GISdevelopment.net ---> AARS ---> ACRS 1990 ---> Global Change of Environment

Study on monitoring vegetation ecological environment over the whole China by using vegetation index from meteorological satellite

Xiao Qiangguang, Cheng Weiying1
Satellite meteorology Centre, State Meteorological
Administration, China


Abstract
The focus of this paper is put on the method (literate method) of processing AVHRR data of NOAA (or FY-1) meteorological satellite in monitoring vegetation ecological environment over the whole area of China. The mosaic image mad with this method eliminates cloud influence. Results of primary analysis of monitoring the ecological environment by utilizing the vegetation maps over the whole land of China is also presented which shows that the mosaic image not only can be used in monitoring vegetation ecology environment, but also can be used in monitoring climatic changes. Mosaic images made in different years, seasons and months apparently exhibit the yearly, seasonal and monthly changes of vegetation over the whole country and it also shows apparent regional division line and climatic lines which indicate the existing relationship between difference of regional vegetative distribution and climate. The mosaic image processing referred in the study has been operated and is becomes the found mental basis for further study on continental and globle changes.

Preface
The vegetation remote sensing in the past is mainly for the studies of the local vegetation status and their types. With the deterioration of the worldwide ecological environment, study in the continental and global changes has become the most noticeable problem. Research work has been conducted in more and more countries. The goal is to protect the global ecological environmental and maintain is balance. People hope earnestly to find an objective scale to measure this change and believe the growth and change of vegetation is an ideal scale. Because the change of vegetation has a close relationship with the changing of climate and environment. The vegetation can exist and reproduce under the certain condition of lightness, temperature and water and existence of vegetation can change the surface reflectance, and influence the local evaporativity and the exchange of energy and mass in ground surface. So vegetation and environment have bidirectional characteristics. Through a simple calculation of the reflectances of visible and near infrared channels of meteorological satellite, we can derive the vegetation index (greenness index), from which not only it exhibits vegetation, but also its amount can be estimated. The vegetation index maps in different years, season, months show the different status of vegetation distribution and amount. Relating it with the environment, we can infer what changes have happened and what changes will happen in the environment Biology), which focus on researches of the relationship between climate, environment and global vegetation is being on the upgrade.

-----------------------------------
1. lu Weijin and Li jing did part of the work; Mr. Cui Haiting participated in analysis of image.

Processing method of meteorological satellite data used in national vegetation ecological monitoring study
Many domestic and overseas remote sensing researchers described in detail about the advantages of NOAA meteorological satellite in macro-dynamic monitoring and acquired very good effect in practical applications, but the work was mainly done regionally, the data used are single time-phase. For continental and even global vegetation dynamic monitoring, the data needed are multi time phase and multi orbits as well, which is difficult in data processing, the challenge is assimilating the spectral data derived at different time, different place. Because the orbit of meteorological satellite drifts east ward 6 degree per day, the sub satellite points repeats every 9 days. To the same place, in different, at the sub satellite point it can achieve 11.km, and is about 4km when fat ways from sub-satellite point. To cover Chinese are, it at lest needs 3 orbits which take about 6 hours. In this time period, solar altitude angles and the intensities of incident sunlight on target change greatly, ad the irradiances reflected on satellite radiometer are also different. To the same place in different date, solar altitude angles are also different. Figure 1 shows this kind of change. It can be seen from the figure that the solar altitude angles change obviously within 9 days, and vary with the seasons' differences. Because cloud covers more than 70% of the areas, we need to mosaic, overlay the images so as to eliminate the cloud influence. The processing method is introduced briefly as follows:


  1. The creation of remote sensing equation
    The normalized output of each NOAA / AVHRR (or FY-1/AVHRR) channel can be calculated following equation:


    Considering the bi-directional reflectent characteristics equation (1) can be changed as:


    Among them,
    Nsr : Output of normalized instrument with the band-width of r
    j(l) : instrument's spectral response function
    l1, l2 : upper, lower limits of spectral channel
    E(l) : incident radiance at a pixel
    t(l) : Atmospheric transmittance
    Lpath : atmospheric path radiation
    q' : satellite zenith angle
    Z : solar zenith angle
    f : solar azimuth angle
    f' : satellite azimuth angle
  1. The definition of greenness in ecology environment dynamic monitoring.
    Vegetation has a relative high reflectance in near infrared wave hand, and the second channel of NOAA/AVHRR (or FY-1 / AVHRR) is just in near infrared band, it is the most ideal channel for vegetation remote sensing. But because of the influences of the solar altitude angle, satellite scanning angle, satellite scanning angle and atmospheric attenuation, the error is very big while only the reflectance of channel 2 "remote sense" the vegetation status. Even though there is little change in vegetation growing, the different between two days is still quite big. It is proved in theory and practice that the greenness index (vegetation index) derived from various combinations between channel 1 and 2 of NOAA/AVHRR (or FR-1/AVHRR) is much better that that derived from single channel. Among the 4 kinds pf greenness are mainly defines :

    1. Normalized greenness index NVI

      NVI = CH2 - CH1 / CH2 + CH1 .......................(3)


    2. Ratio Greenness index RVI

      RVI = CH1 / CH .......................(4)


    3. Difference greenness index DVI

      DVI = CH2-CH1 .......................(5)


    4. Vertical greenness index PVI

      PVI = cosq CH2 - sinq CH1 .......................(6)

    Among them CH1, CH2 respectively represent the reflectance of AVHRR channel 1 and 2.


  2. The vegetation Map processing
    All the greenness mentioned above can be used in making national vegetation map. In our studies we mainly choose the normalized greenness index, because it eliminates part of influence of solar zenith angle, satellite scanning angle a and atmospheric path radiation. On the other side, it is for matching with the greenness index which is popular in current world, so as to exchange data easily and development a worldwide monitoring together. The process of making vegetation map is shown in Fig 2. the first step is to preprocess the three orbits of everyday, calculate the reflectance and the greenness. Respectively them making mosaic image based on calculated greenness value the mosaic image made here is isometric latitude-longitude grid project. In order to match the image with map, the isometric orthographic oblique projection transformations needed. After the finishing to daily mosaic image processing, they need to be overlaid by day. Then have:

    MNVI = MAX (NVI (t)) t=1,2.....10 .......................(7)

    Among them MNVI is the maximum NVI within 10 days. Putting the boundaries of country and provinces based on it, making an image every 10 days, then we overlay three ten-day images of every month, we can derive image per month. Archiving images of every ten days and every month, therefore we can use them to monitor the ecological environment and climate change.

    The geometric construction of the geographic position is very important in above processing. If the everyday's error of the location is big, in the result of overlaying, the primary ground targets 9rivers, lakes) may disappear or magnify greatly. The pixel value does not represent the true station again. For solving this problem, we input the elements of satellite orbits derived from the testing and controlling agency every day, calculate precisely the location of every orbit so as to ensure the location error or preprocessing is less than one pixel.

    In vegetation mosaic processing, we can also derive the country wide true color image of three channels, that is after the preprocessing, put channels 1,2,4 together, then use red color to represent the lightness temperature of channel 4, use green for the reflectance of channel 2, use blue for the reflectance of channel 2, use blue for the reflectance of channel 1. This kind of mosaic is very useful in macro geomorphic studies.
Application analyses
After a series of processing listed above, a country wide vegetation index map (pseudo-color greenness map and quantitative digital data road printing map can be derived every ten days, the extent includes the whole China, the adjacent area of India-China peninsula and India peninsula. This kind of map can be used to realize the dynamic monitoring of vegetation ecological environment, and to study the climate change, macro agriculture management. The focus of this section is on analyzing the yearly, seasonal and monthly changing law of vegetation. It can be seen from the vegetation is quire obvious, the demarcation lines of regions and the separatrixes of climates are also very clear, because the difference of the regional distribution of vegetation represents the relationship between distribution of vegetation and climate. The satellite greenness map can be used to test the sensitive degree of vegetation to anomaly climate, because ever kind of plant combines in a special way under the most beneficial climate factors for its growing, the vegetation always prefers adjusting its own natural status to climate environment. The greenness changing on this kind of map reflects the change of climate factors.

Fig. 3 is the countrywide vegetation map in winter. During this season the most part of China is in the aphyllous period of defoliation, and farming land comes into over wintering period, the greenness value in northern area is very low. The area with maximum of greenness value on the map is the tropical in forests located in Laos, Burma an Thailand, the secondary locates in the southern slope of Simalaya (including cha yu), Ruilee, Cangyuan, Sishuangbana, Hekou in Yunnan Province. It also can be seen from the figure that the separetrix of winter vegetation is from Lingi-Guangyuan along Qing ling (Qingba mountains) Mountains - middle and lower reaches of Yangzi River, this separatrix is roughly identical to the northern boundary of the subtropical zone on China Block Plan, and the practical vegetation map is more objective than the block plan.

Fig. 4 shows the vegetation distribution condition in spring time, spring is the growing period of tress they become green, and the cultivated vegetation begin to revive, greenness value increase widespreadly, the green line begin to move from south to north forming a line which alogn DaXinganLing, through Hebei, Shanxi, northern part of Shaanxi, cross eastern part of Gan Shu, southeast of part of Tibet, this line forms the front of green line.

Fig.5. shows vegetation distribution status in summer. Summer is the most flourishing season of vegetation growing, the law is reflected from the figure that greenness decrease from southeast to northeast, the most luxuriant area locates in eastern part of China; in the Western part of the country, including west part of Inner Mongolia, Gansu, Qinghai and Tibet, the vegetation is sparse, only in Tianshan Mountain and Yily River Valley vegetation well distributed. It can be seen from the vegetation map that there are quite big differences f vegetation greenness among the wet subtropical zone, subtropical zone; semi humid region, semi humid region, dry area and desert area. On the map "The Green Boundary" to the southern. From the map the difference among the prairie, forest prairie and desert prairie can e told. In this season dynamic monitoring for the sensitive area is very important, it can determine the degeneration of the pasture and the process of desertilization.

Over Indian Ocean and Pacific Ocean, summer is the strongest monsoon season, it brings ample rainfall to the continent, the vegetation of all the places past by monsoon is quite well. On vegetation map, the monsoon area and the non-monsoon area are both obvious, the separatrix of monsoon area and non-monsoon area divided by climate region is along a line from Dasinanling Mountain - Yinshan Moutnain - Helanshan Mountain - Bayankalashan - Gangdisshan, roughly identical to the separatrix on vegetation map. Because the strength and time of monsoon are different by year by year, the greenness values reflected on vegetation map are also different.


Fig 6 shows the vegetation status in autumn, the trees come into the duration of fading. On vegetation map the brown line withdraw to Tai-bar mountains and the middle, lower reaches of the Yangzi River, the hilly area with high elevation decrease very fast, only the coniferous forest in Dasinanling Mountains show the spotted high greenness area.

Greenness map is a powerful tool for macroscopic monitoring of yearly change of vegetation. Especially in typical climate years, the change shows on greenness map is very evident. China and East Asia are the areas controlled by monsoon, the time and strength of every year's monsoon is different, so the conditions of light, water vapor and temperature along with it are also different, greenness clue map can reflect this kind of change timely and objective Fig. 7 and Fig. 8 are the summer vegetation maps of 1987 and 1989. Because there is a quite big difference between the climates of these two years, it can be seen form the map of 1987 that the high greenness area of Yily River in Xinjiang, joins together with Barlikun area, the greenness value in Altay Mountain is obviously higher than that of 1989. Because the influence of the rainfall and temperature of 198, the high greenness area of middle part of prairie in Inner Mongolia protrudes nouthward, meanwhile values of the area around the boundary of Hunan and Guizhou are also higher than 1989; but the greenness of the grassland vegetation around Qinghai Lake in 1989 is higher than that of 1987; the greenness values of 1989 in the boundary area of Sichuan, Gansu and Qinghai, the hilly area in Zhenjing, Fujian and Taiwan Island are higher than that of 1987. All of this relies on the climate condition.

References
  • D.F. MCGInnis, Jrand J.D. Tarpley: Advance space Research.Vol. No. 6 pp. 359-369, 1985
  • Compton J. Jucker, John R.G. Townshend ect. Science Vol. 227, No. 4685, 25, January 1985
  • Stantiey R. Schenider, Davicd F. McGInnis ect: INT.J. Remote Sensing Vol.6 No.1 1985.