The Chitose region of Hokkaido, Japan was selected as the study area because of the large number of conifer plantations and old deciduous stands. Japanese Earth Resources Satellite-1 (JERS-1) data taken using the Optical Sensor (OPS) was acquired on April 17, 1993. The background and understory in the study area were still covered with snow, and the deciduous trees and shrubs had not yet opened their leaves for the approaching spring. The objectives of the research were to classify conifer tree species and to evaluate spectral differences between tree age classes and phenological stages (unopened, emerging, open) in coniferous stands. Representative conifer tree species forming closed cover were selected as sample classes. The forest cover type classification was produced with the maximum likelihood classifier. The results of the field survey and analysis indicated the following: (1) The near infrared band (band 3) was very useful for classifying vegetation. The differences between tree species were especially important. (2) There were spectral differences for the Japanese larch (Lalix leptolepis) depending on leaf phenological stage in spring time. (3) The mean VNIR bands' values for old stands of Sakhalin spruce (Picea glehnii) were lower than the mean value for young stands. (4) Bands 1 and 2, and bands 6, 7 and 8 showed a similar spectral pattern in the forest cover classes. (5) OPS has a high potential to classify forest cover type and distinguish between tree ages and phenological stages in coniferous stands.

View full abstract

A system to evaluate and classify the forest's activity by each sub-compartment was studied. The forest's function was investigated mainly based on topographical factors. By distinguishing the stands suitable and unsuitable for cutting, based on topography, site class, stand volume and so forth, they could be evaluated by a score evaluation method. As a consequence of doing a cluster analysis of the scores, it was decided that they should be appropriately divided into 3 types such as a timber production function, on the erosion protection function and a watershed protection function. On the other hand, as for the forest function for public health and recreation, a condition was set that people could use the water for recreational use. As a result, the various functions of the forest were divided into 4 types by adding a forest function for public health and recreation. In order to determine which function each sub-compartment belongs to, a system to classify them through the use of a decision tree method, was developed by making the best of use the GIS function. As a consequence of its adoption by the Shibata District Office of the National Forestry Bureau, (to the forests within its jurisdiction) it was confirmed that an evaluation of the forest's functions by each sub-compartment could be rapidly classified.

View full abstract

A method to evaluate the forest function in a small watershed area in every raster was shown. Concurrently, the influence of the raster size on forest function evaluation was analyzed by changing the raster size (10m, 20m, 50m). As a result of comparison with function classification in a small sub-compartment unit, from conversion accuracy of vector information and presumed result of inclination angle, it was considered that a raster size of 10m×10m is optimum for forest function evaluation in a small watershed area. As for comparison of a forest in a small sub-compartmentunit and raster unit, the areas for the timber production, public health and recreation, and selection cutting water source conservation functions increased. Conversely, a decrease in the areas for erosion production and reserved water source conservation functions was recognized.

View full abstract

This paper outlines three change detection methods, i.e., (1) direct multidate classification (MDC), (2) 12-dimensional multitemporal principal component (MPG) and (3) 2-dimensional multitemporal principal component, to detect specific changes in forest cover due to new artificial regeneration, natural growth and forest cutting. Of these 3 techniques, the MDC technique which simply classifies the multidate bands directly consistently provided better delineation of forest change. In the selection of band dimensionality, the 12-dimensional MPC is the most effective. Not only this technique can reduce the dimensionality of the original data, but also effectively picks up the change of interest and provides nearly as the first technique. For all date-pairs (3- to 7-year intervals) the changes due to forest cutting and new artificial regeneration and tree height growth of young plantations were detected; however, the height growth of the larger trees, i.e., S1-S2/3, P1-P2 and P2-P3, could not be detected. As indicated in this study, as the time interval increases, the ability of multidate TM to detect forest cover change increases. Within the young forest plantation, differences in the density of under-story vegetation of Japanese cedar seedlings sometimes led to misclassification.

View full abstract

In this paper, improvement cutting done in a broad-leaf secondary forest for recreational use, was appraised by a standard presented in this paper. This improvement cutting was done by the 'Forest for Safeguarding Living Environment Project (FSLEP)' in Gifu City. This standard, based on data collected in two plots of the study forest, shows correlations among the average DBHs of the trees remaining after cutting and other parameters of stand structure, when trees having a larger DBH in the descending order are given preference for saving over cutting. The other parameters are the STD of DBH, stand density, the number of species, the diversity index and the spatial pattern of trees. The average DBHs of the two plots after improvement cutting, were 8.7cm and 10.0cm respectively. The parameters of actual stands after cutting - except stand density - were bigger than those simulated by standards. Therefore, it may be concluded that this kind of improvement cutting enhances the variety of in-stand scenic view variation. Nevertheless, the spatial patter of remaining trees belonged to the Ca_1 type, both in the actual stands and the standard.

View full abstract

In pre-industrial societies, forests provide most daily necessities, ranging from fuelwood to construction material and from supplementary food to fertilizer. Forest resources become even more important as the firsthand source of energy and materials for both industrial and household uses when the society tries to industrialize itself, thus making the availability of forest resources one of the key factors for successful industrialization. In this paper, the present endowment of forest resources in the countries of South and Southeast Asia is analyzed in a historical and geophysical perspective so as to assess their potential for industrialization from a natural resources point of view, with a special emphasis on Thailand. Though at the heavy cost of deforestation, Thailand seems to have been one of the successful forerunners in the post-World War II Asian effort of industrialization, and its experience in the ordeal of deforestation and associated socioeconomic conflict should provide a good lesson for the rest of the countries in the region.

View full abstract

We developed a method to reconstruct past stand structures especially diameter distribution, by data obtained by investigations of present stands. On the assumption that regeneration takes place in gaps when upper-story trees die, we reconstructed past upper-story trees in past gaps. Past gaps were estimated from tree ring analysis and spatial distribution analysis of present trees. Past lower-story trees were reconstructed according to shading of the forest floor. It became possible to reconstruct past diameter distributions by repeating this procedure. We applied this method to a broad-leaved secondary forest stand in Tochigi Prefecture, Japan, and obtained apparently satisfactory results in reconstruction.

View full abstract

A project study on the development of new type of yield table has been carried out since 1989. The result of this study was reported as "System Yield Table" in 1992 and "Programs of System Yield Table" in 1995. This report discusses the question "What is the system yield table ?" by overviewing programs developed. The yield table of the type used widely for Japanese forest management and forest planning to date since the 19th century is called a "conventional yield table". The conventional normal yield table or empirical yield table shows the estimated values of forest stand growth, in case of specific tree species in a specific area and under specific management as variables of site, stand age and density. This is called the yield table of the first generation. A growth diagram of the stand prepared with the density effect of the stand as a basis is called a "forest stand density-control diagram". This diagram shows relationship corresponding to stand density, volume, tree height and diameter in the growth process of the stand on the plane. This can estimate the yield in the case of various density-controls. The conventional yield table of the first generation was line information on growth, while this density-control diagram would be plane information on growth, and is called the "yield table of the second generation". The basic characteristic of the system yield table is that it is a computer-program with an algorithm to estimate the growth process in the future in cases of various types of management for the forest stand under various conditions. The system yield table can express countless growth processes in countless stands in a multi-dimensional space formed by variables including tree diameter, height, volume, density, tree age, thus it is called the third generation yield table to indicate the whole space. Evaluation shall be made for the function level of the system yield table by explaining the condition of the forest stand, management method and growth process. Functions are explained by the following four items. (1) Quantity and quality of information on estimating forest stand condition, management and growth (information quantity). (2) Range of applicable forest stand (applicable range). (3) Usability of system yield table program (easiness of use). (4) Rationality of growth model used for estimation (logic).

View full abstract

A verification study has been done using JERS-1/SAR data for forest stand structure analysis. The initial survey conducted at a relatively flat forest at in the Mt.Fuji area showed that there was a relationship between the radar backscatter intensity and the forest tree parameter in the azimuthal slope direction. In the secondary survey, investigation of the relationships between forest tree parameters (tree height, diameter at breast height (DBH), stand volume, stand density, standcoverage etc.) and the SAR data showed that JERS-1/SAR data were highly applicable for analyzing forest stand structure. To obtain normalized backscattering coefficients with exclude topographical effects for any specified direction, we must develop an estimation formula for key parameters of stand analysis.

View full abstract