| As populations grow, the impact of global environmental change by human activity is gradually increasing, and the irrational use of grassland results in global grassland ecosystem is destroyed. The grassland ecosystem in Ili area is fragile, grassland degradation directly reflects in the changes of grassland community structure, which directly restricts the development of grassland husbandry, thus threatening people’s lives who live in the frontier pastureland. Therefore, the study of spatial pattern and evolvement rule of grassland community in Ili area is conducive to further implement of the repair work of grassland degeneration, improve the ecological environment in the pastureland, and improve the living standards of herders.In view of this, this paper took the Ili area of Xinjiang as an example, and established a spatial model by using GIS technology based on measured sample point survey data(community index data and soil index data), remote sensing data(DEM and NDVI) and climate data(temperature, precipitation and humidity, etc.), which achieved the evaluation index of the space; related studies are carried out by this paper from such aspects as genetic driving of grassland community structure, spatial pattern change, trend change and degradation performance, etc., including correlation and regression analysis between grassland community index and ecological indexes, spatial pattern analysis of evaluation index, dynamic change trend and significance of evaluation index, quantitative discrimination of grassland types, and transfer of grassland type and its cause analysis, so as to provide technical support for restoration of grassland degradation in Ili area, and provide reference for research methods for studies on grassland communities in other regions.The main conclusions of studying the grassland community in Ili area are generalized as follows:(1) Grassland community evaluation indexes and ecological factors were significantly correlated, a mathematical regression model between grassland degradation evaluation indexes and elevation, annual temperature, annual precipitation, annual cumulative temperature above 10 ℃, wettability and NDVI were feasible. 1)There was an extremely significant correlation between grassland degradation indexes and grassland average total coverage, grassland soil bulk density, soil total nitrogen content, soil total phosphorus content and soil organic carbon content of grassland(P<0.01); 2)There was a significant correlation between average grass height and elevation, annual accumulative temperature above 10 ℃ and wettability(P<0.05), while there was an extremely significant correlation between grassland degradation indexes between annual average temperature, annual precipitation and NDVI(P<0.01); 3)There was a significant correlation between aboveground biomass and elevation, annual average temperature and annual accumulative temperature above 10 ℃(P<0.05), while there was an extremely significant correlation between grassland degradation indexes between annual precipitation, wettability, and NDVI(P<0.01); 4)There was a significant correlation between belowground biomass and elevation, annual average temperature, annual accumulative temperature above 10 ℃ and NDVI(P<0.05), while there was an extremely significant correlation between grassland degradation indexes between annual precipitation and wettability(P<0.01); 5)There was a significant correlation between soil total carbon content and annual precipitation, wettability,(P<0.05), while there was an extremely significant correlation between grassland degradation indexes between elevation, annual average temperature, annual accumulative temperature above 10 ℃ and NDVI(P<0.01).Grassland community factors of evaluation indexes and ecological linear fitting results indicated that 1) the best fitting model between grassland soil organic carbon content and wettability was ‘s’ type curve model; 2) the best fitting models between belowground biomass and annual accumulative temperature above 10 ℃, grassland soil bulk density and wettability were the reciprocal model; 3) the best fitting models between average grass height and mean annual temperature, soil bulk density and average annual precipitation, grassland soils carbon content and NDVI, grassland soil total nitrogen content and NDVI, soil organic carbon and NDVI were the exponential model; 4) the best fitting models between the other evaluation factors and ecological factors were binomial model.(2) The spatial inversion results of the evaluation indexes can reflect the actual distribution of each evaluation index in Ili area, and the accuracy of the inversion results is high. 1)The correlation between inversion results and the measured results in addition to belowground biomass of small R2=0.28, the other eight evaluation indexes are greater than 0.5; 2) the relative bias of the inversion results and the measured results are less than 10%; 3)the relative accuracy between the inversion results and the measured results are greater than 67%;(4) the mean estimation accuracy between the inversion results and the measured results are greater than 90%. So the comprehensive inversion results of each evaluation index reflected the spatial distribution better in Ili area, which made up the deficiency of the space research of grassland degradation evaluation index.(3) The changes of grassland community indexes from 2000~2013 were different, the inter-annual variation of soil bulk density was smallest, while the inter-annual variation of belowground biomass was largest; trend analysis of evaluation index of grassland degradation from 2000~2013 indicated that in addition to belowground biomass and soil bulk density showed a rising trend, others showed a decreasing trend, but the rise or decline trend was not significant; extremely significant correlation areas were mainly concentrated in the northern region of Qapqal Xibe Autonomous County, and Eastern region of Zhaosu County along Tekes River, but the area of those place accounted for a small proportion of Ili area; there were difference between the sensitive area of spatial fluctuation in each year from 2000~2013, and affected by changes in precipitation and the interference intensity of human activities(over grazing, different rotation modes), the sensitive areas were concentrated on the both sides along Ili River and the Kunes River, north of the southern slope of Ish rick mountain and south of northern slop of Jarque he Wu Shan in Zhaosu county.(4) Accorded to the principle of half peak, we determined the threshold range of each evaluation indexes quantitative identification of grassland types, thenbased on the fixed threshold range, the decision treesewere establishedand the grassland typeswere identificatedquantitatively. The distribution characteristics of different evaluation indexes ofgrassland types were analysed. Comparison of the results of quantitative analysis and remote sensing interpretation indecated that they had the same spatial distribution, and the numerical error between them were small, the accuracy of different kinds of grassland reached 75%, so the result was true and reliable; and compared with the traditional ground investigation and interpretation of remote sensing, wecould save a lot of manpower and financial resources.(5) Grassland degradation was serious in 1980s~2000 Ili area, the most direct manifestation was that the grassland area had decreased, while the non-grassland areahadincreased; between 2000 and 2010, grassland degradation had improved, grassland types transferred into grassland area increased, non-grassland area decreased, but local grassland had been degrading. |
PDF Full Text Request