<p>In this study, we tested the potential of photo-identification as an identification method for bull trout using I<sup>3</sup>S Classic, a program that utilizes natural markings to identify individuals. To determine the ability to identify individual bull trout from spotting patterns, we used photographs of bull trout captured in 2017–2019 by the Idaho Department of Fish and Game at Fourth of July Creek, Idaho. During weir operations, 132 unique bull trout were sampled of which 34 fish were recaptured (<i>n</i> = 166 total captures). Using pictures of these fish, we created “spotting fingerprints” in the I<sup>3</sup>S system for each fish and compared “fingerprints” to identify recaptures. In a blind test, three readers used I<sup>3</sup>S to determine which fish were recaptures and which were not. The three readers correctly assigned a mean of 92.7% (SD = 1.6%) of the bull trout as recaptured fish or fish captured for the first time. For the recaptured fish, the three readers were able to correctly identify 72% (SD = 6%) of the fish; however, the correct identification rate increased to 87% (SD = 7%) after removing fish with poor image quality that were misidentified by all readers. The readers correctly identified 100% of the fish that were not recaptured, i.e., the false-positive rate was zero. If standardization of photograph quality and spot selection can be established, photo-identification could be a complementary tool that could be used to monitor the population dynamics of bull trout.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Utilizing the spotting patterns of bull trout to identify individuals in photo-identification software

  • Svyettlana Parry,
  • Kat Gillies-Rector,
  • Eric J. Billman

摘要

In this study, we tested the potential of photo-identification as an identification method for bull trout using I3S Classic, a program that utilizes natural markings to identify individuals. To determine the ability to identify individual bull trout from spotting patterns, we used photographs of bull trout captured in 2017–2019 by the Idaho Department of Fish and Game at Fourth of July Creek, Idaho. During weir operations, 132 unique bull trout were sampled of which 34 fish were recaptured (n = 166 total captures). Using pictures of these fish, we created “spotting fingerprints” in the I3S system for each fish and compared “fingerprints” to identify recaptures. In a blind test, three readers used I3S to determine which fish were recaptures and which were not. The three readers correctly assigned a mean of 92.7% (SD = 1.6%) of the bull trout as recaptured fish or fish captured for the first time. For the recaptured fish, the three readers were able to correctly identify 72% (SD = 6%) of the fish; however, the correct identification rate increased to 87% (SD = 7%) after removing fish with poor image quality that were misidentified by all readers. The readers correctly identified 100% of the fish that were not recaptured, i.e., the false-positive rate was zero. If standardization of photograph quality and spot selection can be established, photo-identification could be a complementary tool that could be used to monitor the population dynamics of bull trout.