Spatial distributions of variables in marine environmental and fisheries research. Part 1: Geostatistics and autocorrelated environmental and fisheries data

Abstract

The geographical information very often is lost during calculation of indices for the abundance from surveys. Isoplet diagrams showing the distribution of both fish abundance and temperature/salinity are usually produced by hand by skilled personnel. In order to find an objective approach we explored a new stochastic method for abundance calculation and isoplet diagram drawing, using the geostatistics presented in this paper. The potential of geostatistical methods is to analyse the structural pattern of spatially dependent variables, model it and use it to estimate the unknown values at the unsampled locations. Moreover it gives the variance of the estimation error at a location and the variance of the global estimated value which other methods do not give. At each location we can store the estimated values of different spatial dependent variables for the purpose of studying the relationship among them and presenting graphic plots showing the geographic distribution pattem of each variable. Furthermore the structural pattem of each variable can be used to detect the interrelationship among them. This presentation demonstrates these points with the following data: Temperature, salinity, bottom depth, and 0-group cod density in the Barents sea.