Improved seasonal climate forecasting using shark-borne sensor data in a dynamic ocean

Abstract

Accurate ocean forecasts require sufficient observations to resolve key processes, yet conventional observing systems often miss fine-scale variability in dynamic ocean regions. Top predators frequently target these features, offering an opportunity for instrumented animals to sample underrepresented areas. Here, we use sharks equipped with depth- and temperature-sensing satellite tags as opportunistic ocean observers to reduce climate forecast errors in a proof-of-concept model experiment. We compiled >8200 high-resolution shark-derived depth–temperature profiles from the Northwest Atlantic Ocean and used these data to inform an operational forecasting model. Retrospective forecasts incorporating shark-derived observations showed up to 40% lower surface temperature error than control forecasts when compared against reference satellite observations and ocean reanalysis products. Forecast improvements from shark-derived measurements were strongest in dynamic shelf and slope regions that traditional observing approaches often under-sample. These results demonstrate the potential for animal-borne observations to strengthen operational forecasting and capture complex, ecologically important dynamics in a changing ocean.