GISdevelopment.net --> Application --> Urban Planning



Abstract
In this paper, we use multi-temporal remote sensing data to analyze the pattern of surface change of the city of Shanghai and its neighborhoods brought on by the sudden development of the city and analyze how the surface change affects the change in surface temperature. To detect the surface change, the study used the satellite data of Landsat 3 MSS in 1979, Landsat 5 TM in 1989, Landsat ETM+ in 2001. To analyze the surface change of 22 years, the study used each of the satellite data to make the land surface maps, superimposed them together and analyzed the quantitative data for the amount of the surface change. We applied ISODATA unsupervised classification algorithm to classify them and made error matrix to verify the classification accuracy. The classification accuracy for the land cover maps was analyzed to be 93.01%, 90.44%, 88.23%, respectively. We analyzed the amount of the surface between 1979 and 2001 to find the change of 20.40%. That corresponds to the area of about 166,354ha. The rate of change in the urban area was 245.25%. Hence the size of the urban area increased more than two-folds during the period. To analyze the change in the surface temperature due to the change of the surface, we used Landsat thermal infrared data photographed in August, 1989 and July, 2001 to make the surface temperature distribution map. To analyze the impact of the surface change on the surface temperature, we superimposed the land cover map and the surface temperature distribution map to find out the change of the surface temperature according to land cover type. The average surface area in 1989 and 2001 were 23.55. and 28.78. hence increased by about 5. in 12 years. The difference in the temperature of the urban area and the agricultural area was found to be 2.01. in August, 1989 and 2.38. in July, 2001, hence the difference increased by 0.37. This result can be seen to verify the increase in the surface temperature due to urbanization. The increase in temperature due to global warming can be one factor to this, but the change in the material such as concrete, asphalt, and steel that cover roads, buildings and industrial complex built due to the sudden increase of the urban area seem to have a bigger impacts.

1. Introduction
As the biophysical status of the surface of the earth, the land cover is the basis of the most material and energy transfer and of the transfer of the materials and energy and the interaction between the geosphere and the biosphere. The land surface change involved the variety of organism, soil type, running water and the change in the rate of deposition and the cause of it cannon be understood without the knowledge about the change in the usage of the land. Hence, it can be said that the usage of the land and the surface change has a limited and localized environmental implication and also related to the global environment change process. Because of the interconnected nature of the natural environment components, a direct impact on one component can have an indirect impact on another component[1].

The urbanization, which is the change in the surface type due to the popular increase and the economic development, is historically an important pattern in the change of the land cover and has the biggest influence on the climate. The urban area, which is covered by buildings, roads, and impervious materials, absorbs high amount of solar radiation and has a higher heat value and conductivity. As a result, the heat is stored during the day and released during the night. The difference of temperature between the urban area and the island area is not very big - on average, about 1·. But when the urban, geographical, meteorological conditions meet that of generation of urban heat island effect, the difference can increase to several centigrades.

Remote sensing and the spatial analysis technology have been recognized and used as powerful and effective tools to monitor urban land usage and surface changes. Satellite remote sensing collects multi-spectrum, multiresolution, multi-period data and provides valuable information in understanding and monitoring the process of urban land cover change, and in constructing urban land cover databases.

The purpose of the study is to present a method of using remote sensing and spatial analysis technology in environmental topics and for this purpose, we assessed the pattern of the sudden increase in the urban area of Shanghai, China and analyzed its impact on the surface temperature.

2. Materials & Methods

Figure 1. The region covered in this study: Shanghai, China and its neighborhoods.

Our targets of study, Shanghai of China and its surburbs are located on the center of the area between of the western part of the Pacific Ocean and the Chinese coastal line running from south to north. As shown in

, its latitude is between 30°59'N and 31°24'N and its longitude is between 121°17'E and 121°44'E. The city of Shanghai is situated on the banks of the Yangtze River Delta, where Huangpu River meets Yangtze River and contains the wide area around the city center and the island of Chungming near the mouth of the Yangtze River. Them main river in Shanghai is the Suzhou river. The climate is warm and the annual average temperature is 15°~16°, and the average rain fall is 1,100 ~1,200 ·. The population of Shanghai in 2001 is 16 million and 740 thousand. The outskirts of Shanghai belongs to northern subtropic region and has fertile soil with agricultural area of 333,000ha. An area near Shanghai is on a delta with fertile soil and has many small and large tributaries running north-south and east-west, beside Suzhou river[2,3].

The sudden economic growth in Shanghai brought about changes in land usages and the land surface patterns. This study adopts the method of analyzing the pattern of surface changes and impact on surface temperature due to that To detect the surface change, the study used the satellite data of Landsat 3 MSS in 1979, Landsat 5 TM in 1989, Landsat ETM+ in 2001[4]. Using the satellite data, the region was divided into urban areas (including asphalt and concrete road), agricultural regions, hydro regions (rivers, lakes and ocean), and unclassified regions. ISODATA unsupervised classification was used to classify the land cover based on the satellite images. For quantitative analysis of the land cover classification results, we made the error matrix and for quantitative analysis of surface change we made land cover change matrix.

To analyze the change in the surface temperature due to the change of the surface, we used Landsat infra-red heat data photographed in August 11, 1989 and July 3, 2001. 1989 data were acquired by Landsat 5 TM and 2001 data were acquired by Landsat ETM+. In this study, ATCOR2 atmosphere calibration algorithm was used to calculate surface temperatures from satellite data from Landsat 5 TM. In ATCOR2 atmosphere calibration algorithm[5,6], we assume that surface radiation rate e for infra-red spectrum (10.5 - 12.5 ·) in Landsat 5 TM is 0.98.

3. Results and Discussion
Land surface map was constructed based on the classification of land surface of Shanghai, China and its neighborhood using Landsat satellite image. The land surface was classified into urban, agricultural, hydro and unclassified regions. The overall classification accuracy for the years of 1979, 1989, 2001 were 93.01%, 90.44% and 88.23%, respectively. The Kappa coefficients, which measure the reliability of the result, were analyzed to be 0.84, 0.80 and 0.78, respectively

The land surface map with this level of classification accuracy can be used to detect the change of land surface. To analyze the land surface change from 1979 to 2001, land cover change matrix was constructed using the land surface map constructed above. We analyzed the land surface change using the superimposed land surface maps of 1979 and 1989 to find out that there was change of 3.34% in overall area. That corresponds to 27,196ha. The urban area had the greatest increase of 40.06% from 1979 to 1989. On the other hand, the agricultural region decreased by 0.41%. The land surface change from 1989 to 2001 was analyzed to be decreased by 17.07% in overall area. Also in this case, the urban area had the greatest increase of 146.50%. On the other hand the agricultural region decreased by 11.65%. The land surface change from 1979 to 2001 was analyzed to be 20.40% in overall area. That corresponds to 166,354ha. The rate of change in urban regions was 245.25%. Hence the urban region area increased more than two folds during the period. On the other hand the agricultural region decreased by 12.01% in 2001 compared to 1979.

To analyze the change in surface temperatures due to the change of the surface in Shanghai, we used Landsat infra-red heat data photographed in August 11, 1989 and July 3, 2001. The infrared image used in this study came from Landsat 5 TM and ETM+ and each has different sensors and used different temperature transformation equations. We analyzed the relation between DN values and temperatures of Landsat 5 TM and ETM+ infrared sensors to calibrate the temperature value discrepancies due to the difference of the satellite sensors used. The equation between DN values and temperatures of Landsat 5 TM data obtained in August, 1989 and Landsat ETM+ data obtained in July 2001 are as follows.

Temperature(· )· TM = 0.4407 × DNTM - 35.621
Temperature(· )· ETM = 0.4817 × DNETM - 41.121

We obtained the following equation after normalizing two equations.

Temperature(·)·TM_ETM = 0.4687 × DNTM_ETM - 39.401

We constructed the land surface temperature distribution map of August, 1989 and July, 2001 using the temperature transformation equation obtained above. To analyze the impact of the surface change on the surface temperature, we superimposed the land cover map and the surface temperature distribution map to find out the change of the surface temperature according to land cover type. The land surface temperature of urban, agricultural, and hydro region in 1989 were 29.38· ·, 23.37 ·· and 21.92··, respectively. They were 31.80··, 29.42·· and 25.14··, respectively in 2001. The average surface area in 1989 and 2001 were 23.55·· and 28.78·· hence increased by about 5··. The difference in the temperature of the urban area and the agricultural area was found to be 2.01·· in August, 1989 and 2.38·· in July, 2001, hence the difference increased by 0.37··. This result can be seen to verify the increase in the surface temperature due to urbanization. The monthly average temperatures data of Shanghai in July and August in 2003 were 27.8·· and 27.7··, respectively and didn't have a big difference. But the surface temperatures calculated from the satellite data photographed in August, 1989 and July, 2001 had a big difference. The increase in temperature due to global warming can be one factor to this, but the change in the material such as concrete, asphalt, and steel that cover roads, buildings and industrial complex built due to the sudden increase of the urba n area seem to have a bigger impacts


Figure 2. Land cover map extracted from Landsat satellite data using ISODATA classification method

Table 1. Error matrix of land cover map, 1979


Table 2. Error matrix of land cover map, 1989


Table 3. Error matrix of land cover map, 2001


Table 4. Land cover change matrix, 1979-1989 (pixel)


Table 5. Land cover change matrix, 1989-2001 (pixel)


Table 6. Land cover change matrix 1979-2001 (pixel)


Table 7. Average surface temperature(·)_· by land cover type



Figure 3. Land surface temperature distribution in Shanghai, China and its neighborhoods based on Landsat 5 TM and Landsat ETM+ infrared image.

4. Conclusion
This study uses Multitemporal remote sensing data to analyze the pattern of surface change of the City of Shanghai and its suburbs brought on by the sudden development of the city and analyze how the surface change affects the change in surface temperature. To detect the surface change, the study used the satellite data of Landsat 3 MSS in 1979, Landsat 5 TM in 1989, Landsat ETM+ in 2001. As the result of the analysis, we found out that the land surface change between 1979 and 1989 was 3.34%, about 27,196ha and the size of the urban area increased by 40.06%. The land surface change between 1989 and 2001 was 17.07%, about 139,176ha and the size of the urban area increased by 146.50%. To analyze the change in the surface temperature due to the change of the surface, we used Landsat thermal infrared data photographed in August, 1989 and July, 2001 to draw the surface temperature distribution map. We found out that the average surface temperature in 1989 and 2001 were 23.55·· and 28.78·· hence increased by about 5·· in 12 years. The difference in the temperature of the urban area and the agricultural area was found to be 2.01·· in August, 1989 and 2.38·· in July, 2001, hence the difference increased by 0.37··. This result can be seen to verify the increase in the surface temperature due to urbanization. The monthly average temperatures data of Shanghai in July and August in 2003 were 27.8·· and 27.7··, respectively and didn't have a big difference. But the surface temperatures calculated from the satellite data photographed in August, 1989 and July, 2001 had a big difference.

The increase in temperature due to global warming can be one factor to this, but the change in the material such as concrete, asphalt, and steel that cover roads, buildings and industrial complex built due to the sudden increase of the urban area seem to have a bigger impacts.

5. References

1. B. L. Markham, J. C. Seiferth, J. Smid, and J. L. Barker, "Lifetime responsivity
2. Department of Geography, Zhongshan University, 1988, The Land and Water Resources in the Zhujiang Delta (Guangzhou: Zhongshan University Press).
3. Ditu Chubanshe, 1977, Provincial Atlas of the People's Republic of China (Beijing: People's Press).
4. P. M. Teillet, J. L. Barker, B. L. Markham, R. R. Irish, G. Fedosejevs, and J. C. Storey, "Radiometric crosscalibration of the Landsat-7 ETM+ and Landsat-5 TM sensors based on tandem data sets," Remote Sens. Environ., vol. 78, no. 1-2, pp. 39-54, 2001.
5. Roth,M., Oke, T. R., and Emery, W. J., 1989, Satellite derived urban heat islands from three coastal cities and the utilisation of such data in urban climatology. International Journal of Remote Sensing, 10, 1699- 1720.
6. Gallo, K. P., McNab, A. L., Karl, T. R., Brown, J. F., Hood, J. J., and Tarpley, J. D.,1993b, The use of a vegetation index for assessment of the urban heat island eVect. International Journal of Remote Sensing, 14, 2223-2230