Organisation géographique des populations de poissons marins

The management of marine fish stocks has stayed non-spatial, considering spatial distributions and their consequences on demography as implicit. But the impacts of climate change and the anthropisation of habitats now require to consider spatial procedures explicitly. The work presented here starts with the question of evaluating population abundance, which requires the modelling of spatial distributions (section 1). Then an ecological understanding of the multi-scale nature of distributions is investigated (section 2). Last, we study how demographic population dynamics depends on the spatial organisation of life cycles (section 3). In this work, the approach has been to combine data acquisition at sea with statistical characterisation and modelling. The application of geostatistics to fisheries survey data provided a solution to abundance evaluation and mapping for many different survey designs. A variety of models were implemented (stationary, non-stationary, non-linear, multivariate, spatio-temporal), depending on the statistical characteristics of the data and the problem posed. Sampling bias were identified by combining different survey methods. New procedures for assessing stocks based on indicators including spatial ones were developed, which prepared for ecosystem assessments. (...)