa (61 percent), and CLCR2 had the highest percentage of non-insect taxa in samples (3 percent). The ranges of each macroinvertebrate metric were (1) EPT taxa (27–69 percent), (2) Coleoptera taxa (1–19 percent), (3) Diptera taxa (6–61 percent), and (4) non-insect taxa (0–3 percent).

Gerridae tissue samples had the highest average concentration of cadmium (12.47 µg/g), lead (0.13 µg/g), and zinc (333.35 µg/g) in WKCR1 (table 5). All measured chromium concentrations were below reporting limits (<1.2 µg/g). Sample site CLCR2 had the highest average concentrations of arsenic (0.46 µg/g) and nickel (0.63 µg/g). The ranges of concentrations averaged by site for metals are (1) arsenic (0.12–0.46 μg/g); (2) cadmium (1.57–12.47 μg/g); (3) chromium (<1.2 μg/g); (4) copper (30.97–57.80 μg/g); (5) nickel (0.31–0.63 μg/g); (6) lead (0.05–0.13 μg/g); and (7) zinc (205.0–333.35 μg/g; table 5).

Fish and Amphibian Samples

Average concentrations of arsenic (0.49 µg/g) and lead (0.21 µg/g) were highest at GZGL. Sample site WLCR2 had the highest average concentrations of cadmium (0.41 µg/g), copper (8.56 µg/g), and zinc (164.44 µg/g). Sample site WKCR1 had the highest average concentration of nickel (1.43 µg/g). Average chromium concentration was highest (1.7 µg/g) at CCAR, but 95 percent of chromium samples were below reporting limits (<1.1 µg/g). The range of concentrations averaged by site for metals are (1) arsenic (0.13–0.49 μg/g); (2) cadmium (0.02–0.41 μg/g); (3) chromium (<1.1–1.7 μg/g); (4) copper (2.11–8.56 μg/g); (5) nickel (<0.03–1.43 μg/g); (6) lead (0.06–0.21 μg/g); and (7) zinc (67.44–164.44 μg/g; table 6).

A total of seven fish species and one amphibian species were collected during the study (table 7). Riffle sculpin was the most abundant species and was collected at most sites. Of the eight fish taxa collected, only two were non-native species: Bluegill (Lepomis macrochirus) and Spotted Bass (Micropterus punctulatus). Only Spotted Bass was documented in more than one sample. Pacific Giant Salamander were only captured at BRAN1.

Post-Carr Fire concentrations of metals in fish and invertebrate samples were compared with pre-Carr Fire concentrations at two sites, BRAN1 and CCAR. These two sites were the only sites where data were collected in pre- and post-Carr Fire years. Concentrations of metals of biological concern, including arsenic, chromium, cadmium, copper, lead, nickel, and zinc, generally varied among sites and years without a specific pattern in relation to the Carr Fire (figs. 5, 6). Average zinc concentrations in composite invertebrate samples were slightly higher in 2020 than in 2004–05 at BRAN1 and CCAR (fig. 5). Average lead concentrations in fish tissue were higher at both sites in 2020 than in 2004–05. Arsenic concentrations were lowest in 2020 for invertebrate and fish tissues. Chromium and cadmium concentrations in fish tissue were lower in 2020 than in previous years (fig. 6).

In general, review of 2020 post-Carr Fire data compared to previous years demonstrated that concentrations of metals of biological concern varied substantially among sites and years, with few patterns in the data. Concentrations of chromium and nickel in bed sediment increased after the Carr Fire, but most other metals in sediment and water either did not change or demonstrated high variation among sites and years. Concentrations of lead in riffle sculpin tissues were higher post-Carr Fire compared to pre-Carr Fire, and arsenic concentrations were lower in invertebrate and fish tissues in post-Carr Fire samples compared to pre-Carr Fire samples.

Stream-Habitat Characterization

Habitat characteristics, including reach length, elevation, discharge, percentage riffle/run/pool, mean wetted width, mean open canopy, mean riparian canopy closure, mean depth, mean water velocity, and dominant substrate, were assessed for each site (table 8). Sample sites CLCR2 (96 percent) and WLCR2 (89 percent) had the highest percentage of riparian canopy closure, or riparian shade, of all the sites. Clear Creek had the highest average width, highest mean discharge, and lowest gradient sampled. Sample site CLCR2, downstream from Whiskeytown Lake, is affected by operations of Whiskeytown Dam, and has the highest recorded discharge of all the sites. Substrate size and heterogeneity are summarized by site (figs. 7, 8).

Stream-habitat and sediment-size characterization values either did not change substantially or demonstrated high variation among sites and years (figs. 9, 10). Fish and amphibian count data indicated that fewer total fish and amphibians were collected in 2020 than in 2005, but higher numbers of native Sacramento Sucker (Catostomus occidentalis) and Sacramento Pikeminnow (Ptychocheilus grandis) were collected in 2020 than in previous years (fig. 11). Sample site BRAN1 has been subjected to the least amount of historical mining disturbance and frequently had the lowest concentrations of metals of biological concern in water, sediment, and fish and invertebrate tissue samples.