【作者】 刘廷祥；

【导师】 张树文；

【作者基本信息】 中国科学院研究生院(东北地理与农业生态研究所) ， 地图学与地理信息系统， 2012， 博士

【摘要】 东北地区是全球范围内短时间经历人类活动高强度作用的区域之一，特别是耕地和森林分布格局发生了巨大的变化，形成了具有典型变化过程和格局特征的农林交错带。该变化显著地改变了地表覆被状态，并对区域气候系统要素及地表－大气之间的相互作用产生了重要的影响，进而改变了区域降水、温度等的分布格局。对这一格局和过程变化的研究不仅在区域土地利用/覆被变化（LUCC）和气候变化领域具有代表性，并为更大尺度的全球变化研究提供典型案例，而且服务于东北区域发展。本文在总结和评述国内外LUCC研究及其气候效应研究的基础上，以东北地区农林交错带为研究对象，获取了包括1950年代、1970年代和2010年的东北土地利用时空数据序列，并选择嫩江流域农林交错带为典型研究区，从土地利用变化分析和WRF（Weather Research Forecast）模式数值模拟两个方面系统地研究了农林交错带土地利用格局变化及其区域气温变化效应。本文的主要内容和结论包括：（一）东北地区农林交错带的空间分布在过去几十年里发生了显著的变化。本文首先以东北地区为基本研究区，利用基于遥感影像和地形图获取的土地利用数据在东北地区划分出不同的生态交错带，进而主要分析了东北地区农林交错带的时空变化与特征。在东北地区，农林交错带主要分为林缘农林交错带和林间农林交错带，两者的差异取决于区域地貌和气候特征，并形成了不同的土地利用格局；农林交错带中主要的生态问题是水土流失问题，农林交错带的自然环境决定了农业开发中土地系统的脆弱性。（二）农林交错带的景观特征与其他类型的生态交错带之间具有显著的差异，不同地区的农林交错带区域景观剖面表现出不同的土地利用格局。在东北地区不同位置的典型农林交错带中提取了五个景观剖面，分析了不同地区农林交错带横剖面上的土地利用分布格局。结果显示不同地区农林交错带的土地利用格局和过程具有明显的差异。引入景观生态指数计算方法和两个主要的环境因素，对比了大兴安岭东部农林交错带与其他生态交错带之间景观异质性的差异及其时空变化。分析结果表明：（1）坡度2.5～7.5°之间是耕地和林地的主要转换区；（2）从1978～2010年，边界密度（ED）和核面积指数（CPL）的变化表明坡度在2.5～10°的地区林地景观的破碎化程度最高。（3）ED曲线峰值两侧在1978～2010年间的变化表明，森林砍伐具有向着更大坡度地区移动的趋势。（4）区域上不同的生态交错带在干燥度指数上有明显差异，农林交错带中的耕地在区域上所有耕地中的湿润程度最高。（三）在GIS技术支持下，本文对WRF模式输出的不同分辨率尺度上的气温值进行了空间尺度分析。利用气象站观测值与不同分辨率尺度输出结果之间对比，分析表明随着分辨率的提高，站点观测值与WRF模式输出值之间的偏差越来越小。随着分辨率的增加，越来越多的细节被刻画出来，但3km和1km之间信息量的增加有限。在尺度下推中，本文以影响地方气温最显著的因子－高程为例，用高分辨率的DEM将9km分辨率的气温分布下推到3km分辨率；与WRF模式在3km分辨率上的输出结果对比，表明尺度下推结果能够反映出3km尺度上的气温变化趋势，但在空间分布细节上的变化较粗糙。而基于栅格重采样的尺度上推则相对具有更高的精确度。以1km分辨率的气温分布为因变量，以经度、纬度、海拔高度和地形因素为自变量，分析了气温与这些因素之间的关系。其中地形因素为坡度和坡向的综合影响。回归产生的残差分布反映了植被、土壤、小尺度大气过程对气温分布格局的贡献。（四）对嫩江流域过去60年7月平均气温的时间序列分析表明，其整体上具有上升趋势和局部的波动变化，并存在5～12年和21年的周期变化。不同土地覆被条件下基于WRF模式的温度变化敏感性分析表明，嫩江流域农林交错带1950s～1970s的土地覆被变化具有增温效应，而1970s～2010年的土地覆被变化具有降温效应。其中，与气温相比，地表温度对土地覆被类型转变更加敏感，变化的程度明显更显著。同时，土地覆被变化带来的增温或降温效应在区域整体上比较显著，而在不同土地覆被类型转变区上的差异并不十分显著。对土地覆被变化和气温变化的综合模拟分析表明，1978年比1951年气温稍有上升，但上升幅度比土地覆被变化导致的增温效应小，因此气候系统可能具有降温的背景；而2010年比1978年气温显著上升，结合该时期土地覆被变化的降温效应，说明气候系统整体增温的背景趋势更强烈。更多还原

【Abstract】 The Northeast China is one of typical regions experiencing intensive humanactivities within short time worldwide. Particularly as the significant changes offarmland and forest, there were typical characteristics of pattern and process ofagroforestry ecotone change. The intensive land use change of agroforestry ecotonehas made significant change for regional land cover, which had significant impact onthe regional climate system elements and the interactions among them, and modifiedthe distribution pattern of regional precipitation, temperature, etc. Research about thepatterns and processes of these changes is representative at the regional LUCC andclimate change researches, which would enrich typical case studies for global researchon a larger scale, and service for regional development in Northeast China.Based on the summary of domestic and international researches concerning onland use change and its climate effects, this paper takes agroforestry ecotone inNortheast China as study object, and establishes temporal and spatial land use changedata of1950s,1970s and2010year. Then multi-researches are carried out includingland use change analysis and climate simulation with WRF model to study the landuse change pattern and its climate effect of agroforestry ecotone. Several conclusionsare obtained:Firstly, the spatial distribution of agroforestry ecotone in Northeast China hassignificant changes in the past few decades. The paper takes Northeast China as basicstudy area, and identifies several ecotones in Northeast China based on land use datafrom remote sensing images and topographic maps, and analyzes the temporal andspatial change and characteristics of agroforestry ecotone. In Northeast China, theagroforestry ecotone could be classified into two types, agroforestry ecotone at forestedge and agroforestry ecotone inner forest, the differences between them rest withregional geomorphological and climatic characteristics, forming into different landuse pattern. The primary ecological problem in agroforestry ecotone is soil erosion.The natural environment determines the vulnerability of land system in agriculturedevelopment.The landscape characteristics of agroforestry ecotone are significantly differentfrom other ecotones, and landscape profiles at different location of agroforestryecotone in Northeast China showes different land use patterns. Five landscape profilesare extracted in different locations of typical agroforestry ecotone in Northeast China.The land use pattern analysis of different profiles showes that different parts of agroforestry ecotone have different land use pattern and process. Then landscapeindex and two environment factors are introduced to compare the landscapeheterogeneity differences and changes of agroforestry ecotone and other ecotones insoutheastern Greater Khingan Mountains. The results show that:(1) the slope ofprimary transition zone between farmland and forest is2.5-7.5°,(2) from1978to2010, the changes of the edge density (ED) and core area index (CPL) show that theregion with slope between2.5-10°has the highest degree of fragmentation for forestlandscape,(3) according to the change of ED curves besides the peak value between1978-2010, the deforestation has a trend to move toward areas with greater slope,(4)The aridity index of different ecotones and different location is significantly different,and the agriculture land in agroforestry ecotone is the most humid area all over theregional agriculture land.With the support of GIS technology, the temperature values from WRF modelwith different resolutions are analyzed from spatial scale view. The comparisonanalysis between the meteorological stations observed values with WRF model outputwith different resolutions indicats that, with the improvement of the spatial resolution,the deviation between the site observations and WRF model output value is gettingsmaller. As the resolution increases, more detail is portrayed, while the informationincrease from3km to1km is limited. Elevation is the most important factorinfluencing local temperature distribution. This paper uses high-resolution DEM todownscale the temperature distribution with9km spatial resolution to3km spatialresolution. Compare with the WRF output, downscaled data could reflect the basictrend of temperature changes in3km resolution, while detail changes are very rough.Besides, this paper takes temperature with1km spatial resolution as dependentvariable, and the longitude, latitude, altitude and topographic factors as independentvariables, where the topographic factor is the combined effects of slope and aspect, toanalyze of the relationship between temperature and these factors. Distribution ofregression residuals reflects the contribution of the vegetation, soil, small-scaleatmospheric processes to temperature distribution pattern.The temperature time series analysis of Nenjiang River basin shows that, therewere overall rising trend, local fluctuations and cycles of5-12years and21years foraverage temperature in July of last60years. The temperature sensitivity to land coverchange analysis based on WRF model shows that, the land cover change between1950s and1970s results a warming effect, and land cover change between1970s and2010year results a cooling effect in agroforestry ecotone of Nenjiang River basin.Comparing with the air temperature, land surface temperature changes are much moresignificant and sensitive to land cover changes. At the same time, the warming orcooling effect of land cover change is more significant in the region scale rather thanspecific land cover area. The comprehensive simulation analysis combining climatechange and land cover change shows that, the temperature increased between1951 and1978, but the increase is lower than the warming effect of land cover change. Thetemperature increased significantly between1978and2010, but it is the result afterthe offset by the cooling effect of land cover change, and the regional warmingtendency was stronger.更多还原