Integrating satellite imagery and geostatistics of point samples for monitoring spatio-temporal changes of total suspended solids in bay waters: application to Tien Yen Bay (Northern Vietnam)

Abstract

High concentrations of total suspended solids (TSS) in coastal and bay areas strongly affect water quality and aquatic ecosystems. Frequent monitoring of spatiotemporal changes of TSS distribution in such areas is indispensable for understanding sediment fluxes and water cycles, and to preserve ecosystem health. This study aimed to identify appropriate and sufficient tools for assessing changes in TSS distribution in Tien Yen Bay in Northern Vietnam, a typical closed bay, which was selected as the study area because of its rich biodiversity. Initially, a physical based model using the reflectance at the sea surface was developed for estimating TSS concentrations from satellite image data, and a model with an exponential function was identified as suitable for the estimation. This model was fitted appropriately to provide a relationship between reflectance from the MODIS/Terra band 1 (visible red) after the atmospheric correction and the in situ TSS concentrations at 40 points. Ordinary kriging was then shown to be effective in improving the spatial resolution of the MODIS/Terra image-based estimation of the TSS concentration at a 250 m interval, because it could detect TSS variation in detail, in particular in the local estuaries. TSS distributions derived from 12 MODIS/Terra images from November 2009 to October 2010 clarified seasonal changes in TSS during one year. TSS concentrations were high during summer and lower during fall and winter. Such trends were conformable with the hydrodynamics in Tien Yen Bay. Consequently, the proposed method was more effective for TSS estimation than traditional methods using satellite image data only.