Evaluation of mean composition values, trends and correlations of a 4000 km2 north Atlantic area considering data uncertainty: Nutrients, salinity and temperature

Abstract

The assessment of vast and heterogenous areas is demanding due, mainly, to the large uncertainty associated with limited collected information on the system. Larger systems should be more thoroughly characterized, and their models should express a conservative uncertainty of the reported information. This work presents a tool based on Monte Carlo simulation of monitoring data, applicable to up to 100 sampling points, that considers the uncertainty of collected samples and sampling positioning. This tool was successfully applied to assess the spatial distribution, seasonal variation and correlation of characteristic features of an oceanic area with approximately 4000 km² between two seasons. A metrologically significant reduction of temperature, salinity and some nutrient concentrations was observed, consistent with the typical upwelling regime on the western coast of continental Portugal, demonstrating that the developed methodology is robust to the extension and complexity of the spatial distribution of the system.