抄録
In order to monitor ARD area (afforestation, reforestation and deforestation) as defined by the Kyoto Protocol Article 3.3, a method using remote sensing technology based on LANDSAT/TM data was examined. ARD was defined as land-use change, and not land-cover change acquired by TM data analysis, therefore two-step method was adopted. Firstly, land-cover change including ARD (LCC-ARD) was detected from TM data. Secondary, actual ARD as land-use change was detected from photo-interpretation of aerial photographs and SPOT-5/HRV-P data in the close investigation area, and conversion factor to revise LCC-ARD was estimated. LCC-ARD was detected by binarization of channel-7 (TM7) differences image. When accuracy of detection was verified by selecting the LCC-ARD distribution based on the photo-interpretation as truth data, we found that using TM data tended to over-detect LCC-ARD. This is because changing patterns of TM7 in LCC-ARD area and no-change areas were often similar in croplands, grasslands, cut-over lands and the mixed pixel areas of forest boundary. Next, the conversion factor of AR was estimated at 0.0010, and D at 0.0649. ARD area as land-use change was calculated by applying these values to LCC-ARD. This two-step method allowed for monitoring ARD objectively by using remote sensor data.
