Evaluation of in situ and remote sensing sampling methods for SPM concentrations, Belgian continental shelf (southern North Sea)

Abstract

Large sets of suspended particulate matter (SPM) concentration data from in situ and remote sensing (moderate resolution imaging spectroradiometer, MODIS) samplings in the Belgian nearshore area (southern North Sea) are combined in order to evaluate their heterogeneity and the sampling techniques. In situ SPM concentration measurements are from a vessel (tidal cycle) and from a tripod. During the tidal cycle measurements, vertical profiles of SPM concentration have been collected; these profiles have been used as a link between satellite surface and near-bed tripod SPM concentrations. In situ time series at fixed locations using a tripod are excellent witnesses of SPM concentrations under all weather conditions and may catch SPM concentration variability with a much finer scale. The heterogeneity has been statistically assessed by comparing the SPM concentration frequency distributions. Tidal cycle, tripod and MODIS datasets have different distributions and represent a different subpopulation of the whole SPM concentrations population. The differences between the datasets are related to meteorological conditions during the measurements; to near-bed SPM concentration dynamics, which are partially uncoupled from processes higher up in the water column; to the sampling methods or schemes and to measurement uncertainties. In order to explain the differences between the datasets, the tripod data have been subsampled using wave height conditions and satellite and tidal cycle sampling schemes. It was found that satellites and low-frequent tidal cycle measurements are biased towards good weather condition or spring-summer seasons (satellite). The data show that the mean surface SPM concentration derived from satellite data is slightly lower than from in situ tidal cycle measurements, whereas it is significantly lower than the mean SPM concentration interpolated to the water surface from the tripod measurements. This is explained by the errors arising from the interpolation along the vertical profiles, but also by the fact that satellite-measured signal saturates in the visible band used to retrieve SPM concentration in very turbid waters. © 2010 The Author(s).