Patterns and dynamics of coastal waters in multi-temporal satellite images:: support to water quality monitoring in the Archipelago Sea, Finland

The Archipelago Sea in the northern Baltic is a coastal region with A highly dynamic water flow regime, where the need for a methodology to monitor water quality accurately is encountered. In order to contribute to the development of an appropriate strategy to meet this need, the dynamics of the surface waters in the region were analysed using data from six Landsat TM/ETM + images from the late 1990s. Single images were enhanced by principal component transformation and multi-temporal image combination was based on unsupervised classification. The visual patterns discernible in the single images as well as the classification result of the multi-temporal data were compared with the reference data from long-term water quality analyses (Secchi disk depth and chlorophyll-a). The satellite images confirmed dynamic surface flow patterns in the region, indicating gradual and abrupt gradients in water quality, flow directions and forms. These patterns suggest short-term variability in the surface water quality within the region, presenting a challenge for water monitoring. The average Secchi depths increased gradually from the mainland coast until ca. 3-5 in near the open sea. The spatial distribution of chlorophyll-a was found to be more homogeneous, ca. 2-4 mug l(-1). The importance of evaluating the location of an in situ sampling site in relation to the surrounding hydrographic realm is pronounced in regions where field sampling and other monitoring methods remain insufficient in their ability to reflect water quality patterns credibly. In such areas, accurate and cost-effective water quality monitoring and forecasting require an integrated monitoring system, consisting of space and airborne surveillance, field surveys and hydrodynamic modelling. (C) 2004 Elsevier Ltd. All rights reserved.