NatureServe Conservation Status Assessments: Factors for Evaluating Species and Ecosystem Risk NatureServe Report Revised Edition April 2012 NatureServe is a 501(c)3 nonprofit organization dedicated to providing the scientific basis for effective conservation action. NatureServe and its hemisphere-wide network of natural heritage programs are the leading source for information about rare and endangered species and threatened ecosystems. Citation: Master, L. L., D. Faber-Langendoen, R. Bittman, G. A. Hammerson, B. Heidel, L. Ramsay, K. Snow, A. Teucher, and A. Tomaino. 2012. NatureServe Conservation Status Assessments: Factors for Evaluating Species and Ecosystem Risk. NatureServe, Arlington, VA. © NatureServe 2012 Cover photo: Polar bear (Ursus maritimus) © Larry Master (www.MasterImages.org) NatureServe 4600 N. Fairfax Dr., 7th Floor Arlington, VA 22203 703-908-1800 www.natureserve.org NatureServe Conservation Status Assessments: Factors for Evaluating Species and Ecosystem Risk Lawrence L. Master, Don Faber-Langendoen, Roxanne Bittman, Geoffrey A. Hammerson, Bonnie Heidel, Leah Ramsay, Kristin Snow, Andy Teucher, and Adele Tomaino Revised Edition April 2012 ACKNOWLEDGMENTS his 2012 revision to NatureServe workshop in November 2000. The ideas Tconservation status builds on the presented also draw upon discussion 2009 version developed by Element with and input from staff of the member Ranking Work Group. Members active programs of the NatureServe network. during the bulk of the 2009–2012 We especially thank Bill Nichols for work include the authors of this helping improve our global, national and report: Lawrence L. Master, Don subnational rank definitions. In addition, Faber-Langendoen, Roxanne Bittman, external data users and agency staff Geoffrey A. Hammerson, Bonnie Heidel, have provided much useful input. Leah Ramsay, Andy Teucher, and Adele Tomaino. Kristin Snow provides key The revision draws heavily from the data programming skills for the Rank Standards and Petitions Subcommittee Calculator and its interface with Biotics. of the IUCN Species Survival We thanks other members currently Commission and from the IUCN–CMP serving on the ERWG for their support, alliance to develop standard taxonomies including Bruce Young, Margaret Ormes, of threats and actions. Some of the and, most recently, Marilyn Anions. concepts incorporated here draw from Past members include Steve Rust, draft invasiveness assessment factors Paul Hendricks, Bryce Maxell, and Ben (Randall et al. 2001). Wigley. Previous revisions were done in consultation with Syd Cannings, Gwen Funding for 2009 revisions was Davis, Kathy Goodin, Paul Hendricks, generously provided by the National Kat Maybury, Larry Morse, Bryce Council for Air and Stream Improvement Maxwell, Leah Oliver, Donna Reynolds, (NCASI), Office Depot, U.S. Fish & Dale Schweitzer, Steve Taswell, Alan Wildlife Service, USDA Forest Service, Weakley, Troy Weldy, and Ben Wigley; and the Sarah K. de Coizart Article participants at the National Center TENTH Perpetual Charitable Trust. for Ecological Analysis and Synthesis We thank the generous support of an (NCEAS) workshops (2000–2004) anonymous donor for providing funds on methods for assessing extinction to revisit some key issues for this 2012 risk; and NatureServe ecologists at a revision. ii NatureServe PREFACE he NatureServe Conservation Status Response: It would be very challenging TAssessments: Factors for Assessing to assign all species to specific life- Extinction Risk was released in 2009 history strategies, since many species (Master et al. 2009), along with the are intermediate between the two Ranking Methodology document strategies. Furthermore, it was most (Faber-Langendoen et al. 2009). These straightforward to suggest that methods have greatly improved our population size should sometimes not approach to assessing the conservation be used in calculating a conservation status of species and ecosystems. In status for those species or species the years since its release, several groups in which either the life-history issues arose that required additional strategy led to very dynamic population review. The Element Ranking Working numbers from year to year, or for Group (ERWG) (see Acknowledgments) which it was very difficult to estimate was asked to review these issues population size. The calculator draws and propose solutions. Below, we upon other “rarity” characteristics when summarize the main issues and provide population size is not used. We have details on our proposed solutions that reworded the text for the Population guided our revisions. Size rank factor accordingly. There were five main issues that we Trends (Short-Term, Long-Term) felt needed review: Population size/ Issue: The 2009 rank calculator method scales, Trends, Area of Occupancy for for scoring trends is not ideal in that ecosystems, Threats, and the Biotics it does not adequately distinguish Upgrade to incorporate the revised rank decreasing trend from increasing trend, methods. and the calculator method did not seem to be sensitive to changes in trends Population Size unless they are extreme. Issue: We considered whether to incorporate a systematic evaluation Response: We found that the rank of life-histories (r-selected, k-selected calculator method for scoring trends is species), because, for example, not ideal for two reasons: the method some species can have very high, but for scoring trend and the insensitivity of dynamic, population numbers (e.g., the trends factors to changes in trend. annual plants, some insects), and others can have lower, less dynamic, Scoring of Trend population numbers (e.g., whales). Our typical approach to scoring rank Scoring both types with one set of factors sets the maximum score of 5.5, population size scales and weightings which if all factors are scored at that may not accurately assess extinction level, leads to a status of G5 (secure). risk. But the trends scale is different from the Factors for Evaluating Species and Ecosystem Risk iii rest in that the scale goes from varying concerning, the more at risk a species levels of decline (A–F), to stable (G), to is due to rarity factors, the greater the increasing rate (H–I). Thus, in principle, decline has to be in order to calculate a population or ecosystem with an a rank that is more imperiled. And the increasing trend could actually have converse is true as well: a presumptive its rank lowered (e.g., from G3 to G4). G5 with stable trends will get pushed To make that happen, the maximum towards G4. weighting of 5.5 should occur at G in the rating scale, and then should either be Under further review, it also became kept constant for H and I, or continue to apparent that ranks generated using increase above 5.5. the 2009 calculator, if different from the preexisting ranks in Biotics, tended Insensitivity of the Trend Factor to to vary more strongly on the side of Changes in Trend lowering the rank (less imperiled). We After testing the above adjustment now think this is in large part due to the to scoring of trend, a still more method for scoring the trends factors. fundamental issue remained: the trends factors, based on linear changes The Trend Subtraction Method in points, were not sensitive enough For all these reasons, we turned to to changes in trends. The reason for a new method for scoring trend, the lack of sensitivity is that the trend developed by Andy Teucher, a method categories are split finely (6 categories, we call the Trend Subtraction Method. some with only 10% spreads). The The method is based on the assumption result of having so many categories that a negative trend should move is that even with a pretty substantial a rank toward greater imperilment decline (30-50%, or even 50-70%), the (proportional to the size of the decline), assigned points for such a decline are and an increasing trend should likewise quite high (2.75 and 2.06 respectively), push a rank toward a more secure value. contributing to a smaller change The basic approach of this method is to in rank than may be expected for calculate an initial score based on rarity such a decline. These patterns were and threats factors, and then either not changed when various linear or subtract from that score when there is exponential weightings were tested. a negative trend or add to it when there is an increasing trend, to obtain a final Further investigation showed that the rank score. The procedure is as follows: effect of trend on the final rank was very dependent on the values of the other 1. An initial rank score is calculated factors, in that it had a centralizing from a weighted average of rarity tendency: it was biased against the and threats factors, with the rarity extremes, and tended to pull ranks score weighted 0.7, and threats away from G1 or G5 unless the trend weighted 0.3. value was also extreme (>90% decline 2. Trend scores (long- and short-term) or >25% increase). For example, if all are given values (points). These of the rarity factors combine to give are scaled exponentially from -0.50 a calculated value of 1.9, this would (>90% decline) to 0.00 (stable trend) place the species firmly in G2. Using and up to 0.14 (>25% increase). the original trend scoring, a trend of D 3. Long- and short-term trend values (50-70% decline) would actually make are multiplied by their weights, and the final calculated rank LESS imperiled, the weighted values are summed to because the trend score (2.06) is higher give a total overall trend score. Long- than the overall combined score from term Trend has a weight of 1, and the other factors (1.9). A decline of Short-term Trend has a weight of 2. 70% or more would be required to push the rank towards G1. Even more iv NatureServe 4. The total overall trends score is Area of Occupancy for Ecosystems added to the initial rank scorerarity / Issue: The guidance in the 2009 threats to give a final calculated rank description of Area of Occupancy for score.