| dc.description.abstract | Wetland forest ecosystems have significant role in supporting human life, as a sources of water, food, biodiversity concervation, as well as a micro climate controller. Davies et al. (1995) noted that the extent of wetlands in Indonesia is about 38 million hectares. In addition, according to Komite Nasional Pengelolaan Ekosistem Lahan Basah (2004), Indonesia has about 40,5 million hectares of wetlands. While, according to Mulyani and Las (2008), Indonesia has about 40,2 million hectares of wetlands. The records of the extent of wetland in Indonesia is varied significantly due to different methods and data used. Thus, a development of remote sensing technology for monitoring of wetland ecosystems is required. In 2006, the government of Japan had launched ALOS (Advanced Land Observing Satellite) that carry a radar sensor called PALSAR (Phased Array Type L-band Shynyhetic Aperture Radar). ALOS PALSAR imagery is a new imagery and its interpretation methods has not been developed yet. The utilization of radar imagery data for forestry sector is also relatively undeveloped. The examination to evaluate the capability of the imagery is required. Considering the above mentioned matters, a study on the capability of ALOS PALSAR imagery for establishing identification techniques of forest wetland is required. The objective of the study is to evaluate the capability of ALOS PALSAR imagery for identifying wetland forest ecosystems. The research was conducted in Banjar, Barito Kuala, Banjarbaru, Banjarmasin, and Tanah Laut Regency, South Kalimantan Province. Identification of wetland forests was performed using ALOS PALSAR imagery, acquired in June of 2009 with 50 m resolution, with unsupervised classification method, separability analysis by transformed divergence (TD), and accuracy assesment. The hardware used in this study is a computer with Erdas Imagine Ver 9.1 and ArcView GIS Ver 3.2 softwares. This study found that wetland forest ecosystems that could be identified using ALOS PALSAR imagery with 50 m resolution include swamp forest with wet shrubs (bog) and mixed farms and also mangrove forest along with high shrubs. The accuracy evaluation showed that the accuracy for identification techniques of wetland forest are quite good having the overall accuracy and kappa accuracy are 73,49 % and 65,75 %, respectively. | en |
