Available online at www.sciencedirect.com Procedia Environmental Sciences 10 (2011 ) 1466 1472 2011 3rd International Conference on Environmental Science and Information Conference Application Title Technology (ESIAT 2011) Percentage of Vegetation Cover Change Monitoring in Wuhan Region Based on Remote Sensing Tao Chen 1, a, Rui-qing Niu 1, Yi Wang 1, Liang-pei Zhang 2 and Bo Du 3 1 Institute of Geophysics and Geomatics, China University of Geosciences, Wuhan, China 2 State Key Lab of Information Engineering in Surveying, Mapping & Remote Sensing, Wuhan University, Wuhan, China 3 Computer School, Wuhan University, Wuhan, China a chenhhabc@yahoo.cn Abstract In this paper, percentage of vegetation cover in different districts of Wuhan were calculated based on remote sensing images from 1988 to 2002 by employing NDVI method for dimidiate pixel model. Then the vegetation coverage maps in different periods were generated to analyze the temporal change of vegetation coverage of Wuhan. The results showed that, the average vegetation coverage of the whole area was decreased from 8.41% to 0.4%, especially in Jiangxia and urban district. From 1996 to 2002, the vegetation coverage decreased most sharply, while the vegetation coverage increased lightly from 1991 to 1996 in the whole study area. By analyzing spatial changing character in the whole study area we can see that, the whole area was in the decreasing moment, especially in Jiangxia district and the urban district, this was caused by the urban development, and may led to the malign development of the environment, worthy of all aspects of attention. 2011 2011 Published Published by by Elsevier Elsevier Ltd. Ltd. Selection Selection and/or and/or peer-review peer-review under under responsibility responsibility of Conference of [name ESIAT2011 organizer] Organization Committee. Keywords: Percentage of Vegetation Cover Change; NDVI method for dimidiate pixel model; Wuhan; Remote Sensing 1. Introduction The percentage of vegetation cover is defined as area ratio of vegetation and the defined area, such as a pixel [1]. The change of vegetation coverage has important impact on the cycle of global energy and biochemistry of matter, and it is also an important index of regional ecological environment. So the information of ground vegetation coverage and its change has great realistic significance on revealing rule of ground spatial change, discussing the driving factor of change, and evaluating the regional ecological environment [2]. Wuhan lies in the east of Jianghan plain (113 41'E~11 0'E 29 0'N ~31 22'N), which is the largest industrial city in Hubei province and the central China. But with the development in economic, expansion of population and overspread of urban, it brought many problems. Such as the reduction of the function of the city structure, the pollution of the area and water environment and so on, especially in the ecological environment change which is brought by the reduction of the greenbelt of the urban and suburb, 1878-0296 2011 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of Conference ESIAT2011 Organization Committee. doi:10.1016/j.proenv.2011.09.234
Tao Chen et al. / Procedia Environmental Sciences 10 ( 2011 ) 1466 1472 1467 these problems have affected people's quality of life seriously. Therefore, it is very important to carry on the change detection of the percentage of vegetation cover in Wuhan city. In this paper, percentage of vegetation cover in different districts of Wuhan were calculated based on remote sensing images in 1988, 1991, 1996 and 2002 by employing NDVI method for dimidiate pixel model, then the vegetation coverage maps in different periods and different district were generated, and a dynamic analyze was done. The goal of this paper was to provide reference of the city s ecological environmental quality evaluation, city plan and ecological safety assessment by investigating the change of the city surface cover which caused by the development of the city and the change of the region. 2. Methodology Dimidiate Pixel Model Dimidiate Pixel Model [3-] assumes that a pixel consists of two components: pure vegetation and non-vegetation, so the reflectance of any pixel can be presented as follows: R = Rv + Rs (1) Where Rv is the reflectance of pure vegetation while Rs is the reflectance of non-vegetation. We assume that the vegetation coverage proportion of a pixel is fc, that is the vegetation cover, and then the non-vegetation coverage proportion of the pixel is 1- fc. If the whole pixel is covered by vegetation, the reflectance we gain is Rveg; if it has no vegetation coverage, the reflectance is Rsoil, so Rv and Rs of a mixed pixel can be presented as a product of Rveg and fc (showed in Eq. 2), Rsoil and 1- fc (showed in Eq. 3), respectively: Rv = fc * Rveg (2) Rs =(1- fc) * Rsoil (3) Through computing Eq. 1, Eq. 2 and Eq. 3 together, we acquire the equation of calculating percentage of vegetation cover as follows: fc = (R- Rsoil) / (Rveg- Rsoil) (4) Where Rveg and Rsoil are two key parameters of dimidiate pixel model. Obviously, if we get these two parameters, we can compute the vegetation cover by using remote sensing information through Eq. 4. Estimations Vegetation Coverage by Employing NDVI NDVI was defined as (NIR-R)/(NIR+R), where NIR and R is near infrared and visible red band reflectance, respectively. According to dimidiate pixel model, we can express the NDVI of each pixel as Eq. : fc = (NDVI - NDVIsoil) / (NDVIveg - NDVIsoil) () Where, NDVIveg is the NDVI of a pure vegetation pixel while NDVIsoil is the NDVI of a pure soil pixel. In theory, NDVIsoil should be zero for most soil types, but it changes from -0.1 to 0.2 [6-7] because of the influences of many factors. NDVIveg should be the maximum of NDVI, but it will change with the spatial or temporal change because of the influences of vegetation types. Thus, NDVIveg and NDVIsoil can not be fixed values [8] even in the same image. Vegetation type changes with the change of the land use type. To the same land use type, vegetation types are same approximately [9], so the pixels NDVIveg are close to the same vegetation type; the pixels NDVIsoil are close to the same soil type too. Thus, the land use map and the soil map may be used to compute NDVIveg and NDVIsoil.
1468 Tao Chen et al. / Procedia Environmental Sciences 10 ( 2011 ) 1466 1472 3. Data and Results Data Preparation The datasets used in this paper are: 4 scenes of Landsat TM and ETM+ images at spatial resolution of 30 m. Relief map of Wuhan city at 1:0000 scale; administrative map of China at 1:4000000 scale; soil map at 1:4000000 scale; land use map at 1:20000 scale and other related data. Data Pretreatment Firstly the relief map is used as reference to make geometry rectification of the Landsat TM and ETM+ remotely sensed data. Then by using the administrative map, the administrative districts of wuhan are cut out from the image, and then the NDVI images of the administrative districts are calculated. The NDVI maps in four times are shown in Fig. 1. All of these operations above are processed by the Erdas remote sensing software Figure 1 The NDVI maps of Wuhan in 1988,1991,1996,2002 from left to right Estimation Vegetation Cover This paper estimates vegetation coverage of different land use type in Wuhan city based on the NDVI method for dimidiate pixel model indifferent times. Fig. 2 showed the flow chart of percentage of vegetation cover calculation of Wuhan.
Tao Chen et al. / Procedia Environmental Sciences 10 (2011) 1466 1472 Figure 2 Flow chart of percentage of vegetation cover calculation of Wuhan Firstly, the image of the whole study area is obtained by overlapping boundary map of study area and the Landsat TM image, and then NDVI of this map is computed, thirdly, the land use map and the soil map of the drainage basin are used as reference to make a statistic of the whole NDVI value. According to the frequency statistical table, the NDVI frequency of % in the soil map is used as NDVIsoil, and that of 9% in the land use map is used as NDVIveg; at last, NDVIsoil and NDVIveg which obtained above are taken into equation Eq.. So we acquired the vegetation cover maps of the whole study area of different times (Fig. 3). Figure 3 The vegetation fraction map in Wuhan in 1988,1991,1996,2002 from left to right 4. Analysis and Discussion Spatial distribution and temporal change of percentage of vegetation cover change Fig.3 showed the percentage of vegetation cover maps in the 4 different periods. It showed the main character of vegetation coverage in Wuhan, which is, the urban city s vegetation coverage is lower than other districts. From 1988 to 2002, the highest mean vegetation coverage of the entire study area is in 1988, and the lowest one is in 2002 while the mean vegetation coverage of the entire study area decreased from 1988 to 1991. The mean vegetation coverage decreased 2.2% while the total industrial output value 1469
1470 Tao Chen et al. / Procedia Environmental Sciences 10 ( 2011 ) 1466 1472 of Wuhan increased 24%. Although the economic development keeps high speed, the mean vegetation cover of the study area increased 1.6% from 1991 to 1996. From 1996-2002, caused by the government increased the scale and speed of economic development, the mean vegetation coverage decreased sharpenly, which decreased about 7.4% in the entire study area especially in Jiangxia district. The results also showed that the vegetation coverage changed much from 1988 to 2002; the average vegetation coverage of the whole area was decreased from 8.41% to 0.4%, especially in Jiangxia district and the urban district. From 1996 to 2002 is the period which vegetation cover decreased the most sharply in this period. From 1991 to 1996 the vegetation cover was lightly increased in the whole study area. Spatial analyses of percentage of vegetation cover change By using mask module in Erdas 9.2 software, the vegetation coverage maps of four districts were extracted, which were urban city, Hannan district, Caidian district and Jiangxia district. Then the mean percentage of vegetation cover of each district were calculated. The results are shown in Table 1. Table 1 Dynamics of percentage of vegetation cover in different districts of Wuhan in 1988, 1991, 1996 and 2002 Administrative Name Urban Percentage of vegetation cover in different times (%) 1988 1991 1996 2002 0.4 6 1.7 7 3.6 Percentage of vegetation cover change during each period (%)* 1988-1991 1991-1996 43.30 1.31 1.91 1996-2002 - 10.37 Hannan 48.1 2.2 3.3 43.76 4.1 1.12-9.61 7 Caidian 7.3 8.8 8.6 49.99 1.3-0.21-8.69 6 9 8 Jiangxia 7.7 6.6 9.3-47.73-1.1 2.77 1 1 8 11.6 * negative values mean the vegetation cover is decreased (the same below) In this paper, we named Hannan district, Caidian district and Jiangxia district as suburban district, and the area of them was 69.8% of the whole study area. So we analyze the percentage of vegetation cover change of urban city, outskirts and the whole study area, respectively, the results are shown in Table 2. Table 2 Corporation of dynamics change of percentage of vegetation cover in urban city, suburban district and the whole study area in different times Name Urban suburban district whole study area 1 Percentage of vegetation cover in different times (%) 1988 1991 1996 2002 0.4 60.0 8.4 6 6 0 1.7 8.6 6.2 7 4 3.6 60.3 7.8 Percentage of vegetation cover during each period (%)* 1988-1991- 1996-1991 1996 2002 43.30 1.31 1.91-10.37 0.3-1.39 1.68-9.81 0.4-2.21 1.6-7.4
Tao Chen et al. / Procedia Environmental Sciences 10 ( 2011 ) 1466 1472 1471 The results showed that, during the process of conomic development of Wuhan, the urban development center was leant torwards suburbs gradually. The mean percentage of vegetation cover of urban district maintained momentum every year before 1996. The mean percentage of vegetation cover of the suburs increased between 1991 and 1996 except it decreased between 1988 and 1991. But between 1996 to 2002 the percentage of vegetation cover decreased sharply, which caused by the economic development. So from the percentage of vegetation cover change which we discussed above, we can see that, the conflict and challenge between the protection of ecological environment and the economic development always exist in Wuhan, which worthy of all aspects of attention on it.. Conclusion From the analysis which we discussed above, we can conclude that: a) The spatial distribution of vegetation coverage of Wuhan area is that, the urban city s vegetation coverage lower than other districts. b) In the whole area, the percentage of vegetation cover decreased from 8.41% to 0.4%, especially in Jiangxia and urban district. c) In the whole area, the percentage of vegetation cover decreased most sharply from 1996 to 2002 which caused by the economic development, while the vegetation coverage increased lightly from 1991 to 1996 in the whole study area. Acknowledgment This paper is funded by the Special Fund for Basic Scientific Research of Central Colleges, China University of Geosciences (Wuhan) (CUGL100206), the open fund of Key Laboratory of Agrometeorological Safeguard and Applied Technique, CMA (AMF20090), the open fund of State Key Lab Information Engineering in Surveying, Mapping & Remote Sensing, Wuhan University (09R02). References [1] Leprieur. C, Kerr. Y. H, Mastorchio. S, et al, Monitoring Vegetation Cover Across Semi-arid regions, Comparison of Remote Observations from Various Scales, International Journal of Remote Sensing, 2000, 21(2): 281 300. [2] Pan. Yaozhong, Li. Xiaobing, He. Chunyang, Research on comprehensive land cover classification in china: based on NOAA /AVHRR and Holdridge PE index, Quaternary Sciences, 2000, 20: 270 281 (In Chinese). [3] Chen. Jin, Chen. Yunhao, He. Chunyang, et al, Sub-pixel Model for Vegetation Fraction Estimation based on Land Cover Classif ication, Journal of Remote Sensing. 2001, (6):416 423 (In Chinese). [4] Leprieur. C, Verstraete. M. M, Pinty. B, Evaluation of the Performance of Various Vegetation Indices to Retrieve Vegetation Cover from AVHRR Data, Remote Sensing Review, 1994, (10):26 284. [] Zribi. M, Le. Hégarat-Mascle, et al, Derivation of Wild Vegetation Cover Density in Semi-arid Region: ERS2/SAR Evaluation, International Journal of Remote Sensing, 2003, (24):133 132. [6] Carlson. T. N, Ripley. D. A, On the Relation between NDVI, Fractional Vegetation Cover, and Leaf Area Index, Remote Sensing of Environment, 1997, 62(3):241 22. [7] Bradley. C. Rundquist, The Influence of Canopy Green Vegetation Fraction on Spectral Measurements Over Native Tall Grass Prairie, Remote Sensing of Environment, 2002, 81(1):129 13. [8] Kaufman. Y. J, Tanre. D, Atmospherically Resistant Vegetation Index (ARVI) for EOS-MODIS, IEEE Transactions on Geoscience and Remote Sensing, 1992, 30(2): 261 2701.
1472 Tao Chen et al. / Procedia Environmental Sciences 10 ( 2011 ) 1466 1472 [9] Chen. Tao, Niu. Ruiqing, Li. Pingxiang, Zhang. Liangpei, An Artificial Neural Network Method for Vegetation Cover Retrieval with Beijing-1 Microsatellite Data, Remote Sensing Technology and Application, 2010, 2(1): 24 30 (In Chinese)