基于SD和CLUE-S模型的区域土地利用变化对土壤有机碳储量影响研究

土地利用/土地覆被变化改变土壤呼吸条件,进而对土壤有机碳储量变化产生影响,而土壤有机碳储量则是影响农业可持续发展和全球碳平衡领域的重要因素。以上海市崇明岛为例,运用系统动力学模型(System Dynamics Model)预测2020、2030年土地利用需求变化,结合CLUE-S模型(Conversion of Land Use and its Effects at Small region extent Model)得出各种用地类型的空间分布,并引用碳密度法估算三种发展幕景下土地利用变化对土壤有机碳储量的影响。结果表明:2030年三种发展幕景土壤有机碳储量分别为:低速发展幕景为3 093.03×106kg,惯性发展幕景为3 079.47×106kg,高速发展幕景为3 059.81×106kg;研究期内土壤有机碳储量呈现缓慢下降趋势,但人类活动对其扰动较小;SD和CLUE-S耦合模型可以从时间和空间两方面对土壤有机碳储量进行模拟,具有可行性;建议通过加强城镇用地集约利用、农田保护、林地建设来减少人为活动对土壤有机碳储量的影响。

EFFECT OF REGIONAL LAND-USE CHANGE ON SOIL ORGANIC CARBON STORAGE BASED ON SD AND CLUE-S MODEL

Carbon cycle or carbon balance is altered by human activities and carbon storage, especially soil organic carbon storage has become the focus of the world. Therefore, predicting the future carbon storage and selecting proper regional development paths have immense values. By taking Chongming Island as an example, this paper predicts its land use demand from 2011 to 2030 using the SD model, and then simulates landuse spatial distribution by the CLUE-S model. Finally, this study investigated the effect of land use changes on carbon storage under three development scenarios. Total population, GDP, urban construction land, agricultural land, forest and orchard land, urban green land, water and swamp area and other land are the level variables of the SD model. Remote sensing data of 2000, 2005 and 2010 years and statistical data of Chongming County Statistical Yearbook are the basic data of the paper. Testing stage of the SD model is from the year 2000 to 2010 and simulation stage is from 2011 to 2030.Based on the trend of historical development and some regional future plans, three scenarios (L, A and H) have been built up in the next decades. L means the development pattern with a low developing speed, A means the economic development pattern at a current developing speed, and H means the development pattern with a high developing speed. There are 12 driving factors in the CLUE-S model such as the distance of main towns, important towns, rural settlements, main roads, secondary roads, main rivers and wetland reserves, per capita GDP, population density, urbanization level, road density and river density. The results of logistic regression are used to check the availability of driving factors and land-use spatial distribution, which indicates that the value of ROC is appropriate. The results reflect that total soil carbon storage declines slowly and the decreasing rate is 1.40% and 1.32% from 2010 to 2020 and 2020 to 2030 during the study periods, respectively. The influence of human activities on the total soil carbon reserves is weak, but the influence of one single land use type is serious. For instance, in the H scenario, the declining rate of total soil carbon storage is 2.72% from 2010 to 2030,while the decreasing rate of carbon storage in agricultural land and forest and orchard land is 30.13% and 126.60%, respectively. L scenario is selected as the better development pattern with the largest total soil carbon storage. L, A and H scenario in 2030, the soil carbon storage of which are 3093.03×10⁶kg, 3079.47×10⁶kg and 3059.81×10⁶kg, respectively. It is reasonable and practical to simulate and calculate the ecosystem carbon storage by the coupled SD and CLUE-S models in both temporal and spatial scales. This paper then proposes some useful suggestions to reduce the impact of human activities on soil carbon process, such as strengthening farmland protection, promoting forest construction and improving land-use structure.