Remote Sensing Assessment Of Ecosystem Health In Fuzhou Based On VOR Model

Under ecological civilization construction,China’s land space planning system has entered a new stage of paying more attention to and integrating natural elements.The basic analysis and evaluation of ecosystem services have become an important factor in supporting the scientific development of spatial land planning.Stable regional ecosystem health is the premise of providing good ecosystem services.The regional ecosystem health assessment research is of great practical significance for land space ecological assessment,regional ecosystem planning,and policy formulation.Research on regional ecosystem health is one of the hotspots in environmental remote sensing.Although remote sensing technology has been used to evaluate regional ecosystem health,it is still necessary to find a suitable index system under the selected model framework for optimization.The existing remote sensing evaluation index system has not reached a consensus,and there are problems such as insufficient comprehensiveness of dominant vitality factors and weak spatial correlation of resilience factors.This study aims to improve the measurement ability of remote sensing technology in the dynamic evolution process and characteristics of regional ecosystem health.Fuzhou is a typical comprehensive ecological landscape embedded in various ecosystems under the background of rapid urbanization in the eastern coastal region.As a coastal provincial capital city,the ecosystem pattern changes significantly under the land planning,construction and economic development,and the selection of the study area has a certain location value.This research was carried out under the traditional Vigor-Organization-Resilience(VOR)model by optimizing and improving the construction method of sub-indexes,coupled with spectral index analysis,landscape theoretical ecology model and spatial measurement.The regional ecosystem health research route of’model construction-key technologies research-advantage comparison-remote sensing evaluation-spatial-temporal differentiation’was followed.Eleven remote sensing datasets were selected(among them,1996,2008 and 2021were used for spatial-temporal differentiation analysis,and other temporal phases were used for difficult problems research)to carry out remote sensing diagnosis of regional ecosystem health in Fuzhou administrative region.The main research findings and conclusions are as follows:Firstly,the construction method of vigor index based on comprehensive remote sensing index method was studied.A new comprehensive vigor index(CVI)was developed by the principal component analysis(PCA)model based on the four indicators:fractional vegetation cover(FVC),global vegetation moisture index(GVMI),vegetation temperature condition index(VTCI),normalized differential build-up and bare soil index(NDBSI).The advantages between CVI and the traditional method were compared.The results showed that the CVI index developed in this study scientifically diagnoses the health vitality of regional ecosystems than traditional ones and has a stronger characterization ability in analyzing spatiotemporal changes.It also comprehensively reflects the dynamic changes caused by seasonal characteristics of vegetation and reduces the preprocessing steps of terrain radiation correction.Furthermore,both vegetation and aquatic ecosystems are more integrated and comparable.The calculation results show that the changing trend of vitality is that the area of poor level(0.0–0.3)expands continuously(about 538.76 km~2),the area of medium level(0.30–0.70)decreases(about 1626.84 km~2),and the area of good level(0.7–1.0)increases(about 1088.08 km~2).Downscaling vitality indicators(taking Sentinel-2A data as an example)was carried out.This study proposes an improved high-resolution urban thermal sharpener(HUTS)algorithm to improve the inversion of land surface temperature(LST)at the scale that is later extrapolated under Sentinel-2A data.Using the improved HUTS method combined with geographically weighted regression(GWR),the LST inversion showed higher accuracy(8.9%)than the traditional HUTS method.Secondly,the construction method of organization index based on landscape index method was studied.There were two difficult problems,the improvement of land cover classification accuracy and the extraction of coastal tidal flats in wetland ecosystems,were studied in the construction of the organization index.A Support Vector Machines(SVM)method for texture combination of multispectral and Synthetic Aperture Radar(SAR)data and a fast extraction method for wetland(coastal beach)were proposed.The finding showed that the classification accuracy of traditional SVM algorithm based on original spectral features was improved by introducing spectral feature index,SAR texture feature and terrain feature.Regarding the problem that coastal tidal flats cannot be easily recognizable,unlike land landforms,the proposed fusion algorithm of SAR data(low sea level)and multispectral data(high sea level)effectively solved the problem of low-quality data(optical multispectral data alone),especially in cloudy and foggy areas.A new organization index was developed based on the landscape index.Four types of indexes,namely landscape heterogeneity(LH),landscape connectivity(LC),the shape characteristics of forest patches(CS)and the connectivity of forest patches(CC),were used as the main factors for calculating the organization index.The new organization index results were 0.8554,0.7872 and 0.7709 in 1996,2008 and 2021,respectively,showing a significant monotone decreasing trend(7.9%over 25 years)in Fuzhou.Thirdly,the construction method of resilience index based on habitat quality method was studied.The traditional calculation method of ecosystem resilience ignores the influence of spatial proximity.To alleviate this weakness,our study introduced the habitat quality method of In VEST model to calculate the resilience index.The constructed resilience index calculation framework considers the influence of the relationship between landscape patches and considers spatial heterogeneity.Consequently,the habitat quality index reflecting ecosystem resilience is closer to the regional ecosystem health characteristics.This study investigated difficult problems in the construction of resilience index and put forward a method to improve the construction of protection range by taking advantage of the landscape terrain gradient method.Using the dependence of landscape components on terrain conditions,the problem of arbitrariness in delineating the accessibility vector range of habitat threat factors is solved.The results indicated that the resilience index values of Fuzhou were 0.677(1996),0.674(2008)and 0.664(2021),showing a gradual downward trend to the present.Fourthly,the results evaluation and spatio-temporal differentiation analysis of regional ecosystem health were carried out.The temporal and spatial characteristics of multi-period ecosystem health in Fuzhou were evaluated and analyzed from two aspects:landscape and land cover type.At the landscape level,the regional ecosystem health value of the whole region is gradually reduced,and the average values for the three years are 1996(0.3521),2008(0.3445)and 2021(0.3345)(dimensionless),respectively.The average reduction rate of 25 years is 0.0007 per year.A significant negative correlation was found between the change in regional ecosystem health index(EHI)and NDBSI in each district and county.At the land cover type,woodland had the highest average ecosystem health value(86.44%).The range of ecosystem health values varied greatly among land cover types such as woodland(0.00–0.70),grassland(0.00–0.70),water(0.00–0.66),wetland(0.00–0.54),cultivated land(0.00–0.26),and construction land(0.00–0.14).The mean values were 0.67(woodland),0.50(wetland),0.47(grassland),0.40(cultivated land),0.39(water),0.25(construction land)and 0.24(unused land).Regarding the trend of ecosystem health,the wetland and farmland showed a monotonic decrease;the total value of water ecosystem health showed a trend of 2008>2021>1996,which was adjusted in each district and county,and further subdivided into three types of change:decreasing type,increasing type,increasing first and then decreasing type.In addition,Moran’s-I of the ecosystem health values in all years were higher than 0.8,indicating that the ecosystem health had a high spatial positive correlation.At the same time,local autocorrelation showed that it had the characteristics of high or low-value agglomeration.The semi-variance distribution map and statistical values of geostatistical analysis illustrated that the base values of the three years are gradually increasing;in which the base values in 2008 were 6.7%higher than those in 1996 and 5.7%less than those in2021,indicating that the variation range of regional ecosystem health is increasing year by year,and the variation range of the former period(1996–2008)was greater than that of the latter(2008–2021).To sum up,taking Fuzhou City as the research area,based on the VOR framework,a systematic study is carried out from the perspectives of the innovation and optimization construction of the regional ecosystem health index,the research on difficult problems,and the spatial-temporal differentiation.Among them,the index construction proposed comprehensive remote sensing index construction,habitat quality and other innovative methods;in the difficult problems(downscaling of vitality index,feature extraction and accuracy improvement of remote sensing interpretation,and construction of accessibility vector range of resilience,etc.),through experimental research,targeted solutions are found.The spatial-temporal differentiation analysis determined the quantitative characteristics and spatial autocorrelation of landscape and type-level.The proposed remote sensing diagnosis method provides a complete case for solving the dynamic evolution process and regional ecosystem health characteristic measurement.