- Email: [email protected]
Spatial-temporal Change of Sanshui district's Dike-pond from 1979-2009
Available online at www.sciencedirect.com
Physics Procedia 25 (2012) 452 – 458
2012 International Conference on Solid State Devices and Materials Science
Spatial-temporal Change of Sanshui district’s Dike-pond from 1979-2009* Jiaxing Liu, Jianfei Chen1 and Xiaoxuan Wang School of Geographical Sciences Guangzhou University Guangzhou, 510006, China
Abstract Dike-pond is a representative style of ecological agriculture in the PRD(Pearl River Delta). Since 1992, Guangdong quicken its pace of reform and opening-up to the outside world. A mass of factories had been built in the PRD. The dike-ponds have come across some influential changes in the recent 30 years. To detect and study on the changes of dike-ponds, the Remote Sensing and Geography Information System skill was applied in this paper. This article selected Sanshui district as an example and used Landsat TM 1979, 1990, 2000 and SPOT 2009 satellite image as the major data sources. With the help of ITTVIS company newly released software-ENVI EX, object-oriented approach has been used to extract the dike-pond land from each image. The result indicates that the area of dike-pond gained rapidly growth from 1979 to 2000, but decrease critically during 2000-2009. When using Change Detection Analysis to compute each period’s change statistics, the result shown that the increased dike-pond area were mainly from vegetation covered land and other bare land. Then we found out that the mean centre of Sanshui district’s dike-pond was moving from northwest to southeast during 1979-2009. Therefore, it comes to the conclusion that Sanshui district’s dike-pond increased across the southeast of Sanshui district from 1979 to 2009. Last but not least, some suggestions have been put forward to keep the dike-pond land area from decreasing.
© 2012 Elsevier Ltd. B.V.Selection Selectionand/or and/orpeer-review peer-reviewunder underresponsibility responsibility Garry Lee 2011 Published by Elsevier of of [name organizer] Keywords: dike-pond; object-oriented approach; spatial-temporal change; Sanshui district
1. Introduction Sanshui district, located in the South China and situated at the north of PRD(pearl river delta) is famous for its mulberry fish ponds, as well as Shunde district and Nanhai district. There has a South-Subtropics climate with an annual average temperature between 22ć to 26ć. The annual rainfall ranges from 1200 to 1800 mm approximately but varies a lot seasonally. The PRD people benefit a lot *
This work is supported by Guangdong Natural Science Foundation, China (number: 9151051501000030).
1875-3892 © 2012 Published by Elsevier B.V. Selection and/or peer-review under responsibility of Garry Lee doi:10.1016/j.phpro.2012.03.110
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
from the Pearl River and its abundant tributaries. Conditions in the area for developing the dike-pond system are therefore good. Sanshui district utilize the advantage of the rich water resource to develop the fishery economy. The dike-pond system is an excellent artificial ecological agriculture system, and can be suitable for areas resembling to the PRD elsewhere in Asia, Africa and South America, especially for low-lying water-logged land which is normally reclaimed by constructing water conservancy projects[1]. It has an obvious land-water interaction and remarkable eco-economical effects which is regarded as a sustainable agro-ecosystem in the low-lying area for centuries[2-4]. However, since the proposed of reform and opening up policies in 1978, especially after 1992, Guangdong Province has quicken its pace of reform and opening-up. With the rapid urbanization and industrialization, the land-use and land cover in the PRD changes a lot[5-7]. The transfer of land-use type and structure also causes the alterations in the integration of dike-pond system’s structures, spatial patterns, and species diversities[4, 6, 8-12]. To find out the spatial-temporal change regulation of Sanshui district’s dike-pond system, the following base on both the quantitative method and qualitative methods. Remote sensing image is an important data source in dynamic monitoring of land use. Some researchers have done quite a lot of works in analysis the dike-pond land distribution. Wang Xingling and Kang Qing uses Landsat TM image as a data source trying to extract the dike-pond land in the PRD[13], based on the situation of changing dike-pond land. He Zhijian also applied the Remote sensing skill in the study of dike-pond land’s sustainable development of PRD[14]. Meanwhile, Gao Ai and Liu Kai used landscape metric to analysis the dike-pond land use change, they drew the conclusion that the dike-pond land area dropped severely in the past 30 years[15, 16]. However, all of them use a traditional pixel basic method to extract information from the Remote sensing image. It’s noticed that most of the dike-ponds have a similarly area and shape. So we decided to use a classification of object-oriented approach to extract the dike-pond land. The following operations are mainly under the ITTVIS ENVI EX software. The first step is image segmentation, then integrated the information of spectrum, shape, texture, etc to classify the land use type. The classification result then converts to a GIS vector format analyzing Sanshui district’s dike-pond land use spatial-temporal change. If we know the causation and regularities of this dynamic change, a reasonable planning and reliable management would be made out. 2. Methology
2.1
Basic informations
Sanshui district belong to Foshan city, which locate in the south-east of China. It covers the latitude around 22°58ƍ~23°34ƍN, longitude around 112°46ƍ~113°02ƍE, with an area of 874.22km2. Along with Guangzhou, Zhongshan, Jiangmen and other cities crowded in the PRD(Fig.1). With the sub-tropical monsoon climate, Sanshui district has an annual average temperature of around 22ć to 26ć, annual precipitation about 1200 to 1800 mm. Between 2003-2008 Sanshui district’s an uninterrupted average GDP growth of 51.67%. De jure population 531.3 thousand by the end of 2008. Sanshui district distributes a spider-web-like canals, the topography of this area mainly plain. Since the opening-up, the economy booms in China. Sanshui district develops quickly, soon it becomes one of the fastest developing districts around the PRD area. With the remarkable progress of industry, great changes have taken place in this area, which include the remarkable change of the land use types.
453
454
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
Fig. 1 Location of the study area in Guangdong Province, produced by ASTER data.
2.2
Data source
the major data source is Landsat MSS, TM and SPOT5 remote sensing image, the images acquire date are 1979-10-19, 1990-10-13, 2000-01-02, and mosaicked SPOT image between 2007-11-28 and 2009-01-11, all of the four map have been Geometrical Correction, Atmospheric Correction, Image registration etc. The Administrative Map of Guangdong Province, and 30×30 meters resolution DEM data from ASTER GDEM. The survey data includes the updated data of Sanshui district’s land use map, field work, and Google Earth’s high resolution image. 2.3. Object-oriented approach extracting the dike-pond land use dike-pond land use type has a similar spectral Characteristic with other water substance, using a traditional pixel basic classification couldn’t extract the dike-pond exactly[17]. However, judge from the appearance of the fish-scale-like dike-pond spatial distribution, and it’s slightness rivers and canals. It’s easy to separate the other waters from dike-pond water, so we applied an object-oriented approach, integrate with the spectrum, shape, texture, etc. to extract the dike-pond land use[18-21]. In this case, we use an object-oriented approach model under ITTVIS ENVI EX software—the Feature Extraction. To extract the dike-pond land, firstly adopt an edge-based segmentation algorithm to process each image, then separate the water substance from terrene using a spectral based NDWI(Normalize Difference Water Index)[20]. Finally build rules such as area, elongation etc to segment objects like rivers, canals, lakes, etc.. The left are mainly dike-ponds. Then we make further investigation to omit the non dike-pond land. The results are as follow(Fig.2), the average accuracy achieved for this four classification map is 85.5% and the Kappa co-efficient is 0.83.
455
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
Fig. 2 Extraction results of dike-pond area in 1979, 1990, 2000 and 2009
3. Calculation Results The result shows that the dike-pond area and its spatial distribution change a lot during 1979-2009. To analysis the spatial-temporal change of Sanshui district’s dike-pond from 1979-2009, we selected the single land use dynamic index, the model on changes of gravity centre, and the land use conversion matrix as a quantitative study method. 3.1 dike-pond land use dynamic index Land use dynamic index is one of main models for depicting the change rate of land use types. We use this index to calculate the dynamic index of dike-pond, the formula of land use dynamic index is as follow:
K
Ua Ub 1 x x 100% Ua T
(1)
K means the dynamic index in a specify period; Ua, Ub equals to the beginning and the final amount of a certain land use type. T means the Time Bucket, the unit usually year. We known the beginning and the final amount of dike-pond land area, Table I calculated the Land use dynamic index of dike-pond during 1979-1990, 1990-2000, 2000-2009, and 1979-2009. TABLE I Dike-pond land use dynamic index of Sanshui district Time Bucket Land use dynamic index/%
1979-1990 +21.06
1990-2000 +8.68
2000-2009 -4.15
1979-2009 +9.60
1. +: gain; -:loss
From TABLE I, we known that during 1979-2000, the dike-ponds are in an increase trend, the major course is that the structure of agricultural production change from the traditional crop farming to a more economic benefit aquaculture—the so call three high agricultures (high yield, high quality and high efficiency agriculture). In 2009, Sanshui district was selected as a typical rural in Guangdong since the opening up. But during 2000-2009, the dike-pond area decreased 4.15% every year. All in all, between 1979-2009, the dike-pond area increased 9.6% every year. 3.2 Analysis on dike-pond’s gravity centre change
456
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
A region’s land use spatial change can be reflected from the change of gravity centre. The following formula shows how to calculate the gravity centre: n
Xt
n
¦ (C g x ) / ¦ C ti
i 1
Yt
i
i 1
n
n
¦ (C g \ ) / ¦ C ti
i 1
(2)
ti
i
(3)
ti
i 1
In this two formula, Xt, Yt means a certain land use type barycentric coordinates at t year; Cti means the i unit of the whole region a certain land use type’s area; Xi, Yi represent the whole i unit’s barycentric coordinates; n means the quantity of unit over the region. The gravity centre of land use type change, to some extent, reflected the change of a region’s landscape pattern. When calculated the 1979, 1990, 2000, 2009 dike-pond gravity centre, we adopted Euclidian Distance Formula to attain the migration distance(meter) and migration direction(degree) during 1979-2009. The start migration direction points to east, counterclockwise rotation. The results are as following(TABLE II): TABLE II Change migration distance and direction of dike-pond area in Sanshui district period 1979-1990 1990-2000 2000-2009 1979-2009
migration distance(m) 3809 105 1914 2254
migration direction(°) 292.69 350.85 90.05 314.19
During 1979-1990, the migration distance is 3809m, migrating to the south. Later during 1990-2000, its migration distance is 105m, migrating to the east. Then migrate to the North, the distance is 1914m. Totally, the migration distance from 1979 to 2009 is 2254m, migrate to the south-east. The Chancheng district is to the south of Sanshui district, and it’s the center of Foshan’s government, financial, cultural, education, technology, information and logistics. Considering the effects of Chancheng district’s economy development and the advance of transportation network, as well as the population increased, more land is reclaimed for agriculture and aquatic product. 3.3 Land use conversion matrix of dike-pond To specific the detail area of changed dike-pond land, we furthermore classify the changed parts into five types of land use, they are vegetation covered land, dike-pond, building area, river and other land. By using Change Detection function under ENVI software, the initial state select 1979’s land use map, and the final state select 2009’s land use map. The Change Detection Statistics window contains all of the changed statistics, the statistics show that during 1979-2009, 31% of the decrease vegetation cover land changed into construction land, the left change into dike-pond and other vegetation land. Still there are 60% other bare land changed into dike-pond comes from cultivated land and woodland. The above statistics indicated that Sanshui district experienced a period of reclamations of bare land and low lying land. Later, accompany with the urban sprawl, some dike-pond was embed to afford construction land to the city. 4. Conclusion and discussion Through the analysis of area and spatial-temporal change of dike-pond in Sanshui district during 1979-2009, we noticed that both of the area and distribution changed severely. Where during 1979-2000, the area of dike-pond increased, but decreased during 2000-2009. The gravity centre analysis indicates
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
that the dike-pond land gravity centre migrate from north-west to the south-east, which is approaching to the administrative centre of Foshan. And the Change Detection represents the decreased dike-ponds were mainly changed into construction land, and the increased dike-ponds were mainly from vegetation cover land and other bare land. Aim at the rapid change of dike-ponds, we should be guided by the Scientific Concept of Development, putting forward some sustainable development policy. To prevent the dike-pond area from decreasing, firstly we should apply strict implementation of land use planning, adjust a reasonable proportion of dike-pond land is also important, and strengthen the construction and management of the dike-pond area, meanwhile, we should also exploit potentialities of dike-pond productive capacity. To improve land use efficiency, not only increase the area of dike-pond, but also emphasize scientific cultivation. Last but not least, neaten and enlarge the causeway of the dike-pond, plant some deep root agriculture to strength the Dike area, as well as avoid the pond from pollution. Now that the dike-pond area gravity centre is turning to the south-east. The trend of urbanization couldn’t be stopped, but we still should take some solution to reduce the occupation of dike-pond. So secondly, raise the cost of changing dike-pond into nonfarm land and increase the comparative advantage of dike-pond industry. Among the scope of economics, if the cost of utilizing the vacant land and other unused land lower than the cost of utilizing the dike-pond land, they will choose not to occupy the dike-pond land. So increasing the cost of occupying the dike-pond land, the economic lever can prevent and reduce the use of dike-pond land. On the other hand, we can take advantage of the dike-pond land to improve the relative economic benefit of developing the dike-pond’s crop farming and fish farming. To form an aquiculture economies of scale, it is required to promote local aquiculture industrial competition and maintain the dike-pond aquiculture high yield and high quality. The land use conversion matrix indicates the urban expansion rules that construction area expands from the core of the city, then spread to the suburban area. Course by the low cost of land consolidation in this area, and its more significant for a developing city to take up lands close to the downtown. Then the farmers reclaim the cultivated land and woodland to make dike-ponds. For the purpose of land use sustainable development, finally, we suggest that the local government and farmer expand the capital and technical input and realize the dike-pond intensive aquiculture. The fishery of dike-pond is mainly household-based, decentralized management, which are helpless in the market competition. Set up some large dike-pond-product-orient companies, study on the marketing tactic, upgrade the product and package level, adjust product structure and enhance product added value. Only if the farmers income increased, the dike-pond aquaculture could be well developed. Build an international brand of dike-pond tourist amusement attraction and promote the dike-pond ecotourism development. Mulberry fish pond is a unique traditional comprehensive land use system in the PRD, if there is a farming ecological village keeping the typical features of mulberry fish ponds will make you cry out: there should be such a Shangrila in today’s stony forest. Currently, Shunde district opened up a national 4A level Farmhouse theme park. It took the fancy of many visitors. Therefore, we suggest that Sanshui district should widen its dike-pond advertisement, fully utilize the mass media to recommend its mulberry fish ponds landscape, exploit the inner province tourism market, and establish national famous brands. Acknowledgements This work is partially supported by Guangdong Natural Science Foundation, China (number: 9151051501000030) References [1] Z. Gongfu, "The Types, Structure and Results and Results of the Dike-Pond System in South China," GeoJournal, pp. 83-89, 1990.
457
458
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458 [2] J. Ding, et al., "Studies on Sustainable Development of Aquaculture in the Dike-Pond Ecosystem," Chongqing
Environmental Science, vol. 23, pp. 12-14, 2001. [3] C. Nie and H. Li, "The integrated Dike-pond system s: its new developments and ecological problems," Journal of FoshanUniversity (Natural Science Edition), vol. 19, pp. 49-53, 2001. [4] H. Li, et al., "Reconstruction and control of modern in tensive dike-pond system in Shunde," Fisheries Sicence & Technology Information, vol. 1, pp. 23-26, 2005. [5] C. Zhiliang, et al., "Land Use/Cover Changes between 1990 and 2000 based on Remote Sensing and GIS in Pearl River Delta, China," presented at the 16th International Conference on GeoInformatics and the Joint Conference, Guangzhou, 2008. [6] L. Huiqing, "Study on LUCC in Typical County of the Pearl River Delta--Case Study of Shunde District," Guangzhou University, 2007. [7] H. Weiping and H. Jianbang, "Built-up Land-cover Change Monitoring by Remote Sensing of Urban Area in the Pearl River Delta Based on GIS," JOURNAL OF REMOTE SENSING vol. 7, pp. 201-206, 2003. [8] D. Fen, "Sangji fish - the main agricultural characteristics of the Pearl River Delta," AGRICULTURAL ARCHAEOLOGY, pp. 193-196, 2003. [9] Z. Feng and Y. Wen, "Effects of Fish Aquaculture on Water Environment in the Zhujiang River Delta," Acta Scifntiarum Naturalium Universitatis Sunyaatseni, vol. 24, pp. 103-106, 2004. [10] H. Lu, et al., "Energy value evaluation of dike-pond agro-ecological engineer models," Chinese Journal of applied ecology, vol. 14, pp. 1622-1626, 2003. [11] M. Li, et al., "Establishment of indicator system for assessing eco-environment quality on dike-pond system," Journal of Agro-Environment Science, vol. 26, pp. 386-390, 2007. [12] L. Huiqing, "Pearl River Delta region cycle mode of agricultural development," Rural Economy and Science-Technology, pp. 29-30, 2006. [13] W. Xinling and K. Qing, "Application of tm image in investigating changes of dike-pond land on the pearl river delta—samples as shunde, nanhai dike-pond region," REMOTE SENSING FOR LAND & RESOURCES, pp. 8-14, 1997. [14] H. Zhijian, et al., "Sustainable Development of Mulberry Dike and Fish Pond Farming in Pearl River Delta by Remote Sensing Technology," ACTA SCIENTIARUM NATURALIUM UNIVERSITATIS SUNYATSEN, vol. 37, pp. plusˈ64-69, 1998. [15] G. Ai, et al., "Changes of Landscape Pattern in Shunde District of Foshan City," TROPICAL GEOGRAPHY, vol. 27, pp. 327-331,359, 2007. [16] L. Kai, et al., "Spatial Evolution Analysis of Dike-Pond Systems in Foshan City," TROPICAL GEOGRAPHY vol. 28, pp. 513-517, 2008 [17] G. Xue-bin, et al., "Remote sensing monitoring of aquaculture and automatic information extraction," Remote Sensing for Land & Resources, pp. 41-44, 2009. [18] X. Yu-lin, et al., "An Object-oriented Approach for Extracting Farm Waters within Coastal Belts," Remote Sensing Technology and Application, vol. 24, pp. 68-72, 2009. [19] C. Jianyu, et al., "Simple Object Segmentation Near Coastal Belt s in High Spatial Resolution Remote Sensing Data," Science China, vol. 36, pp. 1044-1051, 2006. [20] Y. Mei-chun, "Extraction of information about water area change in Nanjing City in the past ten years based on TM data," Water Resources Protection, vol. 21, pp. 31-33, 2005. [21] S. Xiaoyu, et al., "Analyses on Spatial-Temporal Changes in Aquaculture Land in Coastal Areas of the Pearl River Estuarine," Resources Science, vol. 32, pp. 71-77, 2010.
Physics Procedia 25 (2012) 452 – 458
2012 International Conference on Solid State Devices and Materials Science
Spatial-temporal Change of Sanshui district’s Dike-pond from 1979-2009* Jiaxing Liu, Jianfei Chen1 and Xiaoxuan Wang School of Geographical Sciences Guangzhou University Guangzhou, 510006, China
Abstract Dike-pond is a representative style of ecological agriculture in the PRD(Pearl River Delta). Since 1992, Guangdong quicken its pace of reform and opening-up to the outside world. A mass of factories had been built in the PRD. The dike-ponds have come across some influential changes in the recent 30 years. To detect and study on the changes of dike-ponds, the Remote Sensing and Geography Information System skill was applied in this paper. This article selected Sanshui district as an example and used Landsat TM 1979, 1990, 2000 and SPOT 2009 satellite image as the major data sources. With the help of ITTVIS company newly released software-ENVI EX, object-oriented approach has been used to extract the dike-pond land from each image. The result indicates that the area of dike-pond gained rapidly growth from 1979 to 2000, but decrease critically during 2000-2009. When using Change Detection Analysis to compute each period’s change statistics, the result shown that the increased dike-pond area were mainly from vegetation covered land and other bare land. Then we found out that the mean centre of Sanshui district’s dike-pond was moving from northwest to southeast during 1979-2009. Therefore, it comes to the conclusion that Sanshui district’s dike-pond increased across the southeast of Sanshui district from 1979 to 2009. Last but not least, some suggestions have been put forward to keep the dike-pond land area from decreasing.
© 2012 Elsevier Ltd. B.V.Selection Selectionand/or and/orpeer-review peer-reviewunder underresponsibility responsibility Garry Lee 2011 Published by Elsevier of of [name organizer] Keywords: dike-pond; object-oriented approach; spatial-temporal change; Sanshui district
1. Introduction Sanshui district, located in the South China and situated at the north of PRD(pearl river delta) is famous for its mulberry fish ponds, as well as Shunde district and Nanhai district. There has a South-Subtropics climate with an annual average temperature between 22ć to 26ć. The annual rainfall ranges from 1200 to 1800 mm approximately but varies a lot seasonally. The PRD people benefit a lot *
This work is supported by Guangdong Natural Science Foundation, China (number: 9151051501000030).
1875-3892 © 2012 Published by Elsevier B.V. Selection and/or peer-review under responsibility of Garry Lee doi:10.1016/j.phpro.2012.03.110
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
from the Pearl River and its abundant tributaries. Conditions in the area for developing the dike-pond system are therefore good. Sanshui district utilize the advantage of the rich water resource to develop the fishery economy. The dike-pond system is an excellent artificial ecological agriculture system, and can be suitable for areas resembling to the PRD elsewhere in Asia, Africa and South America, especially for low-lying water-logged land which is normally reclaimed by constructing water conservancy projects[1]. It has an obvious land-water interaction and remarkable eco-economical effects which is regarded as a sustainable agro-ecosystem in the low-lying area for centuries[2-4]. However, since the proposed of reform and opening up policies in 1978, especially after 1992, Guangdong Province has quicken its pace of reform and opening-up. With the rapid urbanization and industrialization, the land-use and land cover in the PRD changes a lot[5-7]. The transfer of land-use type and structure also causes the alterations in the integration of dike-pond system’s structures, spatial patterns, and species diversities[4, 6, 8-12]. To find out the spatial-temporal change regulation of Sanshui district’s dike-pond system, the following base on both the quantitative method and qualitative methods. Remote sensing image is an important data source in dynamic monitoring of land use. Some researchers have done quite a lot of works in analysis the dike-pond land distribution. Wang Xingling and Kang Qing uses Landsat TM image as a data source trying to extract the dike-pond land in the PRD[13], based on the situation of changing dike-pond land. He Zhijian also applied the Remote sensing skill in the study of dike-pond land’s sustainable development of PRD[14]. Meanwhile, Gao Ai and Liu Kai used landscape metric to analysis the dike-pond land use change, they drew the conclusion that the dike-pond land area dropped severely in the past 30 years[15, 16]. However, all of them use a traditional pixel basic method to extract information from the Remote sensing image. It’s noticed that most of the dike-ponds have a similarly area and shape. So we decided to use a classification of object-oriented approach to extract the dike-pond land. The following operations are mainly under the ITTVIS ENVI EX software. The first step is image segmentation, then integrated the information of spectrum, shape, texture, etc to classify the land use type. The classification result then converts to a GIS vector format analyzing Sanshui district’s dike-pond land use spatial-temporal change. If we know the causation and regularities of this dynamic change, a reasonable planning and reliable management would be made out. 2. Methology
2.1
Basic informations
Sanshui district belong to Foshan city, which locate in the south-east of China. It covers the latitude around 22°58ƍ~23°34ƍN, longitude around 112°46ƍ~113°02ƍE, with an area of 874.22km2. Along with Guangzhou, Zhongshan, Jiangmen and other cities crowded in the PRD(Fig.1). With the sub-tropical monsoon climate, Sanshui district has an annual average temperature of around 22ć to 26ć, annual precipitation about 1200 to 1800 mm. Between 2003-2008 Sanshui district’s an uninterrupted average GDP growth of 51.67%. De jure population 531.3 thousand by the end of 2008. Sanshui district distributes a spider-web-like canals, the topography of this area mainly plain. Since the opening-up, the economy booms in China. Sanshui district develops quickly, soon it becomes one of the fastest developing districts around the PRD area. With the remarkable progress of industry, great changes have taken place in this area, which include the remarkable change of the land use types.
453
454
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
Fig. 1 Location of the study area in Guangdong Province, produced by ASTER data.
2.2
Data source
the major data source is Landsat MSS, TM and SPOT5 remote sensing image, the images acquire date are 1979-10-19, 1990-10-13, 2000-01-02, and mosaicked SPOT image between 2007-11-28 and 2009-01-11, all of the four map have been Geometrical Correction, Atmospheric Correction, Image registration etc. The Administrative Map of Guangdong Province, and 30×30 meters resolution DEM data from ASTER GDEM. The survey data includes the updated data of Sanshui district’s land use map, field work, and Google Earth’s high resolution image. 2.3. Object-oriented approach extracting the dike-pond land use dike-pond land use type has a similar spectral Characteristic with other water substance, using a traditional pixel basic classification couldn’t extract the dike-pond exactly[17]. However, judge from the appearance of the fish-scale-like dike-pond spatial distribution, and it’s slightness rivers and canals. It’s easy to separate the other waters from dike-pond water, so we applied an object-oriented approach, integrate with the spectrum, shape, texture, etc. to extract the dike-pond land use[18-21]. In this case, we use an object-oriented approach model under ITTVIS ENVI EX software—the Feature Extraction. To extract the dike-pond land, firstly adopt an edge-based segmentation algorithm to process each image, then separate the water substance from terrene using a spectral based NDWI(Normalize Difference Water Index)[20]. Finally build rules such as area, elongation etc to segment objects like rivers, canals, lakes, etc.. The left are mainly dike-ponds. Then we make further investigation to omit the non dike-pond land. The results are as follow(Fig.2), the average accuracy achieved for this four classification map is 85.5% and the Kappa co-efficient is 0.83.
455
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
Fig. 2 Extraction results of dike-pond area in 1979, 1990, 2000 and 2009
3. Calculation Results The result shows that the dike-pond area and its spatial distribution change a lot during 1979-2009. To analysis the spatial-temporal change of Sanshui district’s dike-pond from 1979-2009, we selected the single land use dynamic index, the model on changes of gravity centre, and the land use conversion matrix as a quantitative study method. 3.1 dike-pond land use dynamic index Land use dynamic index is one of main models for depicting the change rate of land use types. We use this index to calculate the dynamic index of dike-pond, the formula of land use dynamic index is as follow:
K
Ua Ub 1 x x 100% Ua T
(1)
K means the dynamic index in a specify period; Ua, Ub equals to the beginning and the final amount of a certain land use type. T means the Time Bucket, the unit usually year. We known the beginning and the final amount of dike-pond land area, Table I calculated the Land use dynamic index of dike-pond during 1979-1990, 1990-2000, 2000-2009, and 1979-2009. TABLE I Dike-pond land use dynamic index of Sanshui district Time Bucket Land use dynamic index/%
1979-1990 +21.06
1990-2000 +8.68
2000-2009 -4.15
1979-2009 +9.60
1. +: gain; -:loss
From TABLE I, we known that during 1979-2000, the dike-ponds are in an increase trend, the major course is that the structure of agricultural production change from the traditional crop farming to a more economic benefit aquaculture—the so call three high agricultures (high yield, high quality and high efficiency agriculture). In 2009, Sanshui district was selected as a typical rural in Guangdong since the opening up. But during 2000-2009, the dike-pond area decreased 4.15% every year. All in all, between 1979-2009, the dike-pond area increased 9.6% every year. 3.2 Analysis on dike-pond’s gravity centre change
456
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
A region’s land use spatial change can be reflected from the change of gravity centre. The following formula shows how to calculate the gravity centre: n
Xt
n
¦ (C g x ) / ¦ C ti
i 1
Yt
i
i 1
n
n
¦ (C g \ ) / ¦ C ti
i 1
(2)
ti
i
(3)
ti
i 1
In this two formula, Xt, Yt means a certain land use type barycentric coordinates at t year; Cti means the i unit of the whole region a certain land use type’s area; Xi, Yi represent the whole i unit’s barycentric coordinates; n means the quantity of unit over the region. The gravity centre of land use type change, to some extent, reflected the change of a region’s landscape pattern. When calculated the 1979, 1990, 2000, 2009 dike-pond gravity centre, we adopted Euclidian Distance Formula to attain the migration distance(meter) and migration direction(degree) during 1979-2009. The start migration direction points to east, counterclockwise rotation. The results are as following(TABLE II): TABLE II Change migration distance and direction of dike-pond area in Sanshui district period 1979-1990 1990-2000 2000-2009 1979-2009
migration distance(m) 3809 105 1914 2254
migration direction(°) 292.69 350.85 90.05 314.19
During 1979-1990, the migration distance is 3809m, migrating to the south. Later during 1990-2000, its migration distance is 105m, migrating to the east. Then migrate to the North, the distance is 1914m. Totally, the migration distance from 1979 to 2009 is 2254m, migrate to the south-east. The Chancheng district is to the south of Sanshui district, and it’s the center of Foshan’s government, financial, cultural, education, technology, information and logistics. Considering the effects of Chancheng district’s economy development and the advance of transportation network, as well as the population increased, more land is reclaimed for agriculture and aquatic product. 3.3 Land use conversion matrix of dike-pond To specific the detail area of changed dike-pond land, we furthermore classify the changed parts into five types of land use, they are vegetation covered land, dike-pond, building area, river and other land. By using Change Detection function under ENVI software, the initial state select 1979’s land use map, and the final state select 2009’s land use map. The Change Detection Statistics window contains all of the changed statistics, the statistics show that during 1979-2009, 31% of the decrease vegetation cover land changed into construction land, the left change into dike-pond and other vegetation land. Still there are 60% other bare land changed into dike-pond comes from cultivated land and woodland. The above statistics indicated that Sanshui district experienced a period of reclamations of bare land and low lying land. Later, accompany with the urban sprawl, some dike-pond was embed to afford construction land to the city. 4. Conclusion and discussion Through the analysis of area and spatial-temporal change of dike-pond in Sanshui district during 1979-2009, we noticed that both of the area and distribution changed severely. Where during 1979-2000, the area of dike-pond increased, but decreased during 2000-2009. The gravity centre analysis indicates
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458
that the dike-pond land gravity centre migrate from north-west to the south-east, which is approaching to the administrative centre of Foshan. And the Change Detection represents the decreased dike-ponds were mainly changed into construction land, and the increased dike-ponds were mainly from vegetation cover land and other bare land. Aim at the rapid change of dike-ponds, we should be guided by the Scientific Concept of Development, putting forward some sustainable development policy. To prevent the dike-pond area from decreasing, firstly we should apply strict implementation of land use planning, adjust a reasonable proportion of dike-pond land is also important, and strengthen the construction and management of the dike-pond area, meanwhile, we should also exploit potentialities of dike-pond productive capacity. To improve land use efficiency, not only increase the area of dike-pond, but also emphasize scientific cultivation. Last but not least, neaten and enlarge the causeway of the dike-pond, plant some deep root agriculture to strength the Dike area, as well as avoid the pond from pollution. Now that the dike-pond area gravity centre is turning to the south-east. The trend of urbanization couldn’t be stopped, but we still should take some solution to reduce the occupation of dike-pond. So secondly, raise the cost of changing dike-pond into nonfarm land and increase the comparative advantage of dike-pond industry. Among the scope of economics, if the cost of utilizing the vacant land and other unused land lower than the cost of utilizing the dike-pond land, they will choose not to occupy the dike-pond land. So increasing the cost of occupying the dike-pond land, the economic lever can prevent and reduce the use of dike-pond land. On the other hand, we can take advantage of the dike-pond land to improve the relative economic benefit of developing the dike-pond’s crop farming and fish farming. To form an aquiculture economies of scale, it is required to promote local aquiculture industrial competition and maintain the dike-pond aquiculture high yield and high quality. The land use conversion matrix indicates the urban expansion rules that construction area expands from the core of the city, then spread to the suburban area. Course by the low cost of land consolidation in this area, and its more significant for a developing city to take up lands close to the downtown. Then the farmers reclaim the cultivated land and woodland to make dike-ponds. For the purpose of land use sustainable development, finally, we suggest that the local government and farmer expand the capital and technical input and realize the dike-pond intensive aquiculture. The fishery of dike-pond is mainly household-based, decentralized management, which are helpless in the market competition. Set up some large dike-pond-product-orient companies, study on the marketing tactic, upgrade the product and package level, adjust product structure and enhance product added value. Only if the farmers income increased, the dike-pond aquaculture could be well developed. Build an international brand of dike-pond tourist amusement attraction and promote the dike-pond ecotourism development. Mulberry fish pond is a unique traditional comprehensive land use system in the PRD, if there is a farming ecological village keeping the typical features of mulberry fish ponds will make you cry out: there should be such a Shangrila in today’s stony forest. Currently, Shunde district opened up a national 4A level Farmhouse theme park. It took the fancy of many visitors. Therefore, we suggest that Sanshui district should widen its dike-pond advertisement, fully utilize the mass media to recommend its mulberry fish ponds landscape, exploit the inner province tourism market, and establish national famous brands. Acknowledgements This work is partially supported by Guangdong Natural Science Foundation, China (number: 9151051501000030) References [1] Z. Gongfu, "The Types, Structure and Results and Results of the Dike-Pond System in South China," GeoJournal, pp. 83-89, 1990.
457
458
Jiaxing Liu et al. / Physics Procedia 25 (2012) 452 – 458 [2] J. Ding, et al., "Studies on Sustainable Development of Aquaculture in the Dike-Pond Ecosystem," Chongqing
Environmental Science, vol. 23, pp. 12-14, 2001. [3] C. Nie and H. Li, "The integrated Dike-pond system s: its new developments and ecological problems," Journal of FoshanUniversity (Natural Science Edition), vol. 19, pp. 49-53, 2001. [4] H. Li, et al., "Reconstruction and control of modern in tensive dike-pond system in Shunde," Fisheries Sicence & Technology Information, vol. 1, pp. 23-26, 2005. [5] C. Zhiliang, et al., "Land Use/Cover Changes between 1990 and 2000 based on Remote Sensing and GIS in Pearl River Delta, China," presented at the 16th International Conference on GeoInformatics and the Joint Conference, Guangzhou, 2008. [6] L. Huiqing, "Study on LUCC in Typical County of the Pearl River Delta--Case Study of Shunde District," Guangzhou University, 2007. [7] H. Weiping and H. Jianbang, "Built-up Land-cover Change Monitoring by Remote Sensing of Urban Area in the Pearl River Delta Based on GIS," JOURNAL OF REMOTE SENSING vol. 7, pp. 201-206, 2003. [8] D. Fen, "Sangji fish - the main agricultural characteristics of the Pearl River Delta," AGRICULTURAL ARCHAEOLOGY, pp. 193-196, 2003. [9] Z. Feng and Y. Wen, "Effects of Fish Aquaculture on Water Environment in the Zhujiang River Delta," Acta Scifntiarum Naturalium Universitatis Sunyaatseni, vol. 24, pp. 103-106, 2004. [10] H. Lu, et al., "Energy value evaluation of dike-pond agro-ecological engineer models," Chinese Journal of applied ecology, vol. 14, pp. 1622-1626, 2003. [11] M. Li, et al., "Establishment of indicator system for assessing eco-environment quality on dike-pond system," Journal of Agro-Environment Science, vol. 26, pp. 386-390, 2007. [12] L. Huiqing, "Pearl River Delta region cycle mode of agricultural development," Rural Economy and Science-Technology, pp. 29-30, 2006. [13] W. Xinling and K. Qing, "Application of tm image in investigating changes of dike-pond land on the pearl river delta—samples as shunde, nanhai dike-pond region," REMOTE SENSING FOR LAND & RESOURCES, pp. 8-14, 1997. [14] H. Zhijian, et al., "Sustainable Development of Mulberry Dike and Fish Pond Farming in Pearl River Delta by Remote Sensing Technology," ACTA SCIENTIARUM NATURALIUM UNIVERSITATIS SUNYATSEN, vol. 37, pp. plusˈ64-69, 1998. [15] G. Ai, et al., "Changes of Landscape Pattern in Shunde District of Foshan City," TROPICAL GEOGRAPHY, vol. 27, pp. 327-331,359, 2007. [16] L. Kai, et al., "Spatial Evolution Analysis of Dike-Pond Systems in Foshan City," TROPICAL GEOGRAPHY vol. 28, pp. 513-517, 2008 [17] G. Xue-bin, et al., "Remote sensing monitoring of aquaculture and automatic information extraction," Remote Sensing for Land & Resources, pp. 41-44, 2009. [18] X. Yu-lin, et al., "An Object-oriented Approach for Extracting Farm Waters within Coastal Belts," Remote Sensing Technology and Application, vol. 24, pp. 68-72, 2009. [19] C. Jianyu, et al., "Simple Object Segmentation Near Coastal Belt s in High Spatial Resolution Remote Sensing Data," Science China, vol. 36, pp. 1044-1051, 2006. [20] Y. Mei-chun, "Extraction of information about water area change in Nanjing City in the past ten years based on TM data," Water Resources Protection, vol. 21, pp. 31-33, 2005. [21] S. Xiaoyu, et al., "Analyses on Spatial-Temporal Changes in Aquaculture Land in Coastal Areas of the Pearl River Estuarine," Resources Science, vol. 32, pp. 71-77, 2010.