Abstract
This paper presented the methods and results of pine moth decrease investigation in Big Xingan Mountain in the summer of 1990 using AVHRR data of meteorological satellite. Considering the possible effects of natural geographical environment and other diseases to the investigation, the areas and egress of insect pest were determined by analyzing RVI (Ratio Vegetation Index) map combined with field sampling investigation. In final the results of satellite and field investigation were anlaysed by contrast.

Introduction
The plant diseases and inset pests of forest are threatening the living of forest heavily. They reduce the quality of woods and affect the renewing of trees. People call them no-smoke forest fire because the economic losses for them reach or exceed that for forest fire. Therefore the areas and degrees of insect pests must be found out as earlier as possible, in order to make the relevant tactics of prevention and cure. Many advanced countries like USA, Canada and Finland lead advanced spacelight, aircraft and computer technology into the applied field, and have achieved notable results in investigating, forecasting and loss estimating of plant diseases and insect pests, and effect checking of prevention and cure.

The activities in our country started later, but some experiments and studies in the field ahs been finished using Landsat and aircraft data in recent years, and some experiences had been achieved. The paper will present some preliminary results of the investigation using meteorological satellite data.

Principle and Method

1. Main spectral features of forest plants
The radiations of different electromagnetic bands of different objects are theoretic bases of target sounding and distinguishing with Remote Sensing technology. When sunlight shine on leafs of green plants, part of it would be absorbed by leafs and the other would be reflected. The absorption, transmission and reflectance is controlled by physical and physiological mechanism relevant of plant spectrum. Spectral responses of plans are affected by factors as plant kind, growth stage, health condition, leaf structure roughness of leaf surface, chlorophyll content water content of cell and intensity of photosynthesis etc. Figure 1 shows the spectral curves of some kinds of tress changing with wavelength. The photosynthesis carry out with blue band an red band energy absorbed by chlorophyll. A part of green light through leafs are absorbed at landsurface and the greater part are reflected back to space, so plants show us in green color. The light of near - Ir band is reflected most intensively.


Fig.1
Differences of spectral responses between healthy and unhealthy trees.

The insect of pests can be classified in two sorts, as figure change and physiological change of trees, or the two changes at same time. In any change, the difference of spectral feature between healthy and unhealthy trees become greater with chlorophyll reducing or vanishing. Figure 2 shows the spectra changing curves of healthy and unhealthy conifers.


Fig. 2
Data Collection
We collected the data in region along Heilong river in Big Xingman Mountain (51°40' N~53°10'N, 125°E ~ 125°45'E), belonging to low mountain and hilly area with a mean elevation of 4,000 meters. This region includes whole forest land of the 18th Forestry Bureau and an part of Hanjiayuan and Kulunsi Forestry Bureau. The kinds of trees are natural confer mostly.

Pine moth harm was found in the region in early June of 1990, and developed greatly during 6~7 month. We collected AVHRR sounding data of NOAA-11 meteorological satellite on May 25 and July 19 of 1990, and obtained distribution maps of ratio vegetation index (RVI) calculated as the following.

RVI =

Anear --------- Avis

In the formula Anear and Avis is reflectance measured with near - IR channel CH2) and visible channel (CH1), RVI is a sensitive index to the change of chlorophyll content of trees, and show well forest coverage and growing condition.

Results and Discussion
Firstly we made corrections to the geographical positions in the RVI maps, and determined RVI critical values in the disaster region by field sampling checks, insect pest bearing when RVI<2.6 and heavy when RVI<2.3. Then we got slight and heavy disaster areas (see fig. 3). When meteorological satellite is applied to investigate insect pests of forest, the differences and changes of RVI in the map show a synthetical affection of natural geographical environments, as composition and distribution of underlying surface substances and all kinds of diseases. Therefore, firstly, we should understand the conditions of underlying surface, distinguish between two differences of objects and insect pests, and also distinguish differences of objects and insect pests, and also distinguish insect pests from other diseases. In our field investigating area (see fig. 3, dotted line) the vegetation coverage was lower in the last ten days of May, but the rivers and low-lamp objects were quire clear because the ice and snow was melting. We collected RVI data on May 25 as background data, delimited river sources, non forest land non-conifer land, those also showed lower values in the RVI map on July 19 and were removed in figure 3. In addition we investigated conditions of forest fire in the spring of 1990 in the region, an removed fire are as a interfering factor. After the above processions, we estimated the disaster area of pine moth in the regions with do accumulating method using digital data of meteorological satellite, the disaster are 416,000 ha, and the heavy 29,000 ha.


Fig. 3
Contrasting the estimating results to field investigating results (see fig 4) by the local Forest Protection Department, the medium degree or above disaster areas even the heavy areas were mirrored well, but the slight areas were not differentiated easily. The local Forest Protection Department took nearly tow months in investigating the conditions of insect pests, and defined disaster areas wit an index the distribute density (nit ; piece per stem) or pine moth: the slight at 1~19, the medium at 20~40, and the heavy at 40 or more. The biggest density is 2,000. We also found the other destructive insects like trussock moth. The analyzing results by a satellite Remote Sensing showed us the synthetical affection of many kinds of insect pests.


Fig. 4
It proved that the disaster areas and degrees can be estimated using meteorological satellite data through field sampling investigation and analysis. Except the slighter descase, the others can be differentiated upto to or exceeding that of artificial investigation.

Conclusions

1. We investigated the plant diseased and insect pests of forest in large scale using processed data meteorological satellite. The method is efficient, low cost, objective and with high precision.
2. The investigation is based on the change of chlorophyll content of injured tress. The differences and changes caused by natural geographical environment and non-insect diseases must be considered and distinguished sufficiently in the analysis.
3. The slight disaster areas are not easy to be differentiated by meteorological satellite, but the areas in changing of RVI can be done effectively.