Abstract
A 3-step land evaluation system using Landsat multispectral (MSS) data was proposed for regional assessment of agricultural development in the North Sumatera swamp area. The 3-step system enable to choose any of the steps depending upon the existing data and available image analysis facilities. In the first step, analogue analysis was proposed using Landsat images, when image analysis facilities are not available. In the second step, assuming that no ancillary ground truth data are available, 4 kinds of thematic maps all originating from Landsat MSS data were combined and evaluated using the PATTERN (Planning Assistance Through Technical Evaluation of Relevant Numbers) method. In the third step of the synthesized land evaluation system, thematic maps (500 meter mesh) based on existing ground truth were overlaid with maps originating from Landsat to evaluate the suitability of land for agricultural development using the PATTERN method. Results of the analysis were presented in evaluation maps with 6 suitability classes. These are shown in Photos 2, 3 and 4. Judging from the results, the proposed 3-step land evaluation system using the PATTERN method appears to be useful for land selection for agricultural development in developing countries lacking ground truth data.
