GSMP research published in NATURE

Published today in NATURE are in-depth discussions about the global spatial risk assessment of sharks

In NATURE today we discuss our 2019 GSMP paper with two new sets of analyses.

Queiroz et al. (2021) Reply to: Shark mortality cannot be assessed by fishery overlap alone. Nature 595, E8-E16.

  • Using regression analysis on spatially matched data, fishing-induced pelagic shark mortality (CPUE) is greater where fishing exposure risk (shark/fishing overlap intensity) is higher, indicating spatial overlap of sharks and longline fishing is a major driver of capture susceptibility.
  • Fishing exposure risk hotspots overlap with shark density hotspots in 56% of tracked shark space use globally, and overlap with higher-than-average fishing effort in 81% of tracked shark space use.
  • Globally, only one-third of High Seas shark hotspot space was free from AIS-tracked longline fishing effort, indicating fishing effort overlapped with the majority of shark hotspots. Heavily fished regions had even lower levels of spatial refuge, e.g. 13% and 20% in Indian Ocean and N Atlantic.
  • When newer AIS data are used, spatial refuges from AIS longline fishing were even lower, at <25% of shark relative density hotspots, indicating our original spatial refuges were overestimated.


Queiroz et al. (2021) Reply to: Caution over use of ecological big-data for conservation. Nature 595, E20-E28.

  • Using newer releases of AIS fishing effort datasets does not alter the mean monthly spatial overlap (24%) with sharks globally (19-29%) or in Oceania (17-25%).
  • Randomly deleting 1% of grid cells containing longline fishing effort per region to simulate gear reclassifications does not change our results; only 7% of species/region pairs shifted from higher to moderate risk. So, even using much larger numbers of AIS vessel (gear) reclassifications than those proposed, results and conclusions of original paper do not change.
  • Simulating localized gear reclassifications we randomly deleted 1%, 5%, 25% and 75% of fishing effort grid cells – where shark spatial densities and fishing effort were both high and spatially extensive, random data deletion up to 75% of fishing effort grid cells did not change patterns of overlap and fishing exposure risk per region. But where tracking data was less extensive, patterns did change marginally above 25% data deletion.

Thanks to all our colleagues for engaging in these important and useful discussions!