GISdevelopment.net ---> AARS ---> ACRS 1999 ---> Poster Session 5

The Research on the Land cover and Vegetation Degradation Rate In the Karst Area using Remote Sensing

Yang Shengtian, Zhu qijiang
The Center for RS & GIS, Beijing Normal University 100875

Karst, 20 millions km2 in the world, occupies 12 percent of land in the globe. In china, based upon the distribution of Carbonate Rock, the karst area reaches 3.4 millions km2, according to the revealing area of Carbonate Rock, it is 0.9 million km2. Especially two regions of Karst Environment that that locate in Sanxi Province and nearby provinces, and in the west-south part of China including Guizhou, Yunnan and Hubei provinces are the widest karst area in the world. So the Karst environment plays an important role on the aspect of development of economy, society and so on in the world and China. For these reasons, many scientists have been focusing on the problems in the Karst environment. H.E. Legard brought forward the ecological problems on the karst environment in 1973; the karst environment formally was regarded as one of the weakest environment in the 149th advance conference in U.S. in 1983. Chinese geologists and geographers have also been concentrating on studying the change of the Karst environment since 1949. In particular, many institutes and university such as Gechmistry institute, Chinese Academy of Geology; NanJing University and Guizhou Normal University have got achievements on the fields of the relationship between human beings and nature, the changes of Karst environment quality, physiognomy generation in the karst environment, etc. However, an important question- the rate of degradation in the Karst environment-has stillnot been answered satisfactorily. Therefore, in the article, Remote Sensing was applied to study the changes of the Land cover and the Vegetation in the Karst area in order to solve in certain extent the degradation of the Karst environment which is useful to the Karst environment science and the development of society and economy in the Karst regions.

For studing the change rate of the Land cover and the Vegetation in the Karst area, we set a study area in the ma mountains, ZhiYun district, GuiZhou Province, China. The study area is 12km* rectangle; its latitude is about 25o34'41" and its longitude is about 106o18'39", its height is 1100 m altitude. The natural properties are that its rock is Carbonate; its physiognomy is ridges and peaks; it is in the area of middle subtropical class; its vegetation belongs to the class of regrowth forest and shrub in the middle subtropical climate. The average income for people in the study area is lower than 200 Yuan per year; the average food for people is lower than 300 Kg per year; so it is one of the poorest region in China. According to the situation of the land use, vegetation and habitat distribution, the study area was divided into three regions: the human being activity region, the nature- human being intergrade region and the natural domination region. The imagery of study area and its three regions are shown in figure 1.


Figure 1 The images of the study area and its regions

The research on the land cover change

The method of studying the land cover change

The imageries of Landsat MSS in 1973, TM in 1990 and TM in 1995 were chosen to classify the land cover and the result of classification was analyzed for purpose of recognizing the change of land cover in karst area. The steps were as fellows:
  1. The TM imagery in 1995 was geometrically registered to the relief map
  2. Taking register TM imagery in 1995 as basis imagery, matching TM imagery in 1990 to it. After resampling the MSS imagery in 1973 to make its resolution reach 30*30 m, matching it to registered TM imagery in 1995.


    Figure 2 the maps of the land cover classification in 1973, 1990 and 1995

  3. In order to minimize the different radiometric properties caused by season which could affect on the classification, the training sites were chosen at the same spatial position where there were the same land cover class in different year. The classes of training sites included: the cultivated land class; the grass and shrub class; the forest and shrub class. Every class of traing sites was divided to subclass according to shadow, different region.

  4. Respectively applying maximum lkehood classification to classify the Remote Sensing data in 1973, 1990 and 1995. After sieving and combining the classes, the map of land cover class was got.

  5. Checking the result of classification in 1995 with Surveying the land cover in the study area. The precision was estimated above 90%, so the classification method was reliable. The results were shown as figure 2.
The analysis of land cover change
The pixel statistics about every class on the basis of the map of land cover classification was shown as table 1. The data in the table 1 suggested:

Table 1 the pixel statistics about every class on the basis of the map of land cover classification
  The cultivated land cover(pixel) The forest & shrub cover (pixel) The grass & shrub cover (pixel)
1973 49877 25314 83633
1990 53852 17085 87860
1995 47901 16677 93600
  • The pixel number of forest and shrub class in the study area tended to become little. The decreased pixel number of the forest and shrub class from 1973 to 1990 was 8229, which decreased 32.5% comparing with the situation in 1973. Its degradation rate of was 1.9% per year. The decreased pixel number from 1990 to 1995 was 408, which decreased 2.4% comparing with the situation in 1990. Its degradation rate of was 0.48% per year. These results showed that the degradation of forest and shrub class hs been existing in the Karst environment recent years, but after 1990's, its degradation rate of starts to reduce because of executing measurements of natural conservation.

  • The pixel number of cultivated land class in the study area had increased and decreasing trend. The pixel number of cultivated land class from 1973 to 1990 increased 3975, which was 7.9 percent of pixel number of cultivated land cover in 1973. Its increasing rate was 0.46% per year. The cultivated land class pixel, however, from 1990 to 1995 decreased 5952 pixels, which was 11.1 percent of the number of cultivated land cover pixel in 1990. Its decreased rate during these 11.1 percent of the number of cultivated land cover pixel in 1990. Its decreased rate during these years was 2.2% per year. These results showed that the cultivated land area was increased during 1970's and 1980's because of human being reclaiming wasteland and the cultivated land area was decreased during 1990's for the reasons of executing the policy of natural conservation and exchanging cultivated land to forest.

  • The pixel number of shrub and grass class in the study area was increased. The pixel number of the shrub and grass class from 1973 to 1990 increased 4227, which was 5.1 percent of pixel number in 1973. The increasing rate was 0.3% per year. The increasing pixels from 1990 to 1995 was 5740, which was 0.3% per year. The increasing pixels from 1990 to 1995 was 5740, which was 6.5 percent of the pixel number in 1990. The increasing rate of the shrub and grass class was 1.3% per year. These results showed the shrub and grass class was in the increasing trend. Actually, the increasing trend of the shrub and grass class is the result of the Karst environment degradation and the environmental conservation. The forest and shrub class is degraded to the shrub and grass class and the cultivated land cover is also developed to the shrub and grass class for the reason of sending cultivated land back to forest. Therefore, it was recognized that the increasing trend from 1973 to 1990 was relate with the forest and grass cover degradation in some extent and the increasing trend from 1990 to 1995 was relate with the measurement of exchanging cultivated land back to forest in some extent.
The results shown as above corresponded with the results coming from analyzing the land cover change in the human being activities region, the natural-human being intergrade region and nutural domination region. The figure 3 suggested that 91) the cultivated land class in the natural-human being intergrade region and the natural domination region were increasing from 1973 to 1990, but the trend became negative from 1990 to 1995. The first process showed human being opened up wasteland and the latter showed human being started to execute the measurement of natural conservation. (2) Because the area of the forest and shrub class in the human being activities region was small during the time from 1973 to 1990, the change of the forest and shrub class was not great. The decreasing trend in the natural-human being intergrade region and natural region showed that the degradation in these regions was great. (3) The increasing rate of the grass and shrub in the human being activities region from 1973 to 1990 suggested the land degradation resulted in the shrub and grass class increasing. In addition to this process, the increasing rate of the grass and shrub class from 1990 to 1995 mainly suggested that grass and shrub class was the middle stp of land developing to forest and shrub. As the same reason, the two processes were included in the natural-human being intergrade region and natural domination region.


Figure 3 The pixel statistics of every class in three regions

The research on the vegetation change

The method of study the vegetation change

In the order to actually recognizing the vegetation change, the TM imagery data, which were at the same time of October end in 1994 and 1995, were chosen to calculate their NDVI. The steps were as fellows:
  1. Respectively standardizating the TM data in 1994 and 1995 to make them have the same mean and Stdev. The formulate was:


    Xij is actual data. Xij is the mean of j sample. j is the Stdev of j sample. After standardization, the mean of TM data in 1994 or 1995 is 1 and their Stdev is 0.

  2. Taking the histogram of TM3 band and TM4 band in 1995 as basis, respectively matching the histogram of TM3 and TM4 band in 1994 to them.

  3. Calculating the NDVI in 1994 and 1995

  4. The NDVI in 1995 minus the NDVI in 1994 was the changed message M. Through analyzing the changed message M, it was found that it mainly indicated the message of shadow. So the actual NDVI changed message AM which really showed the change of vegetation was obtained when filtering the message of shadow from changed message M.

  5. Checking the results with survey in the study area It was found the most of messages AM indicated the forest or shrub destroyed and some message AM indicated the changes of crops, so the message AM expressing the changes of corps must was eliminated with the manual method. The final result of changed NDVI was shown in figure 5.


    Figure 4 The imagery of the NDVI change region

The analysis of vegetation change
The pixel statistic of NDVI changes caused by vegetation degradation was shown in table 4.

Table 2. The pixel statistic of NDVI change
  The changed pixel The total pixel The changed pixel percentage
The human being activities region 184 17515 1.05 %
The nature-human being intergrade region 1060 60109   1.76 %
The nature domination region 170 9002 1.88 %
Total 1414 86626 1.63 %


The table 4 suggested that: There were 1.63 percent area in which the vegetation was lower than in last year. The high NDVI change in the nature-human being integrade region and nature domination region expressed that the better Karst environment is still the goal region of human being exploiting.

The conclusions
According to the analysis and results above, there are the conclusions as fellows:
  1. The method of the land cover classification and calculating NDVI using Remote Sensing data and analyzing their results are successful in obtaining the change message of land cover and vegetation degradation in Karst environement.
  2. Nowadays the degradation of the karst environment still exist because of human being activities, it is estimated that there is 1.63% area showing the vegetation degradation. Especially in the better quality of Karst area, the affection of human being exploitation is evidently revealed. It makes the degradation rate of vegetation cover reach 1.88%.
  3. The area of the forest and shrub cover is the main object of human being cultivating because of it environmental quality. Therefore many policies on environment conservation are recently executed and the degradationrate has been lower than 1.9% before, but it is estimated that there is still 0.48% the forest and shrub degradated per year. 

Acknowledges
The research was funded by the national science foundation of China (49761003) and the science foundation of GuiZhou Province, China (98-3061)

References
  1. Yuan daosian, Karst Science in China, the geological publishing house, 1993
  2. Yang shengtian, The study of sustainable development in Rock mountain, GuiZhou, Guizhou environment, 1994; 4:1~6
  3. yang mingde, The research on the relationship between the economic development and the Karst environment in GuiZhou Province, Guizhou publishing house, 1988
  4. Chao junming, the statistics of geography, Beijing Normal University publishing house, 1995