| Urbanization is an inevitable trend of social development.With the accelerati ng urbanization of Shanghai,land use/cover has been undergoing dramatic chang es.The original natural landscape dominated by vegetation was gradually replace d by artificial structures, and one of the negative effects is urban thermal envir onment's changes.The changes cause the urban heat island phenomenon and urb an heat waves has become more frequent, resulting in morbidity and mortality increasing.Worsening urban heat environmental has become one of the factors h aving important effects on social and economic sustainable development.Shanghai selected as the researched area in this article, the land use/cover t ypes were extracted using support vector machine method based on the images of 1987 and 2007.And the land surface brightness temperature was retrieved f rom TM6.And then land use/cover changes were monitored on the basis of Ian d use/cover change amplitude,the rate of change and the spatial transformation. Thermal environment's changes were monitored on the basis of spatial pattern of thermal field, spatial and temporal evolution, ratio index of urban heat isla nd.At last the relationship between land use/cover and thermal pollution was st udyed and mitigation measures were proposed.Through this study, the main conclusions were as follows:(1)Woodland, grassland,city and countryside construction land showed a ri sing trend, while the cultivated land, waterbody and unused land showed a dec reasing trend in Shanghai city from 1987 to 2007. The dynamic degree of city and countryside land was the greatest, reaching 6.10%. Grassland and woodla nd were followed, respectively,5.78% and 5.22%. The dynamic degree of culti vated land and unused land was small,only -1.79% and -1.7%. The dynamic de gree of waterbody was medium-sized,which was-2.81%. (2)The land use/cover types transfered each other.And cultivated land was mainly replaced by woodland, grassland, and city and countryside construction land. City and countryside construction land continued expansion.It got rid of d evelopment in central urban area with a single core and gradually turned into t he new urban system pattern with multi-center, multi-level,network.(3)The central urban area of Puxi,Wusong industrial zone,Baoshan industrial zone and every populous town center were the high-temperature zones in 198 7.The high-temperature zones expanded rapidly,the orientation and direction of which was consistent with transportation corridors,and they were dispersed in 2 007.The concentrated form of sheets tended to launch and expanded in local s heet.The area of low temperature regions,sub-medium temperature regions and medium temperature regions decreased,and the area of sub-high temperature reg ions and high temperature regions increased from 1987 to 2007. The study abo ut thermal changes in different space directions showed the area in every grade temperature zone changed significantly. The URI except the urban areas all in creased.(4) City and countryside construction land had the highest temperature,follo wed by unused land,and water had the lowest temperature.The contribution of cultivated land, woodland and waterbody to urban heat island was small,but cit y and countryside construction land had a larger contribution.The reduction of woodland and waterbody and the increase of city and countryside construction land lead to a decrease in low temperature regions,sub-medium temperature reg ions and medium temperature regions and a increase in sub-high temperature re gions and high temperature regions.(5)At class-level pattern, it showed the correlation both the land surface brig htness temperature and IJI, LPI, PD, NP, PLADJ was large,but the correlation both the land surface brightness temperature and SHAPE-MEAN, PLANJ, DIVI SION, FARC-MEAN was small through Grey correlation analysis. At landscape-level pattern, the study showed that the negative correlation be tween PD,LSI,DIVISION,SHDI,SHEI and brightness temperature increased wit h the increasing scale. The positive correlation between LPI,CONTAG,AI and brightness temperature increased with the increasing scale. The correlation betw een SHAPE-MN,FRAC-MN and brightness temperature was small. The positive correlation between AREA-MN and brightness temperature increased on the w hole, the 8km X 8km scale smaller than 6km X 6km scall. the negative correlatio n between SHAPE-AM and brightness temperature was the biggest at 2km X 2k m scale,the followed at 6km×6km scale,the smallest at 8km X 8km scale. Consi dering two indicators named R2 and RMSE, the best linear regression model w as selected.At 2km×2km scale, the best linear regression model was y=-0.3423x +33.9401 (x standing for LSI),R2=0.5455, RMSE=0.6687. At 4km×4km scal e, the best linear regression model was y=-3.250x+34.539 (x standing for SHD I),R2=0.6682, RMSE=0.4851. At 6km×6km scale, the best linear regression m odel was y=-0.0436x+33.3898 (x standing for PD),R2=0.7130, RMSE=0.4809. At 8km×8km scale, the best linear regression model was y=-0.0428x+33.4011(x standing for PD),R2=0.7557, RMSE=0.3996.
|
PDF Full Text Request