Comparing Ecoregional Classifications for Natural Areas Management in the Klamath Region, USA Sarr, D. A., Duff, A., Dinger, E. C., Shafer, S. L., Wing, M., Seavy, N. E., & Alexander, J. D. (2015). Comparing Ecoregional Classifications for Natural Areas Management in the Klamath Region, USA. Natural Areas Journal, 35(3), 360-377. doi:10.3375/043.035.0301 10.3375/043.035.0301 Natural Areas Association Version of Record http://cdss.library.oregonstate.edu/sa-termsofuse R E S E A R C H A R T I C L E ABSTRACT: We compared three existing ecoregional classification schemes (Bailey, Omernik, and World Wildlife Fund) with two derived schemes (Omernik Revised and Climate Zones) to explore their effectiveness in explaining species distributions and to better understand natural resource geography in the Klamath Region, USA. We analyzed presence/absence data derived from digital distribution maps for trees, amphibians, large mammals, small mammals, migrant birds, and resident birds using three statistical analyses of classification accuracy (Analysis of Similarity, Canonical Analysis of Principal • Coordinates, and Classification Strength). The classifications were roughly comparable in classification accuracy, with Omernik Revised showing the best overall performance. Trees showed the strongest fidel- Comparing ity to the classifications, and large mammals showed the weakest fidelity. We discuss the implications for regional biogeography and describe how intermediate resolution ecoregional classifications may be Ecoregional appropriate for use as natural areas management domains. Index terms: Bailey ecoregions, Klamath Region, management domains, Omernik ecoregions, World Classifications Wildlife Fund ecoregions for Natural Areas INTRODUCTION Management in the • Increasingly, natural resource management involves multipartner collaboration across Klamath Region, landscapes and regions, often with complex arrays of administrative and ecological USA 2Wildlife Survey Data Management boundaries. Typically, the conceptual and Wildlife Science Division spatial domains for natural areas manage- Washington Dept. of Fish and Wildlife ment (hereafter management domains) are not clearly defined. Ideally, management 1,7,8 600 Capital Way North Daniel A. Sarr Olympia, WA 98501-1091 domains would be grounded in biogeog- raphy, and potentially relevant to a wide 1Klamath Network–National Park array of ecological patterns, processes, and Service 3US Geological Survey suites of species. When exploring alternate 1250 Siskiyou Boulevard 3200 SW Jefferson Way management domains for a region, natural Ashland, OR 97520 Corvallis, OR 97331 areas managers may wish to know answers to a number of practical questions, such as: (1) Which classifications best describe 2 ecologically significant boundaries? (2) Andrew Duff 4 Engineering, Resources and How do such boundaries correspond to 1 Eric C. Dinger Management Department biotic distributions of different species? Sarah L. Shafer3 Peavy Hall 215 And (3) how do patterns of biogeography Oregon State University Michael Wing4 and biodiversity affect conservation and Corvallis, OR 97331 management? Nathaniel E. Seavy5 John D. Alexander6 Large scale classification of biotic commu- 5Point Blue Conservation Science nities has long been a goal of naturalists and 3820 Cypress Drive #11 biogeographers, whose works have yielded Petaluma, CA 94954 a diverse array of maps and classifications • of the world’s flora and fauna (Wallace 1869; Holdridge 1947; reviewed in Welsh 7 Corresponding author: [email protected]; 6 1994). At the same time, climatologists (928) 556-7250; Fax: (928) 556-7092 Klamath Bird Observatory have been converging upon regional clas- 8 Current address: USGS Grand Canyon PO Box 758 sifications of climate, with a number of Monitoring and Research Center, 2255 N. Ashland, OR 97520 Gemini Dr., Flagstaff, AZ 86001 studies focused in the western United States (Baker 1944; Mitchell 1976; Mock 1996). Formal ecoregional classifications • have been developed in North America by US federal agencies (e.g., Bailey 1983, Natural Areas Journal 35:360–377 1998; Omernik 1987, 2011), with subse- quent refinement by states (Welsh 1994) 360 Natural Areas Journal Volume 35 (3), 2015 and nongovernmental entities such as the into integrated and holistic approaches tance of each factor will vary depending World Wildlife Fund (WWF; Ricketts et that cover many species simultaneously upon the life history of the organisms al. 1999). and span large spatial scales (Johnson et involved. Moreover, any classification al. 1999; Vander Schaaf et al. 2004). For used should ideally offer sufficient gen- The causes of variation among ecoregions managers, this broader focus requires an erality to be easily recognized and used include both biological and physical fac- understanding of the biodiversity and bio- by natural area managers. Thus, lessons tors, of which climate is usually considered geography of major species assemblages learned from evaluating ecoregional clas- to be an overarching control (Bailey 1998). in the areas they manage, as well as the sification schemes in this diverse region For plants, other abiotic factors, such as ability to make sound hypotheses of eco- will provide examples of challenges that geology, have long been known to be fun- managers in many parts of the world may logical responses to management actions. expect to confront. damental to the distributions of species and In complex landscapes with many land- communities (Merriam and Steineger 1890; owners, ecoregional classifications may Our initial survey of ecoregion classifi- Whittaker 1960; Walter 1973; Woodward be essential for developing meaningful cations revealed several alternatives to 1987; Ohmann and Spies 1998). For other ecological units for collaborative manage- compare for potential use in our region. taxa, these biophysical controls have typi- ment (Host et al. 1996) and bioassessment Of these, we chose the Bailey (Bailey cally been less clearly understood, but a (Hawkins et al. 2000). 1983, 1998), Omernik (Omernik 1987, number of studies suggest these controls 2011), and WWF (Ricketts et al. 1999) may be important (Root 1988; Hansen and Managers and applied conservation classifications. We also noted that although Rotella 2002; Shine et al. 2002; Duff and scientists often must develop and apply the existing ecoregional classifications Morrell 2007). Upon the coarse biophysi- standardized inventory and monitoring appeared to capture a number of broad cal template, finer habitat divisions arise techniques across large and heterogeneous landscape boundaries across the Klamath Region as a whole, they did not adequately from vegetation structure, successional landscapes. For instance, the Klamath Net- differentiate climate zones within the most variation, interspecific competition, and a work Inventory and Monitoring Program complex part of our region, the Klamath host of other biologically mediated effects of the National Park Service is currently Mountains subregion, where biodiversity (Grinnell 1917). Animals further subdivide developing a long term monitoring program is particularly high and climatic gradients the landscape into breeding territories, for six park units (three national parks, are especially pronounced. Therefore, in foraging and staging areas, and other two national monuments, and one national addition to evaluating the existing ecore- functional spaces (Palik and Engstrom recreation area) in the Klamath Region of gional classifications, we decided to ana- 1999). Consequently, ecoregional varia- northern California and southern Oregon lyze available climate data to see if more tion is best viewed as a multiscale concept that vary tremendously in climate, geol- refined climate zones could be identified containing a complex array of controls on ogy, plant and animal species, and a host for our study area. Our objective was to spatial patterns. of ecological processes (Sarr et al. 2007). compare the three existing ecoregional An integrated and rigorous monitoring pro- Recent research (Wright et al. 1998; classifications and two derived classifica- gram requires a basic understanding of how McDonald et al. 2005; Thompson et al. tion schemes for the Klamath Region using important functional relationships change 2005) has demonstrated that ecoregional a suite of biological and physical datasets across landscapes and where important boundaries may not correspond closely to compare their effectiveness in represent- boundaries or semi-homogeneous regions with compositional changes in individual ing ecologically significant boundaries for occur. Such an understanding also would be life forms, which may change at dif- our region. More specifically, we sought helpful for fostering collaboration among ferent rates across landscape gradients. to better understand the biophysical and reserves and with adjacent landowners to Therefore, it may be useful to explore address issues of mutual interest. biogeographic patterns within our region how differing ecological classifications by investigating the following four research fit varied species geography. For instance, To evaluate the utility of ecoregional questions: highly vagile organisms (birds) may show classification for describing ecological more dispersed distributions and partition patterns, processes, and suites of species, 1. Which ecoregional classifications best the landscape more coarsely than sessile we investigated the performance of several describe ecologically significant boundar-