Assessing change in the occurrence of rare species using the binomial distribution

The demand for comprehensive biodiversity assessments is increasing through the implementation of the ecosystem-based approach to management (EBM) of marine resources. Regional sea conventions such as the Oslo-Paris Commission (OSPAR) strive to implement EBM by developing an extensive status assessment program for the marine environment. Demersal fish communities are one ecosystem feature assessed by OSPAR through several ecological indicators. One of these indicators assesses the recovery in the population abundance of sensitive fish species, which was initially developed to report on the status of the sensitive fish community as a whole. However, for national reporting obligations, contracting parties of OSPAR (particularly for EU member states reporting to the Marine Strategy Framework Directive) prefer to have assessments for individual species. The previous indicator assessment relied on a suite of demersal fish species, which were caught frequently in scientific groundfish surveys, but did not provide assessment results for rare species caught in low frequencies. This study introduces a new assessment approach, the Binomial Occurrence Assessment (BOA), for the FC-1 indicator now renamed the “Recovery of sensitive fish species”, by applying the binomial distribution to relative occurrence data from scientific fisheries surveys. BOA uses occurrences in a reference period to determine boundaries for the expected occurrences in the recent (six year) assessment period of each survey. Significant changes in occurrence between the reference and assessment period, i.e. declines or recoveries, can then be detected when the observed occurrences in the assessment period fall outside of these boundaries. Methods to integrate the assessment results across multiple surveys are explored and compared since data on occurrences for fish species are available from more than one fisheries survey in each marine region considered by OSPAR. A case study on the sensitive demersal fish species in the North Sea exemplifies the applicability of BOA. Furthermore, assessments by BOA are compared to data from analytical stock assessments for those data-rich sensitive species that can support both approaches. Despite some shortcomings of BOA, such as the inability to detect declines of very rare species and the potential for occurrence metrics to differ from abundance metrics, the BOA allows an assessment of the status of a wide suite of fish species throughout the entire OSPAR region. The low data requirements of BOA allow its generic application to any other monitoring program that has captured occurrences of single species or species suites in the past and present.