Dataset: Mesoscale assessment of sedentary coastal fish density using vertical underwater cameras.
Description
Accurate and precise monitoring of the absolute density (i.e., number of fish per area or volume unit) of exploited fish stocks would be strongly advisable for deriving stock status and for designing proper management plans. Moreover, monitoring should be achieved at relevant (i.e., sufficiently large) temporal and spatial scales. This objective is particularly challenging for data-poor fisheries, as is often the case for recreational fisheries. Therefore, the feasibility of underwater video monitoring (vertical unbaited cameras) for estimating, as a proof of concept, the absolute density (and its ecological drivers) of a coastal sedentary fish species is demonstrated. The absolute density of a small serranid (Serranus scriba) targeted by recreational fishing was estimated along the southern coast of Mallorca Island (nearly 100 km). The median fish density ranged between 111 ind/km2 (Es Molinar) and 14,110 ind/km2 (Cabrera). Absolute density was correlated with fishing exposure, habitat, and depth. Site specific, seemingly long-term, effects of fishing exposure were negatively correlated with fish density, but short-term effects (assessed by the interaction between fishing exposure and before/after the season when recreational fishing occurred in the study area) were not detected. We suggest that the short-term effects of fishing may remain undetected because highly exploited sites could contain fish that are already not vulnerable to recreational fishing gear, irrespective of the short-term fishing pressure exerted. Such a process may explain some hyper-depletion patterns and should preclude the use of fisheries-dependent data for monitoring fish density. The results reported here indicate that monitoring fish abundance with vertical unbaited cameras at large spatial and temporal scales can be a reliable alternative for many species.
Files
Steps to reproduce
The script.R file contains the code for the analysis. The input.RData file is the data to run in the script and the estimates.RData file contains the results of the analysis.