Font Size: a A A

Assessing Land Cover And Ecological Quality Changes Under The Scenario Of New-type Urbanization From Multi-source And Multi-temporal Remote Sensing Observations (2009-2019)

Posted on:2023-05-17Degree:MasterType:Thesis
Country:ChinaCandidate:F ShiFull Text:PDF
GTID:2531306824992759Subject:Forest management
Abstract/Summary:PDF Full Text Request
The ecological environment quality is influenced by urban land cover change and its surroundings,which is highly valued by decision-makers.The core district of Nanjing Jiangbei new area has experienced the development processes of traditional urbanization through new-type urbanization(NTU).It is significant to study how this process impact the changes of ecological quality.Based on four kinds of high spatial resolution remote sensing images(Rapid Eye,GF-1WFV4,GF-2 PMS1 and Sentinel-2A)and field survey data in 2009,2015 and 2019,the object-oriented decision tree classification method and an improved Mask-RCNN deep learning algorithm were applied to classify land cover types and analyze their changes.Then,based on Landsat images in 2009,2015 and 2019,the applicability of Remote Sensing Ecological Index(RSEI)model and Regional Optimized Remote Sensing Ecological Index(RO-RSEI)model in the evaluation of ecological quality was compared,followed by a spatial auto-correlation analysis to characterize the spatio-temporal changes of ecological quality in the study area.Next,the intrinsic causality between land cover change and ecological quality change was discussed,and the attribution analysis of the observed changes in land cover and ecological quality was carried out by examining the socio-economic driving factors.Ultimately,the targeted planning suggestions for the further construction of NTU in the study area was recommended.The proposed method in the current work provides a technical reference for monitoring ecological quality under the development of NTU.The main research results were as follows:Compared with the object-oriented decision tree classification method,the improved Mask-RCNN deep learning classification method achieved higher accuracy in land cover classification of multi-source high-resolution images in the study area.The overall accuracy of land cover data sets generated by the improved Mask-RCNN model in 2009,2015 and 2019 was at 90.75%,91.75%and 92.50%respectively.Compared with RSEI model,RO-RSEI model was more suitable for monitoring ecological quality change in the study area.The mean value of RO-RSEI in 2009,2015and 2019 was at 0.652,0.649 and 0.655 respectively.The Moran’s I values of RO-RSEI in 2009,2015 and 2019 were 0.442,0.312 and 0.411 respectively.The average value of RO-RSEI showed that the ecological quality of the study area experienced a process from deterioration to improvement in the past 10 years,and the overall ecological quality remained stable during the 11years;the local spatial auto-correlation analysis result of RO-RSEI showed that the clustered distribution trend of ecological quality was affected by policies,the hot spot distribution was mainly affected by the fluctuation of ecological quality in forest land,and the change of cold spot distribution was consistent with the development direction of urban construction.Thus,there was a strong spatial agreement between the spatio-temporal changes of ecological quality and land cover.From 2009 to 2019,the urban land use increased by 35.07km~2,the annual change rate(ACR)of land cover decreased by 6.33%,and the density of urban land increased in the construction center area.The bareland first increased under the demand of urban construction,giving way to a decrease with the completion of urban construction.The deterioration of the ecological environment quality caused by the increase of unused land was greater than the increase of urban land.The grassland increased from 16.65km~2 to 33.39km~2 in the past 11 years,mainly encroaching into urban green spaces.The ecological quality of the urban planning and construction center in the study area was greatly improved with the completion of construction and the restoration of urban greening.The cultivated land decreased dramatically by 48.10km~2 in the early stage and became stable in the later stage.This was because the government strictly protected the red line of the cultivated land in order to ensure food security.In the past 11 years,the forest land remained stable,but the internal ecological quality of Laoshan Forest Farm first deteriorated slightly(2009-2015)and then gradually improved(2015-2019).The various promulgated protection and management regulations enhanced the ecological protection and construction of the forest land and promoted the rational development and utilization of forest land resources.The water area decreased from 14.52km~2 to 7.93km~2.In the past 11 years,the main urban area was intensively developed,the urban green space restored and increased,the allocation of unused land was reasonable,the protection of forest land and cultivated land was strengthened,and the ecological environment had been improved.In general,the ecological quality of the study area changed from deterioration to improvement and then remained stable during the study period,which was due to the strict implementation of ecological protection and restoration policies by the government and the improved awareness of natural resources protection and active participation of the public of in the process of NTU construction.In order to promote the NTU construction,strengthening the implementation of policies such as industrial adjustment,spatial planning,green construction,modern comprehensive transportation system and legislative co-construction in terms of social economy is recommended;for natural resources,strengthening the protection of wetland and forest,strictly abiding by the ecological safety line,building ecological isolation zones,rationally developing and constructing the natural resources,and strengthening spatial monitoring of ecological security patterns are suggested.
Keywords/Search Tags:Object-oriented decision tree classification, modified Mask-RCNN model, RO-RSEI, Spatial auto-correlation analysis, New-type urbanization
PDF Full Text Request
Related items