Land suitability classification using different sources of information: Soil maps and predicted soil attributes in Jordan

Preparing large-scale soil maps, with adequate accuracy, is costly and time consuming. The low homogeneity of soil mapping units in small-scale maps limits their use for detailed planning. These limitations call for alternative approaches to support the land use planning processes. Recently, models that utilize terrain attributes derived from digital elevation models (DEM) are used to predict soil attributes. The objective of this research is to explore the accuracy of land suitability classifications derived from predicted soil attributes versus those derived from traditional soil maps. Three suitability maps were derived, one from predicted soil attributes and two from traditional soil maps (scales 1:10,000 and 1:50,000). The accuracy of the three suitability maps was verified using two methods: a set of 1469 field observations or an interpolation from these observations (Thiessen polygons). The results indicated that the accuracy of the suitability classification derived from predicted soil attributes compares favorably with the accuracy of classifications derived from traditional soil maps. The use of an interpolation between field observations provides better chance of estimating the accuracy of suitability map. The accuracy of suitability map is higher for land utilizations that tolerate wider ranges of land characteristics. The spatial distribution of the suitability classification derived from the predicted soil attributes indicated more realistic pattern. This research emphasized that the use of soil attributes derived from the prediction model provides an alternative source of soil information in areas where soil maps are not available.