| Rapid urbanization has resulted in increase on urban population, expansion of construction land area, traffic jams and a series of urban problems, correspondingly, the spatial patterns of urban land use have changed a lot. This paper employed high resolution remote sensing data by means of moving window, transfer matrix, gradient analysis combined with markov chain supported by ArcGIS and FRAGSTATS as well as Visual Basic platform, we quantitatively analyzed the change of urban land use spatial pattern from 1996 to 2009 in Jinan City according to landscape indicators, and then made suitability assessment on land use of Jinan City in 1996, We made the urban land use suitability evaluation results as the initial input parameters of the CA model for simulating the land use change for Jinan City in 2004, then we did repeated correction and test on the parameters of the model using the status classification data of land use in 2004, finally we used cellular automata model to simulate and predict the land use change of Jinan City in 2025 integrated with moving window.The results revealed that urban construction land area was increased by 67.57 km2, agriculture land and green space were reduced by 66.55 km2 in Jinan City during the research phase from 1996 to 2009, the most of reduced agriculture land and green space were replaced by urban built-up land, the substitute rates were 14.23%and 20.13% during two research periods from 1996 to 2004 and from 2004 to 2009 separately; in which the transfer rate from agricultural land to residential land was increased from 4.10% to 14.20% during 1996-2004 and 2004-2009. During the study period, the number of patches, Shannon Diversity Index and patch density metric value were increased from 5921 to 6361, from 1.74 to 1.83, from 11.00 number·km-2 to 11.82 number·km-2 respectively on the landscape level in the study area; on the type level the number of patches, patch density, largest patch index, landscape shape index, the percentage of landscape index of residential land and industrial land as well other building land classification were increased gradually, however, the values of the corresponding landscape metrics of agriculture land and green space as well as other land use types were decreased by degrees. The metric raster maps of landscape shape index and Shannon Diversity Index obtained by moving window showed that the fragmentation and diversity index values of the overall landscape on the landscape level were increasing in the study area; the percentage of landscape values of agriculture land and residential land use on type level were distributed in an opposite status in space. The results of gradient analysis displayed that the maxima of the landscape shape index and Shannon Diversity Index on the landscape level were mainly distributed in the neighbor of the 4km away from urban center and the direction of urban northeast and southwest of 10km away from urban center, which were subject to much human disturbance and key regions of urban planning and development; on the type level the maximum of the percentage of residential land was principally scattered in the range of 6km away from city center, while the max of the corresponding metric was chiefly located near the southeast area 4km away from city center, the northeast and the west as well as the east approximately 14km distance to urban center and northwest region of 10km to the center.We made a preliminary forecast to the change of the future urban land use for Jinan City using the markov chain method, the result demonstrated that the mutual transform into and out of different land use classification will progressively reach a stable state in the next few decades, the urban development will also become more sustainable. At the same time, we took topography, vegetation, waters, accessibility, the percentage of construction land and so on as suitability evaluation factor to assess the suitability of urban land use for Jinan City in 1996 by applying Analytical Hierarchical Analysis (AHP) combined with Delphi method. We made the urban land use suitability evaluation results as the basis and the initial input parameters of the CA model, then we embedded the landscape indicator obtained by moving window as a controlling factor for formulating transition rules in the CA model to other natural and social controlling and limiting factors, which was used to simulate the change of urban land use for Jinan City in 2004, the simulation accuracy reached to 91%. Finally, we adopted the corrected CA model to simulate and predict urban land use in 2025 for Jinan City, the simulation results showed that there was an increasing trend in the construction land of Jinan City, the proportion of other land converted into construction land was larger, the expansion of urban spatial pattern was mainly along the direction from the old districts of urban center to urban east-west areas, which appeared the development of "group" type, and also spread and sprawled to the agriculture land in the south of city and urban suburbs.The moving window method can better mine the process of urban land use change, and connect with spatial pattern change more accurately. The space development of urban construction land was uneven, and urban development took on irregular concentric structure, which was closely related to policy-oriented urban development, the amenity of urban green space and special geographical morphology. The combination of moving window and cellular automata enriched the dynamic model for urban spatial expansion and change for making it easier to calibrate and set the parameters of the model and improving the model simulation accuracy, which also drived the more in-depth exploration and analysis to the dynamic evolution discipline and characteristic of urban spatial sprawl in Jinan City. The simulation results will be favorable for instructing the planning on urban land use and providing management and conservation of urban natural resources.
|
PDF Full Text Request