|
Urbanization is an inevitable trend of social development, with the accelerated process of urbanization, the Shanghai land use / cover is undergoing dramatic changes, the original natural landscape dominated by vegetation gradually been replaced by artificial structures, the most obvious change One of the negative effects of urban thermal environment changes, caused by the urban heat island phenomenon, urban heat waves become more frequent, leading to urban population increased morbidity and mortality, worsening urban thermal environment impact has become an important social and economic sustainable development factors. This paper selected as the study area in Shanghai, based on 1987 and 2007 two remote sensing images, the use of SVM method to extract the land use / cover types and use of anti-TM6 band performances brightness temperature information, and then on the basis of land use / cover change magnitude, rate of change, in terms of space into three land use / cover change monitoring, from the hot field of spatial pattern, spatial and temporal evolution of the district heat island ratio index changes in three aspects of the thermal environment monitor, then to land use / coverage and thermal environment to study the relationship between the last major heat island mitigation measures proposed. Through this study reached the following conclusions: (1) 1987-2007 Shanghai woodland, grassland and urban construction land showing an increasing trend, arable land, water and unused land showed a decreasing trend. Single land use / cover dynamic degree is the largest urban construction land, reaching 6.10%, followed by meadows and woodland, respectively, 5.78% and 5.22%, arable land and unused land was only -1.79% less dynamic degree and - 1.7%, medium-sized dynamic degree waters of -2.81%. (2) The land use / cover types into each other, mainly in land transfer out, the main steering woodland, grassland and urban construction land. Expanding urban construction land in the spatial shape of a single core has gradually get rid of the \pattern. (3) in 1987 as well as downtown Puxi Wusong Industrial Zone, Baoshan Industrial Zone of high temperature area, the suburbs populated urban centers also gather higher temperature zone. 2007 high temperature rapid outward expansion of the area, orientation, and direction of urban expansion and transportation corridors are consistent, and the distribution is more dispersed, so that the sheet tends to be more concentrated and localized emission flake-like extensions. 1987-2007 low-temperature region, sub-region in the temperature, the temperature of area reduction, sub-high temperature and high-temperature zone of the area increases. Different orientations of the thermal field in the spatial variation analysis showed that the temperature of each grade of each quadrant area change significantly. In addition to urban heat island reduction ratio index outside of the remaining districts are showing an upward trend. (4) the mean brightness temperature of urban construction land the highest, followed by unused land, water bodies mean brightness temperature minimum. Farmland, forest land and waters of the \area reduction, sub-high temperature zone, temperature increase in the main area. (5) In the Type Horizontal pattern through the gray correlation method derived characterize changes in landscape heterogeneity distribution and parallel index (IJI), the largest proportion of the area occupied by plaque (LPI), patch density (PD), number of patches (NP) and similar percentage of plaque adjacent (PLADJ) and brightness temperature greater relevance, while the average shape index (SHAPE-MEAN), similar to the proportion of type index near (PLANJ), patch diversity Index (DIVISION), average fractal dimension (FARC-MEAN) and brightness temperature less relevant at the landscape level studies show: PD, LSI, DIVISION, SHDI, SHEI increases with scale negatively correlated with the brightness temperature constantly increases; LPI, CONTAG, AI increases with scale positive correlation with the brightness temperature increasing; SHAPE-MN, FRAC-MN and the brightness temperature less relevant; AREA-MN increases with scale and brightness temperature is correlation on the whole increased in 8km × 8km scale than 6km × 6km scales slightly smaller; SHAPE-AM at 2km × 2km scale negatively correlated with the maximum brightness temperature, followed at 6km × 6km scale, In 8km × 8km negative correlation smallest scales. Consider these two indicators R2 and RMSE select different scales best linear regression model: 2km × 2km scale optimal linear regression model is y =-0.3423x 33.9401 (x is LSI), R2 = 0.5455, RMSE = 0.6687 ; the 4km × 4km scale optimal linear regression model is y =-3.250x 34.539 (x is SHDI), R2 = 0.6682, RMSE = 0.4851; scales at 6km × 6km best linear regression model is y =-0.0436x 33.3898 (x as PD), R2 = 0.7130, RMSE = 0.4809; scales at 8km × 8km best linear regression model is y =-0.0428x 33.4011 (x as PD), R2 = 0.7557, RMSE = 0.3996.
|