Statistical monitoring of spatial patterns of environmental indices for integrated ecosystem assessment: Application to the Bay of Biscay pelagic zone

Abstract

Monitoring the environment of fish is a key component of the ecosystem approach to fisheries management. Here, we propose a methodology to statistically monitor time series of maps of environmental indices. These maps were derived from a 37-year hindcast of a coupled physical–biogeochemical model. The space–time variability in the maps was decomposed using empirical orthogonal functions into time-invariant spatial patterns and time-varying amplitudes for these patterns. A statistical process control chart was then applied to the time series of the amplitudes. In that way, changes were detected with known statistical performance. The monitoring system also specified how long the changes lasted and which years, seasons and zones were affected. Results for all indices were assembled in an integrative dashboard of the detected deviations. For illustration, the procedure was applied to the Bay of Biscay pelagic zone. The selected environmental indices characterized the evolution of hydrological structures such as fronts and river plumes, as well as changes in temperature, water column stratification, horizontal current flow and primary production. A major result was that, in the last decade, sea surface temperature showed repeated significant shifts towards warming, which were largest in the northern half of the Bay, while the spatial extension of river plumes over the shelf alternated between wet and dry years. From 2005, several other indices showed repeated significant deviations: increase in sea bottom temperature, increase in the depth of the pycnocline and changes in coastal currents. The procedure provided an integrated view of ecosystem variability and change for all its components and their spatial organization.