Review of puma density estimates reveals sources of bias and variation, and the need for standardization. Murphy, S. M., Beausoleil, R. A., Stewart, H., & Cox, J. J. Global Ecology and Conservation, 35:e02109, June, 2022. ![Review of puma density estimates reveals sources of bias and variation, and the need for standardization [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex Range-wide status assessments of wildlife are critical to effective species conservation and management. Reliability of these assessments is contingent on having accurate and precise demographic estimates for local populations, but for large carnivores, such estimates are often biased, imprecise, or unavailable. Despite being the most widely distributed large carnivore in the Americas, little is known about the range-wide population status of the puma (Puma concolor). Population density is frequently the primary demographic metric used in puma conservation and management decision-making and policy; therefore, we conducted a comprehensive, range-wide, systematic review of capture-recapture and mark-resight model-based puma density estimates published through 2021 and used Bayesian multilevel models to investigate potential sources of bias and variation. Model-based puma density estimates have been produced in just 8 countries (42% of countries with puma populations) for study areas that cumulatively represent \textless 1% of extant puma range. Most estimates applied to small study areas (median = 265 km2), protected areas (70%), and represented high quality habitats, such as forests and mixed savannas (89%). Nonspatial models likely overestimated puma density by an average of 63%, and inclusion of dependent individuals (e.g., kittens) in detection histories resulted in density estimates that were, on average, ~33% higher than estimates for independent individuals only, highlighting the need for standardization. After correcting for those potential biases, range-wide mean and median densities were 1.81 and 1.63 independent pumas/100 km2 (95% CI = 1.62, 2.02), respectively, with a 95th percentile of 3.64 independent pumas/100 km2. Although puma densities did not differ between North and South America, between protected and unprotected areas, or among human disturbance severities, support existed for puma density varying at the landscape-scale as a function of multiple geographical, environmental, and climatic characteristics (e.g., biome, precipitation, vegetation quality, and elevation). However, most puma density estimates were imprecise (90% had CV \textgreater 0.20) and likely positively biased, primarily because of small study area sizes and issues associated with some sampling and analytical methods; for example, we observed a potential 31–33% overestimation of puma density when spatially unstructured genetic sampling was used. Consequently, the quality of many existing model-based puma density estimates may be inadequate for reliable conservation or management decision-making, and the current number and geographical extent of puma density estimates are likely insufficient to inform useful continental or range-wide status assessments for the species.
@article{murphy_review_2022,
title = {Review of puma density estimates reveals sources of bias and variation, and the need for standardization},
volume = {35},
issn = {2351-9894},
url = {https://www.sciencedirect.com/science/article/pii/S2351989422001111},
doi = {10.1016/j.gecco.2022.e02109},
abstract = {Range-wide status assessments of wildlife are critical to effective species conservation and management. Reliability of these assessments is contingent on having accurate and precise demographic estimates for local populations, but for large carnivores, such estimates are often biased, imprecise, or unavailable. Despite being the most widely distributed large carnivore in the Americas, little is known about the range-wide population status of the puma (Puma concolor). Population density is frequently the primary demographic metric used in puma conservation and management decision-making and policy; therefore, we conducted a comprehensive, range-wide, systematic review of capture-recapture and mark-resight model-based puma density estimates published through 2021 and used Bayesian multilevel models to investigate potential sources of bias and variation. Model-based puma density estimates have been produced in just 8 countries (42\% of countries with puma populations) for study areas that cumulatively represent {\textless} 1\% of extant puma range. Most estimates applied to small study areas (median = 265 km2), protected areas (70\%), and represented high quality habitats, such as forests and mixed savannas (89\%). Nonspatial models likely overestimated puma density by an average of 63\%, and inclusion of dependent individuals (e.g., kittens) in detection histories resulted in density estimates that were, on average, {\textasciitilde}33\% higher than estimates for independent individuals only, highlighting the need for standardization. After correcting for those potential biases, range-wide mean and median densities were 1.81 and 1.63 independent pumas/100 km2 (95\% CI = 1.62, 2.02), respectively, with a 95th percentile of 3.64 independent pumas/100 km2. Although puma densities did not differ between North and South America, between protected and unprotected areas, or among human disturbance severities, support existed for puma density varying at the landscape-scale as a function of multiple geographical, environmental, and climatic characteristics (e.g., biome, precipitation, vegetation quality, and elevation). However, most puma density estimates were imprecise (90\% had CV {\textgreater} 0.20) and likely positively biased, primarily because of small study area sizes and issues associated with some sampling and analytical methods; for example, we observed a potential 31–33\% overestimation of puma density when spatially unstructured genetic sampling was used. Consequently, the quality of many existing model-based puma density estimates may be inadequate for reliable conservation or management decision-making, and the current number and geographical extent of puma density estimates are likely insufficient to inform useful continental or range-wide status assessments for the species.},
language = {en},
urldate = {2023-07-05},
journal = {Global Ecology and Conservation},
author = {Murphy, Sean M. and Beausoleil, Richard A. and Stewart, Haley and Cox, John J.},
month = jun,
year = {2022},
keywords = {Terrestrial Ecoregions (CEC 1997)},
pages = {e02109},
}
Downloads: 0
{"_id":"8srYQMZHaS87AtwMr","bibbaseid":"murphy-beausoleil-stewart-cox-reviewofpumadensityestimatesrevealssourcesofbiasandvariationandtheneedforstandardization-2022","author_short":["Murphy, S. M.","Beausoleil, R. A.","Stewart, H.","Cox, J. J."],"bibdata":{"bibtype":"article","type":"article","title":"Review of puma density estimates reveals sources of bias and variation, and the need for standardization","volume":"35","issn":"2351-9894","url":"https://www.sciencedirect.com/science/article/pii/S2351989422001111","doi":"10.1016/j.gecco.2022.e02109","abstract":"Range-wide status assessments of wildlife are critical to effective species conservation and management. Reliability of these assessments is contingent on having accurate and precise demographic estimates for local populations, but for large carnivores, such estimates are often biased, imprecise, or unavailable. Despite being the most widely distributed large carnivore in the Americas, little is known about the range-wide population status of the puma (Puma concolor). Population density is frequently the primary demographic metric used in puma conservation and management decision-making and policy; therefore, we conducted a comprehensive, range-wide, systematic review of capture-recapture and mark-resight model-based puma density estimates published through 2021 and used Bayesian multilevel models to investigate potential sources of bias and variation. Model-based puma density estimates have been produced in just 8 countries (42% of countries with puma populations) for study areas that cumulatively represent \\textless 1% of extant puma range. Most estimates applied to small study areas (median = 265 km2), protected areas (70%), and represented high quality habitats, such as forests and mixed savannas (89%). Nonspatial models likely overestimated puma density by an average of 63%, and inclusion of dependent individuals (e.g., kittens) in detection histories resulted in density estimates that were, on average, ~33% higher than estimates for independent individuals only, highlighting the need for standardization. After correcting for those potential biases, range-wide mean and median densities were 1.81 and 1.63 independent pumas/100 km2 (95% CI = 1.62, 2.02), respectively, with a 95th percentile of 3.64 independent pumas/100 km2. Although puma densities did not differ between North and South America, between protected and unprotected areas, or among human disturbance severities, support existed for puma density varying at the landscape-scale as a function of multiple geographical, environmental, and climatic characteristics (e.g., biome, precipitation, vegetation quality, and elevation). However, most puma density estimates were imprecise (90% had CV \\textgreater 0.20) and likely positively biased, primarily because of small study area sizes and issues associated with some sampling and analytical methods; for example, we observed a potential 31–33% overestimation of puma density when spatially unstructured genetic sampling was used. Consequently, the quality of many existing model-based puma density estimates may be inadequate for reliable conservation or management decision-making, and the current number and geographical extent of puma density estimates are likely insufficient to inform useful continental or range-wide status assessments for the species.","language":"en","urldate":"2023-07-05","journal":"Global Ecology and Conservation","author":[{"propositions":[],"lastnames":["Murphy"],"firstnames":["Sean","M."],"suffixes":[]},{"propositions":[],"lastnames":["Beausoleil"],"firstnames":["Richard","A."],"suffixes":[]},{"propositions":[],"lastnames":["Stewart"],"firstnames":["Haley"],"suffixes":[]},{"propositions":[],"lastnames":["Cox"],"firstnames":["John","J."],"suffixes":[]}],"month":"June","year":"2022","keywords":"Terrestrial Ecoregions (CEC 1997)","pages":"e02109","bibtex":"@article{murphy_review_2022,\n\ttitle = {Review of puma density estimates reveals sources of bias and variation, and the need for standardization},\n\tvolume = {35},\n\tissn = {2351-9894},\n\turl = {https://www.sciencedirect.com/science/article/pii/S2351989422001111},\n\tdoi = {10.1016/j.gecco.2022.e02109},\n\tabstract = {Range-wide status assessments of wildlife are critical to effective species conservation and management. Reliability of these assessments is contingent on having accurate and precise demographic estimates for local populations, but for large carnivores, such estimates are often biased, imprecise, or unavailable. Despite being the most widely distributed large carnivore in the Americas, little is known about the range-wide population status of the puma (Puma concolor). Population density is frequently the primary demographic metric used in puma conservation and management decision-making and policy; therefore, we conducted a comprehensive, range-wide, systematic review of capture-recapture and mark-resight model-based puma density estimates published through 2021 and used Bayesian multilevel models to investigate potential sources of bias and variation. Model-based puma density estimates have been produced in just 8 countries (42\\% of countries with puma populations) for study areas that cumulatively represent {\\textless} 1\\% of extant puma range. Most estimates applied to small study areas (median = 265 km2), protected areas (70\\%), and represented high quality habitats, such as forests and mixed savannas (89\\%). Nonspatial models likely overestimated puma density by an average of 63\\%, and inclusion of dependent individuals (e.g., kittens) in detection histories resulted in density estimates that were, on average, {\\textasciitilde}33\\% higher than estimates for independent individuals only, highlighting the need for standardization. After correcting for those potential biases, range-wide mean and median densities were 1.81 and 1.63 independent pumas/100 km2 (95\\% CI = 1.62, 2.02), respectively, with a 95th percentile of 3.64 independent pumas/100 km2. Although puma densities did not differ between North and South America, between protected and unprotected areas, or among human disturbance severities, support existed for puma density varying at the landscape-scale as a function of multiple geographical, environmental, and climatic characteristics (e.g., biome, precipitation, vegetation quality, and elevation). However, most puma density estimates were imprecise (90\\% had CV {\\textgreater} 0.20) and likely positively biased, primarily because of small study area sizes and issues associated with some sampling and analytical methods; for example, we observed a potential 31–33\\% overestimation of puma density when spatially unstructured genetic sampling was used. Consequently, the quality of many existing model-based puma density estimates may be inadequate for reliable conservation or management decision-making, and the current number and geographical extent of puma density estimates are likely insufficient to inform useful continental or range-wide status assessments for the species.},\n\tlanguage = {en},\n\turldate = {2023-07-05},\n\tjournal = {Global Ecology and Conservation},\n\tauthor = {Murphy, Sean M. and Beausoleil, Richard A. and Stewart, Haley and Cox, John J.},\n\tmonth = jun,\n\tyear = {2022},\n\tkeywords = {Terrestrial Ecoregions (CEC 1997)},\n\tpages = {e02109},\n}\n\n\n\n","author_short":["Murphy, S. M.","Beausoleil, R. A.","Stewart, H.","Cox, J. J."],"key":"murphy_review_2022","id":"murphy_review_2022","bibbaseid":"murphy-beausoleil-stewart-cox-reviewofpumadensityestimatesrevealssourcesofbiasandvariationandtheneedforstandardization-2022","role":"author","urls":{"Paper":"https://www.sciencedirect.com/science/article/pii/S2351989422001111"},"keyword":["Terrestrial Ecoregions (CEC 1997)"],"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/zotero/NAAtlas2024","dataSources":["qLjf8q88GSLZ5dAmC"],"keywords":["terrestrial ecoregions (cec 1997)"],"search_terms":["review","puma","density","estimates","reveals","sources","bias","variation","need","standardization","murphy","beausoleil","stewart","cox"],"title":"Review of puma density estimates reveals sources of bias and variation, and the need for standardization","year":2022}