【作者】 彭征；

【导师】 谢德体； 魏朝富； 廖和平；

【作者基本信息】 西南大学 ， 农业资源利用， 2009， 博士

【摘要】 近年来,全球城市化引起土地利用/覆盖变化的加剧,从而导致社会、经济与环境的进一步变化,进而对区域生态环境以累积的方式影响全球环境变化。而城市生态系统的脆弱性和不稳定性,导致一系列的城市问题不断显现。其中,城市的热力环境问题就一直受到广泛的关注。由于受遥感数据源精度的限制,目前对土地利用/覆盖变化的研究,大都局限于大、中尺度全球或区域层面,尤其是将城市土地类型作为一个整体对城市土地利用/覆盖变化进行研究。相应的,地表温度的研究受城市土地利用/覆盖类型的分类精度的限制,多是某一区域内一种或两种用地类型,研究其与地表温度的关系,而一个特定区域内各种用地类型对该区域地表温度的影响是一个有机的统一整体。因此,只有对城市土地的二级地类细化研究才能更好的反应城市内部结构和演变规律;只有研究各用地类型和地表温度相互影响的作用机理,才能进一步揭示进行两者关系研究的意义。基于以上论述,本文运用遥感解译和GIS空间分析技术,选取重庆地区经济发展最为迅速、人类活动最为频繁的中心城区小尺度典型区域为研究区,利用1986、1995、2000和2007年的TM/ETM影像数据以及统计数据,将城市土地利用类型细化为14个二级地类,分析其土地利用/覆盖变化,以城市土地利用类型的地表温度定量反演为基础,对城市内部地表温度时空分布特征及其与下垫面的关系进行研究,揭示城市土地利用/覆盖变化对城市热环境的影响,为城市合理规划及缓解城市热环境恶化提供科学参考。论文取得了以下主要的研究成果:1.重庆市中心城区土地利用/覆盖变化的研究应用ERDAS,ARCINFO等遥感、GIS软件,将研究区1987、1995、2000和2007年的4景卫星影像数据在与野外实测资料相互对比的情况下进行分类。据此建立起土地利用/覆盖变化的转移矩阵,并得出研究区近20年来的土地利用/覆盖变化总体趋势:研究区土地利用/土地覆盖变化,存在“一增二减”现象。“一增”是指城镇用地的面积显著增加,城镇用地占土地总面积比例从1986年的11.54%增加到2007年的18.02%,净增加面积达到了130.76km~2。“二减”是指旱地和水田面积的大幅度减少,旱地占地总面积从1986年的581.15km~2下降到2007年的490.33km~2,减少面积达90.82km~2。水田的面积则减少了44.64km~2,占土地总面积比例从1986年的20.10%连续递减到2007年的19.17%。重庆市中心城区的土地利用变化存在着明显的空间差异。城镇用地主要是在沿江两岸周围的较平坦地区,山区则主要发展林地、早地等。1995年以前,山区林地遭到严重的破坏,旱地和水田减少较缓慢。1995年以后,早地、水田急剧减少,而城镇用地急剧增加,从沿江两岸扩展到周边的丘陵山地区。整个研究时间段内,各土地类型相互转化的幅度后期较高,前期较低。归根结底,主要是由于直辖后重庆大面积的土地利用变化,导致各土地类型之间的转变更加频繁。2.中心城区城市土地利用动态分析为了对重庆市中心城区城镇用地进行更深入的研究,利用网格分析法、同心圆理论、GIS的栅格分析技术等理论与方法对土地利用变化的各种因素进行分析,得出以下结论:重庆市中心城区城市形态具有分形特征。其城市分形显示出典型的进化规律:近20年来,重庆市中心城区城市形态和其范围内城市各类用地的分维标准误差逐渐下降,从而意味着重庆市中心城区城市生长具有自组织优化趋势。重庆市中心城区建成区住宅用地扩展强度与模式在不同时期发生了很大变化。住宅用地扩展强度、速度在1997年前后发展趋势相反,1986～1995年,扩展强度持续上升,1995～2000年,扩展强度与速度升至最高值,2000～2007年,扩展强度与速度逐步下降,前后强度最大相差近十倍。1986～2007年,研究区内工业用地由18.0391km~2增至64.4700km~2,净增面积为49.4309km~2,年扩展速度为13.05%;其增加来源主要是水田和旱地,分别占23.99%和39.14%,二者之和为63.13%,超过半数以上。重庆中心城区内不仅商业用地比例过小,而且商业的行业分布不合理。根据我们对国际大城市的研究发现,随着城市化进程的加快,一般而言商业的最佳区位离市中心5km以内,而中心城区商业用地比例为15%左右。重庆市近20年的遥感影像所描绘的城市形态变化过程,主要是从1997年的“极核式”空间布局到2004年“分散集团式”和“集中型同心圆式”组合成的复合式空间格局。这两个扩展形式都体现了城市轨道交通的影响。作为城市交通主要方式的公共交通机动性良好时,容易引导城市沿交通线路轴向扩展(2007年);交通机动性较差之时,则是城市近域的圈层扩展(1986年)。3.中心城区地表温度时空变化分析本文采用单窗算法反演了1986年和2007年重庆市中心城区各土地覆盖类型的地表温度与城市建成区各城市用地类型的地表温度,发现反演出的城镇用地的地表温度值最高(1986年305.16K,2007年305.33k)。城市空间的扩展,使得土地利用覆盖类型发生变化的区域,尤其是转换为城镇用地的区域,在该时段内的地表温度明显的上升。城市空间的扩展引起下垫面性质的改变,尤其是城市的空间面积的增加,使得该区域地表温度相应的增加。研究时段内公园绿地的地表温度总是处于最小的状态,而工业用地的地表温度却总是所有城市用地类型地表温度分布最高的区域,而且各个统计的土地覆盖类型地表温度的标准差都不高,说明城市内部,单一用地类型的地表温度变化幅度较小。比较1986和2007年的数据,发现城市用地的地表温度从高到低都遵循以下的排列顺序:工业用地、住宅用地、商业用地、公园绿地和防护绿地。总的来说,工业用地地表温度最高,住宅和商业用地的地表温度差距不大,防护绿地和公园绿地的地表温度最低。4.重庆中心城区土地利用/覆盖变化对地表温度影响的定量分析深入分析植被覆盖与地表温度的内在关系,将下垫面类型分别与植被覆盖和地表温度进行叠加分析。比较各下垫面类型的地表温度和植被覆盖度均值、标准差,结果显示研究时段内,林地具有最高的植被覆盖度(58.70%～50.95%),城市用地的植被覆盖度仅高于水域(19.07%和17.77%),与之相对应的是,林地的平均地表温度最小,而城镇用地的地表温度最大。选取城镇用地的住宅用地、工业用地、商业用地、公园绿地和防护绿地分别制图,得到中心城区各城市用地植被覆盖度分布图。得出研究时段内公园绿地的植被覆盖度总是处于最高的状态,而工业用地的植被覆盖度却总是所有城市用地类型植被覆盖度分布最低的区域,而且各个统计的土地覆盖类型植被覆盖度的标准差都很高,说明城市内部,单一用地类型的植被覆盖度变化幅度较大。将植被覆盖度与地表温度做相关分析得出,植被覆盖指数与地表温度总体上呈负相关关系,植被覆盖度与地表温度的关系根据斜率的不同,明显的分成了两种:0%～35%和35%～75%,前一种的斜率大于0.2,后一种斜率在0.05-0.1之间。总结了土地覆盖变化对地表温度的影响。随着重庆市中心城区丘陵和山地地区的土地利用的覆盖变化,相应的发生植被覆盖度的变化,进而影响到地表温度。在土地覆盖类型发生变化的区域,大体上植被覆盖度下降10%,地表温度增加0.49K。更多还原

【Abstract】 Bearing a close relation to sustainable development, Land Use/Land Cover Change(LUCC), a key component and a primary cause of global change, has been attracting increasingly worldwide attention from all countries and international organizations. As an important type of land use coverage, urban land has become one of the focus contents in LUCC research. With the rapid development of society and economic, process of urbanization is accelerating. As the place influenced by human drastically and completely, no main character is kept by urban ecosystem except large climate character and topography features. This leads to fragility and instability of urban ecosystem, so series of problems appear during city expansion. Since "Urban Heat Island" phenomena were proposed by Lake Howard in 1833, more and more attention on thermal environment in city were paid.Chongqing is built against mountains at the junction of the Yangtze River and the Chialing. So Chongqing is a typical Mountainous City and water-city combining mountain with water everywhere. It displays its important position during the period of historic transition in the Western Development. As a Mountainous City, Chongqing has a little great difference about Land-Use from any other city, regardless it is from the aspect of social economy or ecological environment. Therefore, the researches of the Land-Use and Land-Cover have practical value. In this dissertation, the author selects Chongqing center district where is economy highly developed and human activity frequently as study area. Base on the TM/ETM images of 1986, 1995, 2000 and 2007, the author analyzed the land use and land cover change (LUCC), the spatial characteristic of land surface temperature and the relationship between land surface temperature and characteristic of land cover. This paper exposes the characteristic of land cover impacts on urban thermal environment, and provides scientific bases for reducing extend of the thermal environment deteriorating. The conclusions can be draw as follows:1 .Research of the land use/land cover change in Chongqing center districtThe land use types of the case study area can be classified into woodland, grassland, cultivated land, urban land, water body and rural residential area, according to the interpretation to the satellite images of 1986, 1995, 2000 and 2007 with reference to the historical maps and GPS positioning survey data on-the-spot. Based on the classification, transfer matrixes of LUCC from 1986 to 1995, 1995 to 2000 and from 2000 to 2007 were made and the features of LUCC were analyzed. The general trends of LUCC in the study area from the matrix show: The research on land use/land coverage change demonstrates the phenomenon of "one increase and two decreases", the former refers to urban land squares significantly increases by 18.02% in 2007 from 11.54% in 1986, expanding 130.76km in net increasing squares .The latter refers to squares of dry land and paddy fields significantly decrease which includes dry land’s decrease by 490.33km~2 in 2007 from 581.15km~2 in 1986 ,with the aggregate decrease of 90.82km~2.and paddy land’s decrease by 19.17% in 2007 from 20.10%,with the aggregate decrease of 44.64km~2.The land use change in Chongqing’s main districts existed deliberate differentiations. Urban land scatters in plain areas along the bank of river, and woodlands and dry lands mainly scatter in hilly areas. Before 1995, woodlands in hilly areas were seriously destroyed and dry lands and paddy lands slowly decreased. After 1995, dry lands and paddy lands dramatically decreased and urban expansion land fast increases which expands from banks of river to hilly areas nearby. During the whole study periods, the interactive change of varied lands is apparently clearer in late period than that in the beginning of the period. The change of varied lands is mainly due to Chongqing was directly administered by central government in 1997.2. Dynamic analysis of urban expansion land in center districtsIn order to study urban expansion land in center districts in Chongqing, the paper uses a series of theory which includes gridding analysis, concentric circle theory of GIS, to study varied ingredients about the change of land use. The paper concludes: the urban appearance of center districts in Chongqing is different according their geographical characteristics. Its evolution law is concluded as below: In recent 20 years , the standard differentiation of varied lands in center districts in Chongqing city gradually have decreased which shows center districts in Chongqing have a trend of self-organization optimization. Residential land expansion intensity in center districts in Chongqing changes in different periods. Its speed and intensity is much more different in the 1997 period. From 1986 to 1995, the expansion intensity gradually increases and the expansion speed and intensity goes up to maximum position, however, the expansion speed and intensity gradually decrease from 2000 to 2007, almost one tenth compared with the former. From 1986 to 2007, industry lands increase from 18.0391km~2 to 64.4700km~2 in studied area. Its net increasing areas are 49.4039km~2 and its annual expansion speed is 13.05%. Its source is mainly paddy land and dry land, with the proportion of 23.99% in paddy land and 39.14% in dry land. Land for commercial use in center districts in Chongqing city not only takes an insignificant proportion, but its distribution is not reasonable. According to our research towards international metropolis, with the development of urbanization, the optimal location of commerce is within 5 km away from the center of city, and the proportion is 15%.Transmission process of urban appearance described by remote sensing image in recent 20 years in Chongqing city, mainly demonstrates urban railway’s influence. If the municipal transportation is well, urban land expands along the transportation route axis (2007), and the condition is not favorable, it expands near the urban land.3. Space-time analysis of land surface’s temperature in center districtThe dissertation calculates the land surface temperature id varied land cover type in center districts as well as varied urban lands in constructed area in the year 1986 and 2007, and then concludes that urban land’s surface temperature is higher. (305.16K in 1986 and 305.33K in 2007).The expansion of urban space makes land use coverage types changes, especially in the changed urban land area, the temperature increases apparently. It also changes underlying surface, especially urban land’s increase which increases regional land surface temperature. In the studied period, the land surface temperature of greenbelt maintains in the minimum status and the land surface temperature of industry land is highest among all the investigated land types. The standard deviation is not so high in respective sample; it shows that the change extent of land for single use is steady in the inner part of the city. With the contrast of data in 1986 and 2007, we find land surface temperature of urban land conform an order: industrial use land, residential land, commercial use land, garden greenbelt and protection greenbelt. Generally speaking, industrial use land’s surface temperature is highest, and residential land’s surface temperature is near to commercial use land’s temperature, and garden greenbelt and protection greenbelt’s temperature are lowest.4. Quantities analysis on land use /coverage change’s influence on surface temperatureThe dissertation prudentially analyzes the inner relationship of plant coverage and surface temperature, with a respective overprinting analysis of underlying surface on land coverage and land surface temperature. After the comparison land surface temperature of varied underlying surface and plant coverage average result and standard deviation, the consequence shows in studied period, woodland has a highest plant coverage degree (58.70%-50.95%); urban land’s plant coverage degree is reluctantly higher than water area. (19.07% and 17.77%), correspondingly, average temperature of woodland’s surface is lowest and urban land’s surface temperature is highest. We select industrial use land, commercial use land residential land, garden greenbelt and protection greenbelt to paint maps which include varied land use type’s plant coverage degree of center districts. The paper thinks in the research duration, garden greenbelt’s plant coverage degree is always at the highest status and industrial use land’s plant coverage is at the lowest status. Each plant coverage degree’s standard deviation in calculated region is relatively higher and it formulates the plant coverage degree of land for single use in the city changes significantly. Through a correlative analysis of plant coverage degree and land surface temperature, index of plant coverage degree is negatively correlated with land surface temperature. According to slope of plant coverage degree and land surface temperature, it is divided into two parts: 0-35% and 35-75%, the former’s slope is more than 0.2 and the latter’s slope is within 0.05-0.1.The dissertation finally concludes the land coverage change’s influence on land surface temperature. With the change of land use on plain, hilly areas in Chongqing’s main districts, the plant coverage degree accordingly changes and then affects the land surface temperature. In the region which land coverage types change, totally speaking, the plant coverage degree decreases 10%, and the land surface temperature increases 0.49K.更多还原