HOME > SCIENCE > ELLSWORTH CREEK > EXPERIMENTAL DESIGN

Experimental Design

Despite the desire of many forest managers to conduct restoration, scarce scientific information exists regarding appropriate restoration methods in the Sitka spruce forest zone or at the spatial scale of an entire drainage basin. Occasionally, attempts at restoration are made but these often lack comprehensive effectiveness monitoring, leaving the efficacy of the restoration actions unknown. For this reason, the Conservancy has developed an experimental design, with the help of nine independent scientists, to test three different restoration pathways at Ellsworth Creek in an attempt to maximize opportunities for learning how coastal forest landscapes respond to restoration treatments.
Proposed restoration pathways include:

1. Control: forest stands will remain unthinned during the initial treatment period (first 10 years) and all roads will be maintained or repaired as needed. This management pathway will be re-evaluated in 10 years in an adaptive management context.
2. Vegetation manipulation: forest stands will be actively thinned (e.g., variable density thinning) to relatively low stem densities during the initial treatment period (first 10 years) and at recurring intervals to promote forest growth and the development of structural complexity. Roads will be maintained to allow for restoration harvests and other management operations.
3. Road removal: forest stands will be left to develop without management intervention and all roads will be permanently abandoned (unless otherwise constrained by management or legal restrictions). 

The restoration study is based on an unbalanced randomized block design. Eight experimental units, ranging from 75 to 221 hectares have been defined within the Ellsworth Creek Adaptive Management Study Area. Each experimental unit includes an entire tributary sub-basin within the Ellsworth Creek watershed. Existing ecological conditions within all eight sub-basins were compared using GIS and those with similar conditions were placed within one of three experimental blocks. Because of existing site constraints, two blocks contain three sub-basins each while one block is limited to two sub-basins. A single restoration pathway has been randomly assigned to each sub-basin within a block.

Prior to treatment, baseline data collection has been collected in 2006 and will continue through 2007. Treatments are planned to be implemented beginning in 2008. Baseline data paints a picture of starting conditions in each sub-basin, thus providing a more clear understanding of changes that may occur following treatment. Included in the baseline data collection are: forest structure and composition, forest birds, headwater stream amphibians, stream macroinvertebrates, spawning populations of coho salmon, marbled murrelets, fish species occurrence and distribution, stream physical habitat, hydrology, and the amount of roads hydrologically connected to streams. Light Detection and Ranging (LiDAR) data will be collected from the entire watershed to aid in the understanding of forest structure, geomorphological features, and natural disturbance regimes.


Each of these is discussed in more detail below:

1. Forest structure and composition
   Forest vegetation sampling began in the summer of 2006 and will continue in the summer and fall of 2007. Each sub-basin contains 28 permanent monitoring plots that are randomly located. These plots are 18m radius and are used to sample tree species composition, tree diameters, tree heights, mistletoe infestation, and large woody debris on the forest floor. Within each plot are four 2m radius subplots. In each subplot, all herbaceous and shrubby species as well as seedling and sapling species are identified and their abundance is estimated.
    
2. Forest birds
   Forest birds are sampled using a variable radius plot count method. Within each sub-basin, 15 plots (the same as those used for forest vegetation surveys) are sampled for forest birds. These plots are visited four times between mid-May and late June. Sampling was done in 2006 and will be completed again in 2007. All birds, regardless of distance detected from the observer, are counted and recorded at each plot. This method allows for estimating bird density at each plot.
    
3. Headwater stream amphibians
   A daylight spotlight survey is used to sample headwater stream amphibians. In each sub-basin, eight 150m reaches are randomly selected and surveyed. Surveys were done in 2006 and will be completed again in 2007. Surveyors walk each reach and use a spotlight to search the stream channel and nearby banks for stream and stream-associated amphibians (e.g., Rhyacotriton kezeri). The spotlight survey is a no-touch method, meaning animals are not handled and substrate is not moved. This minimizes impact on the animals and their habitat. Species are identified by sight and their length is visually estimated. Substrate type, water temperature, and other relevant environmental variables are also recorded.
    
4. Stream macroinvertebrates
   Macroinvertebrates are collected from six stream reaches randomly located throughout each of the experimental sub-basins. Collection was done in 2006 and will continue in 2007. Samples are collected using a standardized method developed by Oregon Department of Environmental Quality and are then sent to an outside lab for identification. Analyses are then conducted to determine community composition, providing information on water and fish habitat quality.
    
5. Spawning populations of coho salmon
   During the fall/winter of 2005 – 2006 and 2006 – 2007, the main-stem of Ellsworth Creek and a subset of tributary streams were surveyed for spawning coho salmon. The surveys were done using the methodology of Washington Department of Fish and Wildlife. Trends in spawning coho populations in Ellsworth will be compared to those in other watersheds to track long-term effects of restoration work on coho salmon.
    
6. Marbled murrelets
   Marbled murrelet abundance is expected to be sampled within the Ellsworth Creek watershed basin using multiple radar stations following the methods of Cooper et al. (2000) (Table 10). The radar stations will be operated during the peak of the nesting season (June – July). Radar uses the murrelets’ speed (>62 km hr1) to detect individual murrelets. Detections are possible within approximately a 1 km circle of the radar. Several sites (e.g., 5) will be selected throughout the watershed and each site will be surveyed for 2 – 3 days during the breeding season. Initially, radar-based surveys of murrelets will only be done at the watershed scale although future efforts may be expanded to assess habitat use of individual forest stands. Because the murrelet is an endangered species strongly associated with late-seral habitat, abundance can provide a useful measure of the availability of suitable late-seral nesting habitat. Due to the high expense of these surveys, they will only be conducted once (summer 2007) prior to treatment.
    
7. Fish species occurrence and distribution
   A snorkel and electrofishing survey of all fish-bearing water will be conducted in Summer 2007 to determine distribution and abundance of salmonid and non-salmonid species in the watershed. Electrofishing will be employed in order to get an abundance estimate of salmonid species as well as identify species such as cutthroat trout, sculpins, and lampreys that are difficult to identify using snorkel techniques. This will provide the first comprehensive information regarding fish species in the watershed and will aid in developing future restoration methods and monitoring.
    
8. Stream physical habitat
   Stream physical habitat surveys will be conducted in 7 randomly selected reaches within each sub-basin. These surveys are conducted using the EMAP protocol developed by the U.S. E.P.A. Surveys will provide information on channel habitat classification, bed stability, substrate size and composition, large woody debris, fish habitat, and other physical stream habitat characteristics.
    
9. Hydrology
   One depth gauge and temperature probe will be located near the furthest downstream point of a perennial stream in each sub-basin. This will provide continuous long-term hydrology data for each sub-basin. Gauges were deployed in February – March, 2006, with measurements being recorded every 15 minutes. Data is collected from gauges on a bi-monthly basis. Hydrographs produced from flow and discharge data will be used to assess changes in peak flows and other hydrologic events in each sub-basin.
    
10. Amount of roads hydrologically connected to streams
    The amount of roads hydrologically connected to 0 through 3rd order streams within each sub-basin will also be measured. This data will permit assessment of the effectiveness of road removal in decreasing the amount of roads with direct connectivity to the stream network. In addition, it will provide information regarding road-stream connectivity that will supplement stream habitat survey and hydrology data.
     
11. LiDAR
    Three-dimensional canopy structure will be assessed using Light Detection and Ranging (LiDAR) technology. From LiDAR data we will measure the following components for the entire Ellsworth Creek Adaptive Management Area including each sub-basin: percent foliar cover, canopy roughness, and size and spatial distribution of canopy gaps (a measure of horizontal structure). Vertical canopy profiles of sub-basins will also be generated. These data will permit comparisons of three-dimensional canopy structure across treatments. Accurate location of permanent vegetation plots may also allow investigation of correlations between field-based under story data (plot measurements) and LiDAR measured canopy data.