Key words: phenology, vegetation index, climatic variations, remote sensing

Abstract
This paper is about the monitoring method of early stage signals of environmental and climatic variations. Because environmental (climatic) variation and vegetation variations have bi-directive indicate effects, hence the vegetational variations are reflected by the vegetation index NDVI in remote sensing image. Thus the vegetation index NDVI is an important indicator in measuring environmental changes caused by globe warm effect. Based on the monitoring result obtained through meteorological satellite last decade (1989-1998), we come to the conclusion. The vegetation index changes, especially greatly in spring, which is very obvious in north and northeast China.

Preface
Resources, environment and population are the three hottest in the current world. Deterioration of globe environment challenges the whole world greatly. In order to prevent its further deteriorating, scientists of multiply subjects all over the world are making researches on global environment changes, and trying to find out the causes of deterioration to prevent it. These studies include routine field observation and model estimation as well. Due to the limitation of observation stations and spots, and the random feature of estimation model factor, it Is hard to overview the whole aspect of environmental changes. Along with the progress of space technology, and the maturity of remote sensing technology, it is getting possible to carry out a consecutive, overall and dynamic monitoring of the whole world environmental changes. This monitoring covers land utilizing changes, dynamic changes of vegetation growth, and energy, water balance procedure, and so on. Since vegetation and environment (climate) can reflect and indicate each other, this article is trying to seek out early stage signals of environment changes at vegetational remote sensing method. Using current climate models (GCM model), we can know the mean temperature of China will increase by 1.5 to 4.5 centigrade, if double the content of Carbon Dioxide. Temperature increase will cause vegetational phonological change, this change can be reflected clearly in remote sensing images, while it is difficult to express directly in GCM model. Based on vegetation remote sensing image of China obtained through meteorological satellite last decade, the monitoring result of vegetational phenology is described in this article. In order to reflect the consecutive changes in our research work, we carried out a series of special processing method to deal with the meteorological satellite data.

1) Vegetation index processing method
The high calculation precision must be ensured in accomplishing remote sensing monitoring of climatic change, so we need to carry out a series of processing work. The definition of NDVI is a follows.

NDVI = (CH₂-CH₁)/(CH₂+CH₁)

Where, NDVI is vegetation index, CH₁, CH₂ is the reflectance of channel 1 and channel 2 of NOAA/AVHRR, respectively. The values of NDVI undulate greatly due to the influence of water vapor and aerosol of atmosphere. We use MVC method to get as mush as possible cloudless sky NDVI data to exclude these influences. We need to calculate the TNDVI of every ten-day out to get phonological changes, that is:

TNDVI = MAX[NDVI {t = 1,2,3,….10)

in the above equation, t is the number of days, and TNDVI is the maximum value of ten-day. Because the value of NDVI will be low when it is cloudy or in turbid atmosphere situation. We use this maximizing value method to remove the influence of cloud in most cases. But in summer or raining season, it is really difficult to get ten -day cloudless NDVI data, and the existence of cloud shields the information of vegetation, the NDVI value of this ten-day will be rather low. We have to remove the cloud influence to obtain the correct NDVI data of this ten-day by using three points insert value method. For example, if it is fully cloudy in the 2nd ten day of one month, we should insert a value between Ist and 3rd ten- day data of the month, and then maximize, that is:

TNDVI (N) = max (TNDVI (N+1) + TNDVI (N-1)/2, TNDVI(N)].

In the above equation. N is number of ten-day (36 ten -day per year). N can range from 1 to 36. For example, if the mean value of Ist and 3rd ten-day is greater than the value of 2nd ten-day, the mean value will replace the value of 2nd ten-day. After this Processing, in summer or raining season, NDIV data can be modified to some extent. We carry out a dynamic monitoring in a specified area in serial years by NDVI smoothed by every ten-day of every year. This monitoring can reveal the TNDVI change regularity of normal climate year and the NDVI exceptional phenomenon of abnormal year.