Suitability Analysis Accuracy Assessment

3.3.2 Suitability Analysis

The suitability assessment was conducted by considering the land system analysis. According to Hardjowigeno et al. 2001, the suitability criteria for paddy crops are divided into highly suitable S1, moderately suitable S2, marginally suitable S3, and not suitable N. The factors that analyzed to classify the suitability class are limited only the constricted factors for paddy field in swamps area namely peat depth, pH, slope and salinity. Table 8 describes the suitability criteria of the constricted factors for paddy field in swamps area. Figure 10 shows the suitability analysis framework. Figure 10 Suitability analysis framework 30 Table 8 Suitability criteria for paddy field Factors S1 S2 S3 N Peat depth cm 50 100 100-150 150-200 pH 5.5-7.0 7.0-8.0 or 4.5-5.5 8.0-8.5 or 4.0-4.5 8.5 or 4.0 Slope 3 3-8 8-15 15-25 Salinity mmhoscm 3.5 3.5-5.0 5.0-6.6 6.6-8.0 Source : Hardjowigeno et al., 2001 Overlaying for vector datasets is conducted to obtain an aggregate of layers that determines the suitability. In vector data, the map overlay operation is done in pairs. For a more than two layers to be overlaid, it will be taken several steps. In this research includes four map layers. So, to make its process integrated into single process, the each vector data were converted into raster data in ArcGIS 9.3. After that by using weighted overlay processing with the same weighting for each suitability maps, the combination layer map was produced and can be treated as single map to the next processing.

3.3.3 Accuracy Assessment

Methods for automated land features mapping are concluded with an accuracy assessment of their result Congalton, 1991. The validation sampling points is used on this account. Accuracy assessment is done to verify the analysis and classification result using “point sampling accuracy” approach according to “confusion matrix”. The confusion matrix summarizes the relationship between two datasets that are classification map and reference information or alternative model. The accuracy assessment for this research use formulation below:

a. Omission Error as known as Producer’s Accuracy