Summary
The paper fits a Bayesian population model to data on abundance, bycatch and fishing effort to estimate the chance that each of three proposed gillnet closures, set out in Mexico's 2008 conservation plan, would let the vaquita population grow over ten years. The existing Refuge Area had a probability of success of 0.08, a larger second option 0.35, and a third option covering the whole range more than 0.99. The model estimates the population fell from 664 in 1993 to 209 in 2009.
Why it matters
It gave a quantified case for banning gillnets across the vaquita's whole range, years before such a ban was adopted.
Key findings
- Probability that abundance increases after 10 years: Option 1 (existing Refuge Area) 0.08, Option 2 (larger area) 0.35, Option 3 (whole range) >0.99.
- Median population fell from 664 in 1993 to 209 in 2009; in 2008 it was 214 (95% PI 135-326).
- Under Option 1 the population in 2018 was projected at 163 (95% PI 77-319); under Option 3 at 302 (95% PI 174-536).
- Median 2007 abundance was 226 vaquitas (95% PI 145-340). A 2008 PhD thesis by Jaramillo-Legorreta gave medians of 5,015 for 1920 and 93 for 2007 (p.E120).
- Success probability under Option 3 fell if bycatch ended late: 0.97, 0.89, 0.76 and 0.58 for ending one to four years after 2010.
Caveats
The data were sparse and imprecise, so the probabilities rest on a model with wide prior and posterior ranges. The 1920 median of 5,015 is not a result of this paper; it is quoted from Jaramillo-Legorreta's 2008 thesis, and the authors note that their own 2007 figure is higher and more precise because it includes the 1993 and 2008 abundance data.
Original abstract English
Bycatch in artisanal gill nets threatens the vaquita, Phocoena sinus, with extinction. In 2008 the Mexican government announced a conservation action plan for this porpoise, with three options for a protected area closed to gill net fishing. The probability of success of each of the three options was estimated with a Bayesian population model, where success was defined as an increase in vaquita abundance after 10 yr. The model was fitted to data on abundance, bycatch, and fishing effort, although data were sparse and imprecise. Under the first protected area option, the existing Refuge Area for the Protection of the Vaquita, bycatch was about 7% of population size, and probability of success was 0.08. Under the second option with a larger protected area, the probability of success was 0.35. The third option was large enough to eliminate vaquita bycatch and had a probability of success >0.99. Probability of success was reduced if elimination of vaquita bycatch was delayed or incomplete. Despite considerable efforts by the Mexican government to support vaquita conservation, abundance will probably continue to decline unless additional measures to reduce vaquita bycatch are taken, such as banning gill nets within the vaquita's range and developing effective alternative fishing gear.