Quantitative Study For Subclass Of Comprehensive Sequential Classification System(CSCS)

Comprehensive and Sequential Classification System of Grassland(CSCS), one of wellknown grassland classification systems proposed since1960s in China, represents the science ofGrassland Ecosystem Study for its perfect theoretical framework and quantitative potentiality. Inthe basic level quantitative classification study of CSCS, rich achievement has been made and anindex chart for determining grassland class has been established. But owing to the limitation oftechnique means, the quantitative subclass study does not develop for a long time.Based on analysis of the classification indexes of subclass, including geomorphology andsoil indexes, this dissertation put forward the quantitative principles and method of subclassgrassland and classified subclass in Gansu Province. The main achievements are as following:1. Using for reference from recent research results of geomorphologic classification and soilclassification, this study points out the relationship of geomorphologic classification, soilclassification and grassland subclass classification.2. Proposes the classification principles and method of grassland subclass:1) The clalssification and naming of geomorphology are performed according to "1:1,000,000Geomorphologic Maps of China"(1987),and modified with recent research results.2) The clalssification and naming of soil are performed according to the Soil CategoriseSystem in “China Soil”(by Academician Xi Chengfan,1998),and reference to Chinese SoilTaxonomy.3) Grasslands are grouped into subclasses by geomorphologic indexes in mountain areasand by soil indexes in plain areas. Plain areas mean flat ground, not just low attitude region. InGansu Province, on the scale of1:250000, where the slope is>3°and the area is greater than6km2,is divided into mountain, other is divided into plain. 4) In Gansu Province, take the basic geomorphologic types as the classifying unit, so dothe soil subtypes.3. Establishes the quantitative index system and index chart for grassland subclass.4. Groups the grasslands in Gansu Province into23first class according to CSCS, usingArtificial Netural Networks (ANN) and Kringing interpolation analyze the≥0℃a nnualaccumulated temperature and annual precipitation data.5. Divides20basic geomorphologic types in Gansu Province by extracting the slope, attitude,roughness, mountain and plain factors from SRTM-DEM data.6. Using supervised classification, divides36soil subtypes in Gansu Province by extractingthe slope, aspect, curvature factors from SRTM-DEM data and soil brightness, vegetation index,NDWI,NDSI and NDMI from MODIS data.7. Divides568subclass of grassland by operating the grid algebra with the grassland firstclass map, the basic geomorphologic types map and the soil subtypes map.8. Builds spatial database of grassland subclass.9. Other results:1) Improves the division of grassland first class by clustering station meteorogical datawith K-means, interpolating station meteorogical data with ANN and then interpolation analysiswith Kringing.2) Extracts soil information from MODIS data.3) Studies the divided limit between mountain and plain in Gansu Province based onSRTM data, puts forward the threshold value should be3°。4) Explores the macroscopic geomorphologic classification, determines the optimumsampling unit of roughness is16*16grids based on SRTM-DEM data.