Pedosphere 16(4): 477-488, 2006 ISSN 1002-0160/CN 32-1315/P @ 2006 Soil Science Society of China Published by Elsevier Limited and Science Press Spatial-Temporal Pattern and Driving Forces of Land Use Changes in Xiamen*' QUAN Bin', CHEN Jian-Fei2, QIU Hong-Lie3, M. J. M. ROMKENS4, YANG Xiao-Qi5, JIANG Shi-Feng5 and LI Bi-Cheng' 'Research Center of Soil and Water Conservation and Ecological Environment, Chinese Academy of Sciences and Mi- nistry of Education, Northwest Sci-tech University of Agriculture and Forestry, Yangling, Shaanxi, 712100 (China). E-mail: [email protected]. cn 2School of Geographical Sciences, Guangzhou University, Guangzhou 510405 (China) Department of Geography and Urban Analysis, California State University, Los Angeles 90032-8253 (USA) USDA/ARS, National Sedimentation Laboratory, P.O.Box 1157, Oxford, MS 38655 (USA) 5School of Sciences, Jimei University, Xiamen 361021 (China) (Received August 11, 2005; revised February 14, 2006) ABSTRACT Using Landsat TM data of 1988, 1998 and 2001, the dynamic process of the spatial-temporal characteristics of land use changes during 13 years from 1988 to 2001 in the special economic zone of Xiamen, China was analyzed to improve understanding and to find the driving forces of land use change so that sustainable land utilization could be practiced. During the 13 years cropland decreased remarkably by nearly 11 304.95 ha. The areas of rural-urban construction and water body increased by 10152.24 ha and 848.94 ha, respectively. From 1988 to 2001, 52.5% of the lost cropland was converted into rural-urban industrial land. Rapid urbanization contributed to a great change in the rate of cropland land use during these years. Land-reclamation also contributed to a decrease in water body area as well as marine ecological and environmental destruction. In the study area 1) urbanization and industrialization, 2) infrastructure and agricultural intensification, 3) increased affluence of the farming community, and 4) policy factors have driven the land use changes. Possible sustainable land use measures included construction of a land management system, land planning, development of potential land resources, new technology applications, and marine ecological and environmental protection. Key Words: driving force, GIs, land use change, remote sensing, Xiamen INTRODUCTION Land use and land cover are prominent ecological symbols within the surface system of the earth. Land use refers to human manipulation of the land to fulfill a need or want. Meanwhile, land use change may involve either a shift to a different use, such as from rice paddy to aquaculture, or an expansion and intensification of an existing form, such as from subsistence to commercial farming (Matson et al., 1997). Land cover, defined as the physical surface condition of the land, is likely to change as a result of land use change (Turner and Meyer, 1991). Furthermore, land use influences the environment mainly by land cover, and thus land use and land cover are inter-related. Land use/cover change (LUCC) is a core project of the International Global and Biology Plan (IGBP). It aims to improve understanding of the global dynamics of LUCC with a focus to improve the ability to project such change (Turner et al.; 1997). More and more people believe that it is a timely project to comprehensively assess the global environmental changes (Liu and Buhe, 2000a). LUCC studies the changes of natural land, socio-economic conditions, and human activities. Therefore, it requires the cooperation of natural and social sciences to link LUCC to global change (Turner, 1994). LUCC revolves around core problems of regional population, resources, environment, and development. Since the 199Os, the study of LUCC has been a subject of intense interest in academic circles. In recent years, some researchers have made great progress in LUCC studies (Meyer and Turner, 1996; Luo and *lProject supported by the Fujian Provincial Natural Science Foundation of China (No. D0210010). 478 B. QUAN et al. Ni, 2000; Shi et al., 2002). However, few studies have been done to date in the southeastcrn part of Fujian Province, which experienced major economic development during the past 20 years. Currently, the rate of conversion of agricultural land in the southeastern coastal area of China to non-agricultural uses is increasing (Liu et al., 2003). Consequently, there is a need for more research in the southeastern Fujian Province, where rapid development has led to swift changes in land use patterns. In this work, land use spatial changes during 1988, 1998, and 2001 in Xiamen were studied using remote sensing (RS) and geographical information system (GIS) tools. The characteristics and rules of land use changes and their driving forces were analyzed quantitatively through models, which provided a scientific basis for decisions in regional resource and coordinated environmental development, whilst offered a typical case in land use change in one of China’s “hot spots” of economic development. MATERIALS AND METHODS Survey of region Xiamen, with an area of 1638 km2, is located in the southeastern part of Fujian Province, facing the Taiwan Straits. It has a southern subtropical monsoon climate, an annual mean temperature of 20.8 “C,and an annual precipitation of 1143.5 mm. The natural vegetation is a southern-subtropical monsoon rainforest, but human activities have destroyed most of this. Masson Pine (Pinus massoniana Lamb.) and Taiwan Acacia (Acacia confusa Merr.) have been planted in the upland and bottom flat land, under which a lateritic red soil has developed over time (Quan et al., 2004b and 2005a). In 2001, Xiamen consisted of seven administrative districts including Siming, Kaiyuan, Gulangyu, Huli, Jimei, Xinglin, and Tong’an Districts with a total population of 1.31 million. When China began a policy of opening up to the world, Xiamen became one of the first four special economic zones because of its advantageous location. Since then the economy developed rapidly. Data and classification Land use data were obtained from Landsat satellite images from 1988 to 1998 and 2001 with a spatial resolution of 30 m x 30 m. In addition, maps of Xiamen’s vegetation distribution, Xiamen City remote sensing images, Xiamen administrative district (2004), and land use etc. were collected for image-interpretation. Land resource investigation data (1988-2001) were also gathered for consultation. Two grades were established for the land use classification system. The first grade was divided into six classes: cropland (l),orchard (2), forestland (3), rural-urban industrial land (4), water body (5) and unused land (6). The second grade was divided into 12 classes with the following names and codes: paddy field (ll),dry land (12), orchard (21), forest (31), urban, town and separated industrial land (41), rural land (42), salt field (43), reservoir (51), other water bodies (52), coastal beach (53), barren land (61), and other unused land (62). Procedures For land use data and conversions from 1988 to 2001, first Landsat TM images of three different periods were acquired, and then the GCP (ground control points) works module of Canadian PCI software was applied for making geometric corrections. More than forty ground control points were selected as references on a topography map of scale 1:50000. The Gauss-Kruger projection, which belonged to a kind of transverse equal-angle cylindrical projection, was used to correct images with its projection parameter as follows: central longitude 117”, Krassovsky ellipsoid, and false easting 500 km (Chang, 2002; Chen et al., 1998). Color composites were generated displaying bands 5, 4 and 3 as red, green and blue, respectively. An image enhancement was performed to increase the visual distinction between features in order to increase the amount of information that can be visually interpreted from the data. After image enhancement, based on field investigations, image interpretation LAND USE CHANGE IN XIAMEN 479 symbols of different image elements were added. A global positioning system (GPS) receiver was used to collect the coordinates of sample sites. Additionally, the land use map of 1996 was digitized in GIS ArcView software before image-interpretation. This could be consulted in the process of person- machine alternating visual operations. The land use types were interpreted visually in the screen based on the TM images. Also, some additional errors were corrected based on auxiliary reference data and fieldwork. In the end, the smallest plot on the interpreted map corresponded to a scale of 1:100000, and field checking verified the accuracy of image interpretation of up to 90%. To determine the rate of land use change, the study period 1988-2001 was divided into two sub- periods and the land use changes of the two sub-periods were compared. The first sub-period was from 1988 to 1998, called the earlier stage, and the second sub-period was from 1998 to 2001, called the later stage. The comparative analysis on land use change focused on the two sub-periods. Regional difference in land use characteristics were determined using the land use dynamic degree model that could be mathematically expressed by the following relationship (Liu and Buhe, 2OOOb): where S is the land use change rate over time t, Si is the ith type land use area at the beginning of the monitoring period, n is the number of the land use type, and ASi-j is the total area of the ith type land use that is converted into the other types of land use. The land use dynamic degree was thus defined as the time rate change of land use that was converted into the other types of land uses and that at the beginning of the monitoring period was part of the land use area subject to change. The dynamic degree represented, in a comprehensive manner, the change of land use in a given region.