HGM and Bioassessment Development in Oregon

Last Updated: March 2000

Contact Information

Paul Adamus
Adamus Resource Assessment, Inc.
6028 NW Burgundy Drive
Corvallis, OR 97330
(541) 745-7092
Email: adamus7@attbi.com

Dana Field
Oregon Division of State Lands
775 Summer St. NE
Salem, OR 97310
(541) 378-3805 ext. 238
Email: Dana.Field@dsl.state.or.us

Simultaneous Development, Calibration, and Testing of Hydrogeomorphic-Based (HGM) Assessment Procedures and Biological Assessment Procedures in Oregon

Purpose(s) of Project

To collect and analyze field data on wetland and riparian plant communities and hydrogeomorphic (HGM) features to define reference standard condition and performance criteria for two subclasses of wetlands in for the Willamette Valley in western Oregon. To use that information to develop quantitative but rapid HGM-based assessment models and procedures for those subclasses, as well as to identify new wetland plant community metrics that have demonstrated relationships with landscape condition and with internal alterations of wetlands in this region. To compare results of using HGM models and plant metrics, to verify the results on an independent data set, and demonstrate implementation of the methods for assessing progress of restored wetlands and for helping recommend compensation ratios in a wetland mitigation bank context.

Project History

This project is being done cooperatively with the Oregon Division of State Lands, which in 1997 identified a need, in the context of Section 404 and 401 responsibilities, for a more quantitative assessment method tailored specifically to two subclasses of wetland and riparian systems in the Willamette Valley. The project began in 1998 with formation of an interagency oversight committee, intensive review of local literature, and reconnaissance and preliminary selection of 124 reference sites throughout the ecoregion.

In 1999, the list of sites was narrowed; workshops of local experts were held to develop function models and field procedures; over 38 volunteers were recruited and trained to assist with field work; and two assessment teams ("A-teams") of scientists with help from the volunteers collected data from the selected sites.

The first round of data analysis has been completed. Additional sites will be assessed during summer 2000 to test wider applicability of the results from the first field season. No funding commitments have been made beyond the current year (2000).

Study Design

Plant and hydrogeomorphic data were collected from 69 wetland and riparian sites in the Willamette Valley ecoregion (33 riverine impounding subclasses, 29 slope/flat subclasses, 7 depressional subclasses). Data from the depressional sites were not analyzed because time and resources did not permit a wider sampling of this subclass. The "riverine impounding" sites include all wetland and riparian areas within a two-year river flood plain, e.g., sloughs, oxbows, cut-off channels, beaver impoundments, stream-fed ponds with water control structures. "Slope/Flat" sites include most ash swales, wet prairies, springs, and foothill seeps. From visual inspection alone, most sites in the "slope" HGM class could not be reliably separated from most sites in the "flat" HGM class, due to the flat topography and varied geology of the region, so the two were combined into a "slope/flat" subclass.

Within the subclasses, assessment sites were selected in a manner intended to span gradients of human disturbance, size, and plant community succession. Restored/created sites of various ages were also included. Nearly all sites were on public lands, and ranged in size from 0.1 to 233 acres. Most sites were nominated by local wetland experts. From the collected data, five least-disturbed riverine sites were selected as reference standards, as were six least-disturbed slope/flat sites. Use of cluster analysis and other statistical analysis methods verified that sites belonging to the two main subclasses were significantly different based on landscape position and mapped soil characteristics important to ecosystem function.

For defining the disturbance gradient (the "x axis"), information specific to each candidate site was collected during the reconnaissance phase. This information pertained to past management practices, surrounding land use, and recent physical alterations. Physical alterations were noted visually during reconnaissance visits. At each site, each type of alteration was categorized as: (a) absent, (b) physically affects less than 10 percent of site, (c) affects greater than 10 percent of site but not all, or (d) affects entire site. The categories that were used for alterations were:

Flow-impounding – berms, dikes, dams
Flow-impounding – excavations, pits
Flow-redirecting
Water subsidy (e.g., storm water pipes)
Drainage-inducing (ditches, tile)
Soil compacting (e.g., fill, machinery, cows)
Soil mixing (e.g., plowing)
Soil grading (e.g., flattening)
Vegetation removal (e.g., extreme grazing, logging)

Assemblages Monitored: Plants

Sampling Methods: Plants

At each site, the A-teams identified plants in each of potentially three hydrologic zones: permanent water zone, seasonally inundated zone, saturated-only zone. For woody plants, the team walked the entire site and made an overall estimate of relative percent of the area of each zone occupied by each shrub species (understory and open) and tree species. For herbaceous plants, the team assessed relative cover of each species in 1m x 1m quadrats.

The number of quadrats used was proportional to the area of each zone, with no more than nine quadrats used at any site (and fewer at smaller sites and sites with fewer hydrologic zones). Larger numbers of quadrats were not used because of time and resource constraints. Within zones, quadrats were located so as to maximize the cumulative number of species found, inasmuch as our goal was to develop metrics based mainly on community composition rather than quantitative measures of abundance or cover. In 1999, two teams assessed all 69 sites in 30 work days. Although the field work occurred in late September and October, which was less than optimal for detecting some plant species, 383 taxa were found.

Analytical Methods: Plants

From the collected data, the following plant metrics were compiled for each site:

• Number of native herb species, relative to the intensity of sampling (number of plots at each site) (based on analysis using species-area curves and regression).
• Percent of herb species that are native.
• Percent of the herb species are "remnant-dependent" (i.e., reputedly most characteristic of unaltered sites).
• Percent of the dominant herb species in any plot that are natives.
• Percent of the dominant herb species in any plot that are remnant-dependent.
• Percent of the true wetland species (facultative or wetter) that are natives.
• Percent of shrub species that are natives (when shrubs present).
• Percent of native shrub species that are at least moderately dominant.
• Percent cover of nonnative shrubs.

Data analysis is ongoing and other metrics may yet be identified. Preliminary analysis suggests many of the above metrics had a statistically significant association with categorical observations of partial physical degradation of sites (see Study Design, above) and/or with amount and proximity of surrounding land cover categories (agriculture, urban, natural) that were assessed visually during field work as well as by a GIS analysis of existing digital imagery. Analyses of the HGM and plant data sets, each containing over 4,000 records, were performed on a PC using Excel, PC-ORD, and NCSS.

Lessons Learned

• Random or systematic sampling, whether within a region or within a site, is not always appropriate for use in identifying good biological indicators or developing rapid models for assessing wetland condition and function.
  
  
• Systematic, repeatable, rapid procedures can be developed for assessing some of the disturbance gradients. This is a necessary precursor to selecting reference sites that will yield the most useful data.
  
  
• The biological metrics investigated or used should be appropriate to the study design and measurement protocols.
  
  
• Data suitable for identifying biological indicators of wetland condition can be collected simultaneously with data collected for calibration of HGM models. This does not necessarily require a great deal of additional training or field time.
  
  
• Shared field experiences are a good forum for sharing wetland knowledge among agencies, and among agencies and consultants and citizens. Shared field experiences lead to participants feeling more vested in the process of developing models and multimetric indexes. This informal "buy-in" can lead to greater willingness of participants to use the methods that ultimately result.