| With the rapid development of economic and urbanization,more and more human activities have brought greater challenges to the ecological environment of different scales around the globe.Efficient and accurate monitoring of ecological conditions is of positive significance for improving the regional ecological environment,formulating environmental protection policies and achieving sustainable development goals.The proposal of Remote Sensing based Ecological Index(RSEI)and large-scale application of RSEI showed that ecological indicators extracted from remote sensing images have great potential in characterizing the quality of regional ecological environment.However,RSEI still suffers from some uncertainties in its application in multiple scenarios,since it is random in the direction of eigenvector and not limited by spatial scale,as well the influence of regional ecological environment has not been considered.In addition,due to the lack of sufficient cloud-free and clear time-series remote sensing data,RSEI is intermittently discontinuous in time,and susceptible to the interference of climatic and seasonal variations therefore results in "pseudo-variations",which cannot accurately reveal the continuous evolution of regional features.Therefore,in this paper,based on Google Earth Engine platform,we have reconstructed the Landsat time series dataset from 2000 to 2020 through Continuous Change Detection and Classification(CCDC)surface reflectance algorithm.At the same time,the existing RESI model has been optimized,a method of calculating RSEI based on grid traversal was proposed by combining the idea of scale in landscape ecology.The experimental results show that the stability in the RSEI results has been enhanced by improving the data dimensionless processing method,the principal component analysis model has been optimized to solve the problem of randomness in the direction of eigenvector during the process of principal component analysis.Combined with the reconstructed Landsat dataset,the improved RSEI has provided an effective method to evaluate long time series ecological environmental quality changes in a large scale,batch and accurate manner.In this paper we took Changting County,Fujian Province as the study area,and have used Landsat images with cloud coverage below 80% in the Landsat Collection 2 dataset for time series image reconstruction,and analyzed the ecological and environmental conditions of the study area from 2000 to 2020 by using the optimized RSEI,and the results show that:(1)A high correlation between the original images and reconstructed Landsat time series image dataset of Changting through CCDC surface reflectance model has been shown.As well as a low root mean square error(RMSE)of each band.The NDVI time series fitting results derived from the reconstructed data were close to the real observations.The reconstructed images can clearly show the surface differences and spatial information;the change process of the surface coverage also maintained a high consistency with the original images.The dense,clear and cloud-free reconstructed images can be used in the study of long time series dynamic change monitoring of ecological environment quality,thus improving the accuracy and continuity of ecological environment monitoring.(2)In the process of RSEI inversion,0-mean standardization has been used for dimensionless processing of ecological evaluation indexes instead of linear function normalization,has shown a better stability and anti-interference ability,which can exclude the influence of image outliers.By improving the principal component analysis algorithm in RSEI,the problem of heteroscedasticity of RSEI batch calculation results has been solved,which meets the demand for large-scale and high-volume monitoring of ecological environment quality under the background trend of remote sensing big data.The scale effect has been considered during the grid-based RSEI calculation therefore the RSEI calculation window has been limited to be more consistent with the first law of geography.The optimized RSEI is more sensitive to surface changes,and the ecological monitoring results in the intersection area of different surface covers are more in line with the actual situation.(3)During 2000-2020,the overall ecological environment quality of Changting showed a trend of "slow rise-brief decline-slow rise-repeated fluctuations".The average value of RSEI in the whole Changting area rose from 0.572(January 3,2000)to 0.627(May 16,2008),and then experienced a continuous decline for 21 months to 0.611(February 15,2010),after which the average value of RSEI showed a continuous upward trend and reached 0.654(January 7,2019),fluctuated between 0.63 and 0.65 until 0.647(December 27,2020).The ecological quality of Changting County has generally improved during the last 20 years.Among them,the improved area accounts for about 27.53% of the total area,mainly occurring in Hetian town,Zewu town,Sanzhou town and Maundian town.In particular,the ecological environment quality of Hetian town was the best,followed by Maundian town and Cebu town,and the ecological environment quality of Sanzhou town was relatively poor.However,the average value of RSEI in Sanzhou Town and Hetian Town has increased the most,and the ecological environment quality has improved significantly.The area of deteriorating ecological environment quality caused by urbanization expansion and road network construction is about 27.36%,mainly concentrated in Changting County,Tingzhou Town,Datong Town and Cewu Town,while other townships are scattered.Within the period of 2000-2020,the ecological environment condition of Changting has been improved,and ecological restoration and soil erosion control were effective.However,ecological environment deterioration still exists in a few areas of Changting,and it is still necessary to continue and strengthen the ecological protection work. |
PDF Full Text Request