Changes In Land Cover,forest Spatial Patterns And Attributes Under An Intensive Urbanization Scenario

Rapid urbanization has led to tremendous changes in land use types,particularly the urban forest distribution,which highly alters the urban environments including the water cycling and micro-climate,negatively affecting the life quality of urban dwellers.In this study,based on the SPOT images acquired in 2002 and 2016 and the RapidEye images acquired in 2009 and 2017,land cover classifications at Laoshan forest park area of Nanjing were implemented by combining the Support Vector Machines(SVM)algorithm,visual interpretation and on-screen edits.Based on the land cover maps in 2002 and 2009,the CA-Markov model and the state transition model were applied to simulate the land cover patterns in 2016,followed by a spatial agreement evaluation through a pixel-by-pixel comparative analysis between the 2016 simulated land cover map and the 2016 actual land cover map.Further,the better model,CA-Markov determined by the spatial agreement evaluation process,was calibrated by the 2016 actual land cover map and used to project the 2023 land cover pattern based on the land cover maps in 2009 and 2016,Furthermore,a landscape mosaic model was applied to analyze the landscape,particularly the forest landscape changes,followed by the evaluation of the ecological service value of the entire study area.In order to further explore the influence of urbanization on forests,this study analyzed the forest biomass of the study area by using the Zone model developed based on Landsat image in the same period.The results showed that after an intensive validation of the classifications,an overall classification accuracy over 84% was achieved for multi-temporal land cover classifications.And urban land coverage increased from 7.27% in 2002 to 31.95% in 2017,simultaneously forest dropped from 88.36% to 64.60%.The spatial agreement analysis indicated that CA-Markov had an agreement of 75.17%,higher than that of 71.55% for the state transition model.Besides,from the perspective of landscape mosaic model,forest has been heavily encroached by agricultural and urban land uses in recent 2-3 years.According to Nanjing's ecological service value evaluation system,the ecological service value of Laoshan national forest park and its surrounding area was estimated at 42.62 billion yuan in 2002,and at 31.37 billion yuan in 2017.In terms of biomass,the interior young forests gained an increase in biomass while the interior mature forests remained almost stable in biomass accumulation,whereas there was a decrease in biomass over the surrounding areas of the park,which might be attributed to the increased urbanization,encroaching into the forest.The prejected land cover pattern in 2023 provides an important data basis for near future decision-making on land management.The study concludes that high resolution remotely sensed imagery is an effective dataset to quantify land use,particularly urban forest change,and the CA-Markov model is also a reliable tool to predict future land use change,especially for the small areas with high resolution images.The findings from the work deliver valuable insights and implications on urban management,particularly the urban forest management,which are quite essential for developing reasonable land use strategies and urban planning practices.