Determination of Land Productivity Dynamic Trend for and Degradation in Two Adjacent Micro Catchments

摘要

Land productivity is an expression of the bio-productivity resulting from all land components and their interactions for region-wide assessment, not just those related to human activities and direct use. Land productivity is therefore not to be confused with just agricultural productivity (Cherlet et al., 2013). The main goals of this study were to determine i) changing of land use and land cover, ii) vegetation cover density iii) trends in land productivity and iv) the situation of soil organic carbon (SOC) stocks in different land uses and land covers. This study was carried out in two adjacent micro catchments located in the Gediz River Basin in western Turkey. The study area covers about 16647 ha and its elevation changes between 70 m and 760 mabove sea level. In order to determination land cover and land use changes and NDVI, Landsat7 ETM+ and Landsat8 OLI/TIRS satellite images for May 2001 and May 2015 having about 30 m spatial resolution were used. For determination of the status of soil organic carbon levels, 320 soil samples were systematically collected from surface soils (0-30 cm) and geostatistical methods were used to generate an SOC distribution map for the study area. After image analysis, the amount of area that had changed for land use and land cover types was determined. From 2001 to 2015, forests increased about 321.8 ha (1.93%) due to afforestation, whereas shrubs, grasslands, and sparsely vegetated areas decreased about 6.35% due to conversion to artificial lands and croplands. Accuracy assessment is critical for maps generated from any remotely sensed data, and was the final step of the classification process. An error matrix is the most common way to present the accuracy of the classification results (Fan et al., 2007). The overall accuracy, user's and producer's accuracies, and the Kappa statistic were then derived from the error matrices for the various land use classes. Kappa analysis is a discrete multivariate technique used in accuracy assessments (Moller-Jensen, 1997). According to Congalton (1996), accuracy of the classification analysis result showed that the producer's accuracy and user's accuracy for individual classes of the 2001 and 2015 maps were greater than 80%, indicating strong agreement in this study.Percentage and hectare distributions of the land productivity dynamic (LPD) showed a comparable pattern within all land cover classes (Table 1, Figure 1). 21.86% of the territory showed a stable, not stressed, LPD.