COA Methodology

ODFW COAs were originally developed for the 2006 Conservation Strategy with the best available information at the time and an intention for them to be updated as new information became available. Agencies, conservation organizations, and stakeholder groups have since indicated that COAs were one of the most heavily utilized components of the Strategy (10-Year Report), helping to prioritize on-the-ground conservation actions statewide. In response to staff and partner requests, the ODFW re-analyzed COA boundaries for the 2016 Conservation Strategy, using new and updated science, data, and resources.

To continue the success of the 2006 COAs, the same definition, concept, and general datasets were used in the 2016 analysis. Improvements focused on:

  • Incorporating new and updated data within the modeling analysis
  • Conducting a more precise and better documented analysis
  • Working with internal and external technical experts statewide
  • Standardizing COA profiles to provide clearer connections between COAs and Strategy Species, Strategy Habitats, and Key Conservation Issues

This section presents a description of each step in the COA revision process: Spatial Modeling Analysis, Spatial Modeling Analysis Review, COA Boundary Delineation, and COA Profile Development.

Spatial Modeling Analysis

The process to update the 2016 COA boundaries began with a rigorous spatial analysis, using a conservation prioritization and spatial modeling program called Marxan. Marxan provided decision support with the design of conservation areas, using best available data to focus on concentrations of Strategy Species, Strategy Habitats, and additional datasets related to selected Key Conservation Issues.

The Marxan spatial modeling analysis involved the following steps:

  • Compiling data into assessment units
  • Setting Marxan goals for all Strategy Species, Strategy Habitats, and additional datasets
  • Establishing a suitability index
  • Variable calibration
  • Marxan analysis

Compiling Data Into Assessment Units

The first step in the planning process was to select conservation targets, based on best available datasets. The conservation targets are elements of biodiversity (i.e., plants, animals, and habitats) and were limited to those elements included in the list of Strategy Species and Habitats, or related to Strategy Key Conservation Issues. In November 2014, ODFW Conservation Strategy development staff convened a Technical Advisory Team, including representatives from the Institute for Applied Ecology, Klamath Bird Observatory, Natural Resources Conservation Service, ODFW Fish Division, Oregon Biodiversity Information Center (ORBIC), TNC, The Wetlands Conservancy, and U.S. Fish and Wildlife Service (USFWS). The Technical Advisory Team discussed available conservation targets and associated datasets to emphasize within the 2016 COA analysis.

The following datasets were determined to provide the best available information on each associated Conservation Strategy component:

Strategy Species: Wildlife (Amphibians, Birds, Mammals, Reptiles)
Strategy Species: Fish
Strategy Habitats
Key Conservation Issue: Climate Change
Key Conservation Issue: Disruption of Disturbance Regimes
Key Conservation Issue: Barriers to Animal Movement
Key Conservation Issue: Land Use Changes

All available data for the established conservation targets were then attributed to assessment units, wall-to-wall polygonal features from which the conservation portfolio was constructed. The 2006 COA analysis used hydrologic unit code (HUC) level 6 watersheds, which are relatively coarse and vary greatly in size and shape across the landscape. The 2016 COA analysis used a finer resolution (1 square mile hexagon grid of assessment units) for a more statistically sound analysis, which is the same base data used in the Western Association of Fish and Wildlife Agencies Crucial Habitat Assessment Tool (WAFWA CHAT), the ODFW Crucial Habitat Assessment work presented in the ODFW Compass, and the decision support system currently under development within the Oregon Sage-Grouse Conservation Partnership.

Setting Marxan Goals for all Strategy Species, Strategy Habitats, and Additional Datasets

After selecting targets and compiling all available data, the next step of the spatial modeling process involved setting goals for the number of occurrences and geographic distribution of each target. These goals are based on the amount and distribution of each target across the geography, factoring in target rarity and degree of endangerment to ensure that each target is treated equally. Targets should be represented in multiple COAs (where possible) as a hedge against stochastic events (e.g., disease, fire) and to buffer against the anticipated impacts of climate change, with an overarching intention to provide for long-term viability of Strategy Species and Habitats.

Strategy Species and Habitat goals were generated using an overall range of 30 percent (recommended minimum amount of habitat needed to sustain imperiled populations) to 60 percent (recommended maximum amount of habitat to be included while still prioritizing distinct areas). The goal range was further adjusted to account for additional data characteristics if needed, with more reliable data assigned higher goal percentages. Specific goals were then established based on the following matrices, with an emphasis on assigning higher percentages to Strategy Species and Habitats declared in fewer ecoregions. The overall objective was to normalize Strategy Species and Habitats throughout the COA analysis, providing higher goals for Strategy Species and Habitats that were represented in only one ecoregion, and lower goals for Strategy Species and Habitats represented in all ecoregions.

Strategy Species: Wildlife Goals

The total number of ecoregions where each species was designated a Strategy Species compared to the species’ conservation status. NatureServe State Rank (“S-rank”) was used as a consistent measure for conservation status (NatureServe). Goals were established for both species observations (higher percentage) and species habitat distribution modeling (lower percentage). ‘XX’ indicates that no Strategy Species occurred within a particular combination.

S-rank Number of Ecoregions
1 2 3 4 5 6 7 8
S1 60%/30% XX XX XX XX 50%/25% XX XX
S2 60%/30% 50%/25% 50%/25% 50%/25% 50%/25% 40%/20% 40%/20% 40%/20%
S3 60%/30% 50%/25% 50%/25% 50%/25% 40%/20% 40%/20% 40%/20% 40%/20%
S4 50%/25% 50%/25% XX XX 40%/20% XX XX XX
S5 50%/25% XX 40%/20% XX XX XX XX XX

Strategy Species: Fish Goals

The total number of ecoregions where each species was designated a Strategy Species compared to the species’ conservation status. The priority rank established within the Aquatic Species of Concern Crucial Habitat Assessment was used as a consistent measure for conservation status.

 Aquatic SOC (CHAT) Priority Number of Ecoregions
1 2 3 4 5 6 7
1 50% 40% 40% 40% 30% 30% 30%
2 50% 40% 40% 30% 30% 30% 30%
3 40% 40% 40% 30% 30% 30% 30%

Strategy Habitat Goals

Habitat goals were established for each Strategy Habitat based on qualitative assessments of conservation status, impairment, data quality, and the number of ecoregions containing each habitat. Additional data related to sagebrush and wetland habitats were incorporated to further prioritize especially high value habitat locations.

 Habitat Ecoregion
BM CR CP EC KM NBR WC WV
Aspen Woodlands 60% 60% 60%
Coastal Dunes 50%
Estuaries 40%
Natural Lakes 40% 40% 40% 40% 40% 40% 40% 40%
Grasslands 50% 50% 50% 40% 40% 40%
Late Successional Mixed Conifer Forests 40% 40% 40% 40% 40%
Oak Woodlands 60% 60% 60% 60% 60%
Ponderosa Pine Woodlands 40% 40% 50%
Riparian and Flowing Water Habitats 40% 40% 40% 40% 40% 40% 40% 40%
Sagebrush Habitats 50% 40% 50% 50%
Sagebrush Habitats – Impacted by Fire, Invasives, or Landscape Integrity 30% NA 30% 30%
Wetlands 40% 30% 40% 40% 30% 40% 30% 30%
Wetlands – High Conservation Significance 60% 50% 60% 60% 50% 60% 50% 50%

Remaining dataset goals were established primarily based on data quality: confluences (30 percent), floodplains (20 percent), and topo-climate diversity (40 percent).

Establishing a Suitability Index

Suitability values denote the “cost” of conservation, or the impediments to conservation, and allow the analysis to emphasize areas most suitable to conservation. Suitability was estimated for each assessment unit, using a consistent formula factoring in values based on barriers to animal movement and land conversion/use (i.e., TNC Resistance and Species Permeability Models) and land protection status (i.e., USGS GAP Protected Areas Database) within each assessment unit. Suitability ranges were scaled among ecoregions to provide a standardized minimum, maximum, and range of suitability values within all ecoregions across Oregon.

Variable Calibration

Marxan spatial modeling software requires variables to customize the modeling analysis. Two of the most important variables are the: 1) Number of Iterations, which defines the number of times that Marxan will duplicate a random seeding and analysis and 2) Boundary Length Modifier, which is used to determine the relationship of the size of conservation areas versus number of distinct conservation areas (e.g., few large areas, or many smaller areas). Scripts were utilized to find the optimal calibration settings for these variables within each ecoregion.

 

Calibration results:

Ecoregion Number of Iterations Boundary Length Modifier
Blue Mountains 2,000,000,000 0.16
Coast Range 600,000,000 0.11
Columbia Plateau 1,000,000,000 0.11
East Cascades 650,000,000 0.14
Klamath Mountains 900,000,000 0.14
Northern Basin & Range 2,000,000,000 0.16
West Cascades 500,000,000 0.19
Willamette Valley 600,000,000 0.11
Statewide 2,000,000,000 0.14

Marxan Analysis

Separate Marxan runs were first performed within each ecoregion. These results were then used as a seed to run a final statewide Marxan model.

ODFW would like to acknowledge:

TNC, Portland, OR: provided guidance, data processing, and general assistance with the data, modeling analysis, and documentation. Some of the text within the COA Methodology page has been taken verbatim, with permission, from the TNC report “Conserving Nature’s Stage: Identifying Resilient Terrestrial Landscapes in the Pacific Northwest.”

USFWS, Portland, OR: provided software training, automated scripts and tools, data processing, and general assistance with the data, modeling analysis, and documentation

Spatial Modeling Analysis Review

The results of the spatial modeling analysis were reviewed by ODFW Fish and Wildlife Biologists as well as the Stakeholder Advisory Committee convened by the ODFW for the Conservation Strategy update. Numerous review sessions were held across the state with detailed presentations and in-depth discussions. An online mapping application and PDF map of model results were provided to ODFW field staff and the Stakeholder Advisory Committee, along with multiple formats to provide review and input. All comments for specific areas were converted into a spatial format and assigned to specific assessment units within the COA analysis.

COA Boundary Delineation

The final step in producing the revised COA boundaries was to compile all available information into a single boundary dataset. Layers utilized during this process included:

The ODFW Conservation Opportunity Area Revision Team examined the Oregon landscape and selected the final square-mile assessment units to include in the 2016 COA boundaries using the following criteria:

  • Maintain any assessment units that were highlighted within the solution of the ODFW 2016 COA Spatial Modeling Analysis and were included as a 2006 ODFW COA.
  • Include assessment units that were highlighted within the solution of the ODFW 2016 COA Spatial Modeling Analysis and were approved during the review process by ODFW field staff and/or Conservation Strategy stakeholders.
  • Maintain any assessment units that were highlighted within the solution of the ODFW 2016 COA Spatial Modeling Analysis and are included as an ODFW Sage-Grouse Core Area.
  • Include 2006 ODFW COAs that were requested to be maintained by ODFW field staff and/or Conservation Strategy stakeholders.
  • Include assessment units that included at least three of the layers listed in the paragraph above.
  • Remove all Wilderness Areas from COAs.
  • Connect COAs as much as possible, often following streams or similar Strategy Habitat corridors.
  • Do not exceed more than 35 percent of the Oregon landscape.
  • The 2006 ODFW COAs encompassed approximately 26 percent of the Oregon landscape. Because of recent data improvements, a higher percentage was deemed acceptable for the 2016 analysis. Through discussions with agency partners and the Technical Advisory Team, it was determined that anything exceeding 35 percent of the landscape would negate the purpose of COAs and reduce their chance for success.
  • Do not exceed more than 40 percent within an individual ecoregion, to ensure that focusing on specific areas was not minimized within a given ecoregion.
  • Maintain a mix of public and private land (include at least 30 percent private land, statewide).
  • The final COA boundaries should meet at least 75 percent of the Marxan goals set during the spatial modeling analysis.

After selecting hexagons to include in the final solution, boundary lines were then drawn to divide areas into distinct COAs, following similar habitats and/or watersheds where possible. Each COA was given a unique name and COA ID. Some 2016 COAs follow the same boundaries as the 2006 COA, or TNC Willamette Synthesis COAs, instead of having a hexagonal boundary. All 2016 COA boundaries were clipped to the Oregon state boundary.

 

Figure depicting the ODFW COA boundary delineation process.
Photo Credit: ODFW. 1. Layers including the 2006 ODFW COAs (depicted in black), COA Modeling Analysis results (orange), and ODFW Sage-Grouse Core Areas (red) were overlaid to determine commonly prioritized landscapes.
2. COA boundaries (shown here in purple) were then delineated where the modeling analysis results overlapped areas highlighted in previous prioritization efforts.

COA Profile Development

Initial profiles for each COA were developed using the data from the COA analysis, providing detailed lists of Strategy Species and Habitats that have been documented within each COA. ODFW staff manually recorded remaining profile information. COA profiles will continue to be refined with new and updated information.