LANDSCAPE CHANGE BASED ON RS AND GIS IN THE JINGHE BASIN, CHINA

Based on digital Landsat Thematic Mapper (TM/ETM+) (1986, 1995, and 2000) imagery interpretation, landscape changes were analyzed in the Jinghe basin, one of the soil erosion regions in China, to provide basic data for local decision-making as well as sustainable landscape use and management. Results showed that Grassland area was always largest among all landscape types from 1986 to 2000. The second was cropland, the third was shrubland, and the fifth was forestland. So cropland and grassland were the main landscape types in the Jinghe basin and had an important effect on the direction of landscape change, ecological environment change, regional ecology safety, etc. Forest has an important function in keeping environment quality, preventing soil erosion, maintaining ecology balance, etc. in the region. But its area (including forestland, shrubland, scattered forest, and other forestland) was less than 11% of the total basin area. Areas of different landscape types altered in a certain extent. Only built-up land area increased from 1986 to 2000 always. The area of cropland, scattered forest, other forestland, and unused land increased between 1986 and 1995 and decreased between 1995 and 2000. And the increased area of cropland, other forestland, and unused land during 1986 to 1995 were more than the decreased area of which during 1995 to 2000. Among increased area types, the cropland area increased most, up to about 4165 ha from 1986 to 1995, and the changed area proportion of which about 0.89% during 1986 to 1995. The second was shrubland, up to about 2207 ha, and the changed area proportion of which about 0.47% between 1995 and 2000. Unused land from 1986 to 1995 and grassland from 1995 to 2000 increased about 849 ha and 816 ha respectively. The increased area amount of water was less than 94 ha from 1995 to 2000. Among decreased area types, the grassland area decreased most, up to about 3125 ha from 1986 to 1995, and the changed area proportion of which about 0.67% during 1986 to 1995. The second was shrubland, up to about 2781 ha, and the changed area proportion of which about 0.60% between 1986 and 1995. Cropland from 1995 to 2000 and scattered forest from 1995 to 2000 decreased about 2097 ha and 1086 ha respectively. The decreased area amount of other forestland was less than about 200 ha and the declined proportion in the total area was 0.04% between 1995 and 2000.

The probability matrix of area transformation illustrates the transformation direction and changes of different landscape types. Landscape change occurred among cropland, grassland, build-up land, unused land, shrubland, and scattered forest in the study area. Compared to during 1995 to 2000, alteration in the total landscape area between 1986 and 1995 was more obvious. The exchange between cropland and grassland was the greatest during two periods, and that between grassland, shrubland and unused land was the second. The reciprocal change between cropland and built-up land was the third. The change among cropland, grassland, shrubland, and unused land dominated change in the Jinghe basin, which influenced the direction and rate of the total landscape change. The GIS analysis further indicated that five land-use conversions were the most prominent: the conversion of landscape from cropland to grassland and built-up land and from grassland to cropland, shrubland, and scattered forest. The change between cropland and grassland during 1986 to 1995 were more than that during 1995 to 2000. The above results showed that the landscape change during 1986 to 1995 was more than during 1995 to 2000, and the changes the landscape heterogeneity and fragmentation increased and the landscape connectivity decreased between 1986 and 1995 were more than those between 1995 and 2000 at regional scale.

Change of land resources amount could be described by landscape dynamic index (e.g. land use dynamic index, landscape departure index, etc.), which could quantificationally explain rate of landscape change, and help to predict trend of landscape change. The absolute values of cropland and grassland were less than that of others landscape types, and causes of little dynamic degree of cropland and grassland were that the total areas of them were more than those of other types. The LUDI of other forestland was most during two periods because the area of which was small and the changed area of which was relatively bigger. Cropland changed uniformly during two periods. Changes of forestland, shrubland, sparse forestland, and unused land during 1986 to 1995 were more than those during 1995 to 2000, but the changes of other forestland, grassland, built-up land, and water were inverse to those of forestland, shrubland, sparse forestland, and unused land. The area of forestland, shrubland, grassland, and water decreased between 1986 and 1995 and increased between 1995 and 2000. But changes of cropland, sparse forestland, other forestland, and unused land were contrary to those of forestland, shrubland, grassland, and water. Bidirectional change between 1995 and 2000 was more obvious than between 1986 and 1995. But landscape use extent between 1986 and 1995 was bigger than between 1995 and 2000. Landscape developed continuously and transformed obviously before 1995, and which were in the regulation stage after 1995. Since 1998, China has developed the Natural Forest Protection and Returning Cropland to Forest and Grass programs with the intention of decreasing soil and water loss and improving ecological and environment quality. The policies had some effect on changes of cropland, forestland, shrubland, and grassland during 1995 to 2000. Landscape departure index refers to the extent of departure from original landscape for human activities, which equals to the value that acreage sum of built-up land, cropland, and other forestland is divided by the researched regional acreage. LDI increased from 1986(0.436) to 1995(0.448) and decreased from 1995 to 2000 (0.443), which showed that effect of human activities on landscape increased from 1986 to 2000, but effect extent of which decreased after 1995.

Remote sending data were widely used in different studies. Some big satellite data such as MSS, TM, ETM+, and Modis etc. usually were used by researchers in different research scale such as regional land use and land cover change, global environment change, etc. But big satellite data had some disadvantage, for example: price was expensive, satellite cost too high, run cycle too long, resolution low, etc. With the development of scientific technique, more and more little satellites were launched because of its more merit. Little satellite cost was few, run cycle short, price low, resolution high, and gain easy, etc. So we thought that small satellite could be applied in landscape ecology and land use etc for its merit. In the future landscape change research, little satellite data would be widely used by researchers because of its obvious characteristics.