Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6348640 | International Journal of Applied Earth Observation and Geoinformation | 2015 | 4 Pages |
Abstract
Tracking water level fluctuations in small lakes and reservoirs is important in order to better understand and manage these ecosystems. A geographic object-based image analysis (GEOBIA) method using very high spatial and temporal resolution optical (Pléiades) and radar (COSMO-SkyMed and TerraSAR-X) remote sensing imagery is presented here which (1) tracks water level fluctuations via variations in water surface area and (2) avoids common difficulties found in using single-band radar images for water-land image classification. Results are robust, with over 98% of image surface area correctly classified into land or water, R2Â =Â 0.963 and RMSEÂ =Â 0.42Â m for a total water level fluctuation range of 5.94Â m. Multispectral optical imagery is found to be more straightforward in producing results than single-band radar imagery, but the latter crucially increase temporal resolution to the point where fluctuations can be satisfactorily tracked in time. Moreover, an analysis suggests that high and medium spatial resolution imagery is sufficient, in at least some cases, in tracking the water level fluctuations of small inland reservoirs. Finally, limitations of the methodology presented here are briefly discussed along with potential solutions to overcome them.
Keywords
Related Topics
Physical Sciences and Engineering
Earth and Planetary Sciences
Computers in Earth Sciences
Authors
R.N. Simon, T. Tormos, P.-A. Danis,