Abstract
Previous studies have assessed artificial landform transformation by measuring anthropogenic impacts on landform changes such as erosion and deposition. On the other hand, almost no previous studies have quantitatively evaluated the difference between modified landform and natural landform except Yamada (2001b)’s “topographic naturalness” assessed morphology mountains and hills. However, the method proposed by Yamada (2001) can only assesses morphology of mountains and hills partitioned by closed contour lines. This study, therefore, investigated a method of assessing and analyzing “topographic naturalness” which is not partitioned by closed contour lines, based on the geomorphometry using digital elevation model (DEM). The study area is northwestern Tama hills which have been modified artificially.
In this study, we paid attention to the fact that the anomaly (TSD) and increment (I) of trend surface analysis have negative correlation in mountains and hills. We then carried out regression analysis to explain the increment (I) by the anomaly (TSD), by using past DEM in 1962 as reference data. We regarded the residual between the calculated anomaly (TSD) and observed value as the standard to evaluate “topographic naturalness”. If the calculated residual is larger than -2RSE (RSE: twice of the standard error of the residual), we evaluated the topography as “natural”. If it is smaller than -2RSE, we evaluated it as “unnatural”.
By applying this regression equation to the present DEM in 2003, we obtained the following three results. (1) At hills, “unnatural” topography occupied about 41.9% of all of the grid points in the present DEM whose topography is “unnatural” with the residual smaller than -2RSE. This characteristic was not found in the past DEM in 1962. (2) In the southern part of the study area, those “unnatural” grid points occupied almost all of them, while in the northern part, they were mainly distributed in Nan’youdai 1-3 Choume, and Koshino and Horinouchi. (3) All of these “unnatural” areas have been developed for housing lands, which implies that the increase of “unnatural” topography in the study area is derived from anthropogenic landform changes.
In conclusion, we can evaluate the “topographic naturalness” with satisfactory accuracy if the study area is wide enough. Namely, this method is superior to the one proposed by Yamada (2001) which is only applicable to the area partitioned by closed contour lines.
