Remote Sensing and GIS isn’t just for analyzing terrestrial environments. A study by Vahtmae and Kutser (2013) use geospatial data and regression models to gain an understanding of the health of the Baltic sea. The authors point out that, as one of the most polluted, the Baltic Sea’s ecological conditions need to be protected. Due to the lack of information, insufficient data, and scientific resources, environmental planning has been limited and not very effective (Vahtmae, Kutser, 2013).
The health of a marine environment can often be evaluated by measuring benthic macrophyte habitats. The authors indicate a reduction of the abundance of certain macroalgae, an indicator species in the Baltic Sea. By using large-scale analysis of marine habitats, they will be able to gain a broader understanding of marine habitats and provide evidence with which to gain environmental change.
Since the Baltic Sea is a shallow marine environment with several terrestrial imputs, particulate matter, dissolved organic matter and phytoplankton blooms stay suspended large areas of the sea. Image-based supervised classification techniques such as Maximum Likelihood, spectral Angle Mapper, can make it very difficult for certain types of remote sensing, not to mention time consuming and costly.
This article describes several alternative approaches. One such approach uses a modeled spectral library, in which remote sensing reflectances are compared to simulated reflectances. Though the known properties of the substrates and the optical properties of the water are necessary for this form of image classification and may negatively affect the accuracy. The third approach, is a basic classification method such as Spectral Angle Mapper, or SAM which has no sensitivity to illumination or albedo effects. With this method, one can change the accuracy and compare data with spectral library in the same instance.
The study tested both image-based approach and spectral library approach. Though the image-based method performed better than the spectral library method, they had to take into account that the turbidity of the water could change the results. In evaluation of the data, the suitability of the two methods for shallow water habitat mapping show basically no difference between airborne hyperspectral and satellite multispectral data (Vahtmae, Kutser, 2013).
Vahtmäe, Ele, and Tiit Kutser. "Classifying the Baltic Sea shallow water habitats using image-based and spectral library methods." Remote Sensing 5, no. 5(2013): 2451-2474.