Delineating wetlands using geographic information system and remote sensing technologies

During the last century wetlands have considerably decreased. The principal cause is urbanization, especially in large urban regions such as the Houston area. In order to protect the remaining wetlands, they have to be monitored carefully. However monitoring wetland is a difficult and time-demanding task because it has to be done repetitively on large areas to be effective. This study was conducted to determine if Geographical Information System (GIS) and remote sensing technologies would allow accurate monitoring of wetland as a less time-consuming method. With this idea, a suitability model was developed to delineate wetlands in the Houston area. This model combined GIS and remote sensing technologies. The data used for this study were as high spatial resolution as possible and were generally easy to obtain. This suitability model consisted of four submodels: hydrology, soil, vegetation and multi- attribute. Each submodel generated a Wetland Suitability Index (WSI). Those WSI were summed to obtain a general WSI. The suitability model was calibrated using half of the study area. During calibration, the general model was evaluated as well as each individual index. Generally, the model showed a lack of sensitivity to changes. However, the model was slightly modified to improve the delineation of upland wet- lands by increasing the weight of the soil submodel. This model was validated using the second half of the study area. The validation results improved a bit compared to the calibration results; however they remained weak. It was demonstrated that the model does not favor riverine wetlands over upland wetlands, nor large size wetlands. The model ground truth data were evaluated and were suffciently proven to be up to date. Those results indicated that the weakness of the model must come from inac- curacy in the input data. Therefore, the study showed that while existing computing capacity supports remote delineation, spatial accuracy is still insuffcient to perform correct wetland delineation using remote sensing and GIS technologies.