A facial recognition tool and legislative changes for improved enforcement of the CITES Appendix II listing of the humphead wrasse, Cheilinus undulatus

Humphead, or Napoleon, wrasse, (Cheilinus undulatus), is a large reef fish naturally vulnerable to over‐exploitation and highly prized as luxury seafood. It is predominantly consumed in Hong Kong and mainland China. Hong Kong is the global trade hub for this species. In 2004 (effective 2005), the humphead wrasse was listed on the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) Appendix II that substantially improved understanding and management of the species although illegal, unregulated and/or undocumented trade (IUU) continues. To understand patterns of trade in humphead wrasse, evaluate implementation of the CITES listing and identify ways to tackle IUU, especially in Hong Kong, data from multiple sources, including government databases between 1999 and 2017 and independent retail surveys between 2014 and 2017, were compiled. IUU trade into and through Hong Kong was detected, but declined following increased government enforcement in 2016. Nonetheless, fish numbers on sale continue to exceed imports with CITES permits and mainland China only sparsely reports imports of the species. Approximately 0.5–1.5 metric tonnes were traded internationally annually. Several legislative loopholes undermine enforcement into and within Hong Kong, including poor oversight of live fish carrier vessels and air imports, excessively long validity period for Licences to Possess in the city, non‐compliant traders, and legislative exemptions or omissions. The introduction of a ranched fish category in 2018 by Indonesia created additional challenges. More intelligence and risk analyses could help enforcers target their work. A major enforcement challenge is the inability to distinguish legally from illegally imported fish within Hong Kong. To improve fish tracking, identifying individuals using their complex facial patterns as fingerprints was explored; proof‐of‐concept was demonstrated using a mobile phone application: Saving Face. Facial marks were determined to be long‐lasting and different enough to distinguish among multiple individuals.