Project Title: Stream Recovery Following Watershed Restoration

Investigator: Mary Ann Madej USGS-BRD Redwood Field Station Arcata, CA USA 707-464-6101 Ext. 5490 FAX: 707-822-8904 mary_ann_madej@usgs.gov Partners: Redwood National and State Parks Arcata, CA USA 707-464-6101
	Bridge Creek (left), a major tributary of Redwood Creek and an important anadromous fisheries stream, was extensively disturbed by timber harvest and road construction during the 1970's. Recovery of such disturbed systems to approximate pristine streams (right) is a focus of current research.

Project Description:

Many anadromous fisheries streams in California and the Pacific Northwest have been damaged by various land use activities, including timber harvest, road construction, and grazing. A watershed restoration program at Redwood National Park has been correcting erosional problems and revegetating previously logged hillslopes for over 25 years. The Redwood Field Station is evaluating the effects of such rehabilitation work on stream channels, based on comparisons with previous surveys conducted 30 years ago. Physical stream recovery is being evaluated in several ways: by a return to former pool depths and pool frequency, water temperature and canopy closure, channel bed elevation, channel width and area, and sediment loads. These physical characteristics influence habitat and, consequently, the aquatic and riparian communities associated with these streams. Biological stream recovery is being evaluated through analyses of benthic macroinvertebrate abundance and diversity, periphyton growth, presence and abundance of amphibians and presence and condition of salmonids. Scientists are assessing the degree of stream recovery and the rates of recovery following watershed restoration efforts. The study will also model various restoration scenarios to determine the most effective strategy for stream improvement.

Purpose:

The purpose of the research is to assess the effectiveness of current watershed restoration practices on improving conditions in aquatic and riparian habitats, and to recommend procedures for future restoration efforts. Guidelines for improved restoration techniques will be distributed to park managers and private landowners. The results of this study should be applicable to other damaged stream systems in the western United States as well.

Progress/Results:

In order to assess the impacts of upslope restoration efforts on stream health, aquatic biological and physical data were gathered from 21 sites within the Redwood Creek drainage basin during the spring and summer of 2004. These surveys repeated data collection that was conducted in 1973-75. Streams with both extensive upslope road restoration and untreated road networks were sampled, as well as control streams. More specific mapping of road networks within each basin will be conducted in 2005. A 100-m study reach was designated at each site near previously established gaging stations. Biological data collected included the abundance and diversity of benthic macroinvertebrates and aquatic amphibians, biomass of periphyton and relative abundance of juvenile salmonids. Physical habitat parameters were also measured at each site.

1) Macroinvertebrates:
In order to compare current macroinvertebrate abundance and diversity with that from 1973-75 (Averett and Iwatsubo 1995), ten sites were sampled for benthic macroinvertebrates using a 250 µm Surber sampler in late May through June. Three riffles from each 100m study reach were randomly selected to sample. One sample was taken from each selected riffle at an area within the riffle where the Surber sampler could be placed flush with the stream bed, for a total of three samples per chosen site. The Surber sampler encompasses an area of 0.3 m2, which was disturbed for a timed period using a stopwatch. The sampling period averaged one minute. Invertebrates were carefully removed from the Surber sampler and washed into a 250 um sieve. Samples were preserved in 50% ethanol and stored in whirl-pak® bags for later identification. Nineteen sites were sampled in September following the above sampling procedure, except that the samples were preserved in 85% ethanol.

In addition, the same nineteen sites were sampled for benthic macroinvertebrates in September following the California Stream Bioassessment Procedure (Harrington and Born 2000), a regional adaptation of the United States Environmental Protection Agency’s Rapid Bioassessment Protocol. Three riffles from each 100-m study reach were randomly selected to sample. Starting with the downstream riffle within each stream reach, a measuring tape was placed along the entire length of the riffle. The number of meter marks in the upstream third of the riffle was divided by three and a transect was randomly selected from this third to sample. Along this transect the right side margin, thalweg and left side margin were each sampled by placing the kick net on the substrate perpendicular to flow and disturbing a 0.3m x 0.6m portion of the substrate upstream of the net to an approximate depth of 10 cm for a timed period which averaged between 3 and 4 minutes. If the stream was not wide enough to take three kicks along the transect, three kicks were taken longitudinally within the chosen riffle. The three collections within the kick net were combined to make one composite sample per riffle for a total of three samples per study reach.

Invertebrates collected with the kick net were carefully removed from the net and washed into a 500 µm sieve. Samples were preserved in 85% ethanol in whirlpak® bags for laboratory identification. The preserved samples from each site were cleaned and combined in the laboratory for a total of one sample per study reach. Subsamples from each of these samples were prepared by spreading the sample evenly across a tray marked with 5 x 5cm grids to an approximate thickness of 1 cm. Each grid was removed from the tray with a razor blade and placed in a petri dish containing 85% ethanol for laboratory identification. Invertebrates were identified to the genus level when possible. Sample identification is still in progress. Enough grids were randomly chosen to be subsampled in order to identify 500 macroinvertebrates per site. A voucher specimen will be kept for each species identified. Voucher specimens will be stored in a glass vial with a label containing 85% ethanol.

2) Amphibians:
Stream amphibian abundance and diversity were measured at twelve sites in late May through June. Amphibians were sampled using a 800 µm D-frame kick net. Two people conducted the survey at six randomly chosen riffles per study reach. At each riffle, one person held the net while the other crew member stood upstream of the net and used both their hands and feet to dislodge stream rocks and debris downstream into the net for a timed period averaging five 5 minutes. The person holding the net continued to move the net upstream as the other person moved upstream during the sampling period for an average distance of two meters per riffle. Amphibians collected from each riffle were identified and snout vent length, total length, mass and degree of larval development of each individual were recorded. Amphibians were released once processed.

3) Periphyton:
Periphyton ash-free dry mass (AFDM) was measured at six sites in the spring and at nine sites in late summer to replicate previous sampling by the U.S.G.S. Plexiglass plates cut into 10 x 10cm squares were placed in the chosen study reaches and allowed to colonize for 70 days. Two plates each were nailed into the stream bed at three randomly chosen riffles per reach for a total of six plates per site. The plates were installed parallel to the stream bed using 15cm nails. Spring sampling plates were placed in the sites in late May through June and late summer sampling plates were placed in the sites in mid-August through early September. Dates were chosen to correspond with previous measurements made by the USGS in 1973-75. Velocity and canopy measurements were taken at each plate. After the colonization period, plates were scraped with a plastic bristled brush to remove periphyton and rinsed with stream water filtered through a PUR® water filter into a collecting tray. Periphyton scraped from the two plates in each riffle was composited for a total of one sample per riffle. The sample was filtered through a mesh net into a graduated cylinder to remove debris, mixed well and filtered onto pre-combusted 25 µm glass fiber filters for AFDM determination, with volumes noted. Filters were kept frozen until time of analysis. Filters were oven-dried at 100° C for 24 hours, allowed to cool in a dessicator and weighed on an analytical balance to the nearest 0.0001g to obtain a dried mass. The filters were then ashed at 500° C for 1 hour, allowed to cool in a dessicator, rewet with a few drops of distilled water to restore the waters of hydration, oven-dried at 100° C for 24 hours and weighed to obtain an ashed weight.

4) Fish condition:
Length (mm) and weight (g) data for steelhead trout (Oncorhynchus mykiss) were used to compare recent fish condition to that documented thirty years ago for selected study reaches. Both sets of data were collected from surveys conducted in tributaries to Redwood Creek that included Bridge Creek, Little Lost Man Creek, Harry-Weir (Emerald) Creek, and Tom McDonald Creek. Fish collected during 1974 and 1975 surveys were collected via back-pack electrofisher and seine (Iwatsubo et al. 1976). Fish collected during 1994 and 1995 were collected solely by electrofishing techniques (unpublished data, Redwood National and State Parks). The relationship between the mean length and weight relationship of two fish populations can be used to compare fish health across both spatial and temporal scales. Fish condition indices were calculated using Fulton’s condition factor, K and ordinary least squares regression techniques (Cone, 1989). The results of these calculations are summarized in Tables 1 and 2.

Snorkel surveys were conducted at 21 sites in July in order to determine the presence/absence of juvenile salmonids. Pool and run habitats within each 100m study reach were snorkeled. Two snorkelers were present at each site, although most habitats were easily snorkeled with only one person. Surveys began at the downstream end of each reach and the observer moved upstream to observe the entire study reach. The total number and species of fish observed in each habitat were recorded using a hand counter. The width and depth of each unit were recorded and the visibility of each habitat unit observed was given a subjective rating of poor, fair, or good.

5) Physical habitat parameters:
Physical habitat parameters were measured at 21 sites in July. The average stream gradient for each study reach was based on longitudinal profiles taken with a hand level and stadia rod in steep channels (> 2 percent), or with a self-leveling level in more gentle gradient streams. Cross sectional transects were surveyed with a stadia rod and measuring tape at previously established sites, where they existed. New cross section survey monuments were established at each site as well. Discharge was measured with a pygmy flow meter at each site. Air and stream temperatures were measured with a hand-held thermometer. Substrate along the length of each reach was measured using a modified Wolman pebble count technique, by systematic sampling along a tape at one meter spacing (Bunte and Abt, 2001). All large woody debris over 10cm in diameter that was in the active channel was recorded at each reach. Percent canopy coverage was determined using a spherical densiometer. Photographs were taken at the upstream and downstream end of each study reach.

In the Bridge Creek watershed within Redwood National and State Parks, 30 kilometers of roads that have been rehabilitated were inventoried before and after a 12-year flood. This mapping documented the effectiveness of various rehabilitation techniques in reducing sediment loads to perennial stream channels. Several kilometers of perennial stream channel were surveyed to document pool depth and frequency, areas of bank erosion, stream canopy, etc. Air and stream temperatures were measured continuously throughout two summer low flow seasons at several locations within Redwood National and State Parks.

Citations:

Much of the biological and geomorphic work is still in progress.

Bloom, Anna. 1997. "An Assessment of erosion Control Treatment Effectiveness: A Comparison of 1997 Storm Erosion Response between Treated and Untreated Roads." Masters thesis in Geology, Humboldt State University.

Madej, M.A. 1996. Measures of stream recovery after watershed restoration. p. 46-56 in Proceedings on Watershed Restoration Management: Physical, Chemical and Biological Considerations. J.J. McDonnell, J. B. Stribling, L. R. Neville, and D. J. Leopold, eds. American Water Resources Association. Syracuse, New York. 524 pp.

Madej, M. A. 2001. Erosion and sediment delivery following removal of forest roads. Earth Surface Processes and Landforms. Vol. 26. No. 2 p. 175-190.

Madej, M.A. and G. Gibbs. 1998. Bridge Creek Stream Channel Monitoring Progress Report. In-house report. U.S. Geological Survey Western Ecological Research Center. 100 p.

Keywords: watershed restoration, land use, rehabilitation, stream restoration, aquatic habitat, watersheds