In this paper, we present an automated classification method of landform elements using an application of SAGA GIS software. The spatial assessment was done on the Yambaru forest area (YFA) in the northernmost part of Okinawa Island, Japan. This task is performed through the detailed elevation grid analyses from DTM of YFA with a spatial scale of 10 × 10 m2 supported by The Geospatial Information Authority of Japan. The classification has ten classes; high ridges, midslope ridges, upland drainage, upper slopes, open slopes, plains, valleys, local ridges, midslope drainage and streams. Classes were defined using the ‘topographical position index’ module and selected terrain variables were integrated to vegetation data for site evaluation. Information on terrain characteristics is very important to explain geographical constraints and map variability of natural resources in maintaining sustainable forest management as well as supporting decision making processes. Taking this into account, we adapted a traditional concept of forest terrain introduced by Sai On, a council member of the Ryukyu Kingdom (former name of Okinawa Island) when evaluating the potential site for forestry use.
Conventional forest inventory practice took huge of effort, and is time- and cost- consuming. With the aid of remote sensing technology by light detection and ranging (LiDAR), those unbearable factors could be minimized. LiDAR is able to capture forest characteristic information and is well known for estimating forest structure accurately in many studies. Forest monitoring related to forest resource inventory (FRI) becomes more effective by utilizing LiDAR data and it is tremendously useful, especially to distinguish information on density, growth and distribution of trees in a selected area. In this study, LiDAR data was utilized aimed to delineate crown cover and estimate upper-storey canopy area in Yambaru Forest using object-based segmentation and classification techniques. Agreement between field survey and LiDAR data analysis showed that only 33.7% of upper-storey canopy area was successfully delineated. The low accuracy level of canopy detection in Yambaru Forest area was expected mainly due to tree structure, density and topographic condition.