For the past 20 years, increases in products and income of Thailand resulted from the increase of agricultural area rather than that of products per area due to the fast population growing rate correlated with unexpandable land for cropping and housing. Such situation of poverty and food scarcity forced villagers to migrate into the forest reserve area and then destroy the forest areas by means of shifting cultivation, especially in the watershed areas in the northern part of Thailand. During the period 1961-1985 the forest area was decreasing at an average rate of 0.5 million has annually. Among the major causes of forest depletion are illegal logging, cultivated area expansion, slash and burn activities. In 2001, the Royal Forest Department declared the forest area of Thailand being only 172,050.05 sqkm or 33.40 percent of the country's total area.

The method of forest inventory used in Thailand was systematic or non-random sampling. During the period 1955-1960 the "Camp-Unit System" in regional forest inventory in North Thailand was used by Loetsch (LOETSCH, 1957a) and after that the Royal Forest Department applied the Camp-Unit System and called it "Modified Camp-Unit System". This forest inventory technique was used as the main one between 1969-1976 and 1977-1981 (WACHARAKITTI, 1982) and in some areas "Line-Plot Systemi" and "Point Sampling" have been used with large area afterwards. However, all forest inventory techniques required many people, had high budgets and took a long time in the field for data measurement, but nevertheless mostly forest protected areas require the development of more rapid forest inventory methods for the national forest planning efforts to reflect accurately the rapidly changing forest parameters, patterns of land-use and shifting cultivation.

Remote sensing techniques have great utility for extensive land surveys where current forest resource information is scarce, too expensive or infeasible to acquire by alternative methods, the main aim of forest inventory being to define potential areas of the forest for future development. Remote sensing can reduce cost of forest inventory and monitoring if remotely sensed data are well correlated with the forest stand parameters such as diameter at breast height (DBH), percent crown cover, basal area and volume, and remotely sensed data are available when needed in the sampling design. The most frequently used remote sensing products continue to be from optical sensors that have a moderate spatial resolution (10 - 120 m). Examples include Landsat Thematic Mapper (TM) and Multi-Spectral Sensor (MSS), and SPOT High Resolution Visible (HRV), which are all multispectral sensors with three to six broad spectral bands. In this research, we have used Landsat-7 Enhanced Thematic Mapper Plus (ETM+) imagery as an example.

Khao Ang Rue Nai Wildlife Sanctuary, in which all the data will be collected for this study, includes some of the last remaining habitats of a number of wildlife species. As such these are indicators of high biodiversity, although the paucity of information on the sanctuary makes it difficult to be more specific. The sanctuary consists of an area of 103,000 ha. There are no people living within the sanctuary other than sanctuary staff and their families. Nevertheless pressure on the sanctuary is rising due to population pressure around the edges resulting in forest degradation through hunting, timber collection, fire and other agricultural or gathering activities. Development of forest inventory techniques with remote sensing for forest resources assessment should be established in Khao Ang Rue Nai Wildlife Sanctuary in order to find new forest inventory techniques which contribute with remote sensing data from a high resolution satellite and informative data applicable to the forest master plan or national forest policy planning with more efficiency for sustainable forest management.

The first purpose of the present study is to develop a sampling design and statistics of forest inventory in a tropical forest and in addition to contribute the remote sensing for measuring and assessing forest characteristics. The methods are based on measurement of ordinary forest characteristics completed with additional measurements and information about the occurrence in a small area.

The second objective is to estimate forest variables using satellite imagery such as Landsat-7 ETM+ and compare sample plot size and arrangement of old forest inventory techniques in Thailand with the new technique from the study.

The third objective is to achieve a new pattern of forest resource assessment by using new forest inventory techniques and remote sensing in order to better understand and improve the management of the forest area in the future.