Spatio-temporal separating analysis of NDVI evolution and driving factors: a case study in Nanchang, China

Spatio-temporal separating analysis of NDVI evolution and driving factors: a case study in Nanchang, China | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Spatio-temporal separating analysis of NDVI evolution and driving factors: a case study in Nanchang, China Jiatong Li, Hua Wu, Qiyun Guo, Yue Xu, Huishan Li, Sihang Liu, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5366943/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Investigating vegetation coverage and quantitatively evaluating environmental changes can serve as the science knowledge in ecological protection, resource management, and policy-making, promoting harmonious coexistence between human and nature. In this study, we had explored the separation in space and time of evolutionary characteristics and driving factors of NDVI in Nanchang City from 2000 to 2022 based on Hurst Exponent, ReliefF feature selection algorithm, Geographical detector and so on. The results are: (1) From temporal dimension, the average NDVI in Nanchang City was 0.453, showing an overall upward trend. Although the growth rate gradually slowed over time. (2) In terms of spatial changes, vegetation in Nanchang City overall exhibited a characteristic of reverse sustained development, showing trends of "improvement around rivers and lakes" and "large-scale degradation of urban land." (3) The ReliefF proved to be more suitable among the three algorithms in the temporal dimension-driven analysis. Human factors are the dominant factors significantly influencing the changes in NDVI, while meteorological factors are not as significant. (4) The driver-analysis of geographical dector shows that population density, nighttime lights, and land cover types emerged as significant driving factors. Regions where NDVI and human factors have negative correlation are primarily centred in the heart of Nanchang City and Jinxian County; while the positive correlations are found around rivers and lakes. This study delves into the changing patterns of vegetation cover in Nanchang City, providing scientific guidance for the protection and regulation the regional ecological environment to bring about a sustainable development. climate change driving analysis NDVI temporal dimension spatial dimension Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 1. Introduction Vegetation has vital relevance in climate regulation, material cycling, information transmission and other aspects, which is dominating component in terrestical ecosystems (Terefe et al. 2017 ; Jin et al. 2021 ). The vegetation coverage in secular change reflexes the eco-environment not only the alterations, but also the influences in natural factors(such as weather) or human activities indirectly (Zhao et al. 2024 ). Therefore, monitoring vegetation dynamics which is in domain research has become one of the hot topics. Normalized Difference Vegetation Index(NDVI) is a key variable for monitoring climate change, studying ecological balance, and exploring regional phenological patterns. It has been widely used for estimating vegetation productivity and assessing ecological vulnerability (Wang et al. 2024; Das et al. 2024). Vegetation coverage monitor is crucial for ecological recover and governance. Many scholars have conducted extensive researches in spatio-temporal vegetation variations and its driving factors on the study area using different scales of NDVI. Previous scholars have used methods such as Sen + Mann-Kendall trend analysis, residual analysis, correlation analysis or partial correlation analysis, Hurst index, geographical detector, etc., to study the regional changes and driving factors of NDVI in different regions globally, including Africa (Yang et al. 2022), Europe (Novillo et al. 2019 ), China (Xu et al. 2023 ), Loess Plateau (Li et al. 2021; Kong et al. 2024 ), Qinghai-Tibet Plateau (Wang et al. 2023 ), Yangtze River Delta (Tian et al. 2024 ) and so on. Inter-annual vegetation prediction is discussed by Theil-Sen Median and MK-tendency analysis in the Yellow River Basin and acquiring preferably results (Cao et al. 2014 ). Yang et al.(2019) investigated the evolution trend of vegetation, revealing the global NDVI trend from 1982 to 2015. The Hurst index can detect whether there is super-long periodicity in long time sequences and can simulate future vegetation growth trends (Ahmad et al. 2023 ; Tong et al. 2018 ). For example, Qu et al.(2020) forecasted the future vegetation growth trends by Hearst index over the Yangtze River Basin, and Zhang et al.(2024) calculated the future growth trends of vegetation in Chengdu-Chongqing region with the Hurst exponent of high-dimensional fractals. However, so far, it has not been used for vegetation trend prediction in Nanchang City. Research on the driving forces of surface vegetation, it can be roughly divided into meteorological factors, topographic factors, human factors and other factors. Numerous scholars have used a range of analytical tools to explore the causes of vegetation change, including correlation analysis (Zhang et al. 2017 ), linear and nonlinear regression models (Li et al. 2024 ), and geographical detector (Zhang et al. 2023 ). And the same applies to machine learning in the screening of drivers.Teixeira et al.(2013) applied the RF algorithm to the application of predicting the standard enthalpy of formation of hydrocarbons by downscaling the high-dimensional feature data selectively; Tang et al.(2018) utilized MIC algorithm for linear and nonlinear correlation acquisition for bearing fault diagnosis in feature selection algorithm; Zhang et al.(2016) performed electroencephalogram sensor sentiment recognition based on ReliefF feature selection algorithm, which quantified the importance of each sensor channel by obtaining all the feature weights of the analyzed channels through the algorithm. Previous studies had often failed to adequately consider the role of soil temperature when considering the effects of meteorological factors on vegetation growth (Luo et al. 2016 ), and had also neglected the potential effects in economic level of population (Zhang et al. 2023 ), among others, on NDVI.Most of these studies used qualitative methods to analyze spatial and temporal changes together (Zhong et al. 2018; Sun et al. 2019), and lack in-depth exploration of how much the climate change and human activities quantitatively affect vegetation changes.In addition, there was a relative lack of research on the sustained development and future vegetation growth trend in Nanchang City. In manuscript, we temporally and spatially divide up the NDVI evolution and driver analysis of Nanchang City from 2000 to 2022 systematically. Hurst Exponent, Sen + Mann-Kendall trend analysis and other methods are used to quest its evolutionary characteristics and trends of NDVI; ReliefF, MIC, RF algorithms are used for the selection of meteorological driving factors of NDVI, and geographical detector and complex correlation analysis can quantitatively study the impact of anthropogenic and meteorological factors to NDVI. The central objective of this study is to divide the analysis of vegetation evolution and its drivers into two independent dimensions, temporal and spatial, and to explore each dimension in detail.In the temporal dimension, we focus on the dynamic changes of NDVI over time, and reveal its long-term evolution patterns through trend analysis. In the spatial dimension, we examine vegetation conditions in different geographic locations to explore the effects of spatial heterogeneity on vegetation distribution and pattern. Through this detailed spatial and temporal segmentation method, we can more accurately grasp the pattern of vegetation change, and more precisely identify the key factors affecting the evolution of vegetation. It can bring a scientific foundation in ambulatory monitoring and vegetation protecting in Nanchang City, as well as theoretical guidance for the formulation and implementation of sustainable advancement plans. 2. Materials and Methods 2.1.Study area Nanchang, locating at north-central of Jiangxi Province, East China (115°27'E-116°35'E, 28°09'N-29°11'N), is the core city at Poyang Lake ecological economic zone (Zhu et al. 2023), with be in Ganjiang River, Fuhe River downstream (Fig. 1 ). In the mean time, it is the only provincial capital designated with low-carbon economic development pilot city, which is important comprehensive demonstration area for ecological cities (Liu et al. 2023 ). Nanchang City is a subtropical monsoon climate zone, characterized by both rain and heat in the same season, with four distinct seasons and abundant rainfall, making it one of the rainy areas in China (Zhu et al. 2023). There are notable differences in the distribution of rainfall, which can easily lead to droughts or flood disasters. Nanchang has abundant resources which is about ecological, including a part of Poyang Lake. For the past few years,the number of occurrences in extreme weather events continue to increase because of the ecological destruction (Zhang et al. 2022 ), exacerbating the impact on NDVI evolution. In the preliminary statistics from the Nanchang City Disaster Reduction Committee, the 3-hour rainfall of the two extremely heavy rainstorms in 2012 exceeded 100mm, more than the overall rainfall of the August in previous years. 298.5 hectares of crops area had been affected, with pecuniary losses exceeding tens of millions (Jiangxi Bureau of Statistics 2021). 2.2.Data sources The research data mainly contain MODIS13 NDVI, meteorological, terrain, land cover type, and anthropogenic data (Table 1 ). ArcGIS 10.8 was used to perform spatial transformating, masking and clipping, grid creation, and multi-value extraction to points for modeling and detecting each year's data; when performing data resampling, we standardized the resolution to 1km; Matlab R2018b was used to standardize and normalize modeling of long time series data, and to perform machine learning model-driven analysis. Table 1 Data source and data preprocessing. Data Type Code Data Content Time resolution Spatial resolution .Data Source Product Remote sensing data Y MODIS NDVI year 1km/500m NASA website( https://search.earthdata.nasa.gov/search ) Didan, K.(2015) Meteorological data X1 Annual mean temperature year 1km National Tibetan Plateau / Third Pole Environment Data Center.( https://data.tpdc.ac.cn/zh-hans/data/ ) Peng, S.(2019) X2 Total annual precipitation year 1km Third Pole Environment Data Center. ( https://data.tpdc.ac.cn/ ) Peng, S.(2020) X3 Evaporation and transpiration year 1km NASAwebsite( https://ladsweb.modaps.eosdis.nasa.gov/ ) MODIS16A2 A1-A24 Meteorological data from weather stations month China Meteorological Data Service Centre( https://data.cma.cn/ ) Data from the National Climatic Observatory, Nanchang Anthropogenic data X4 Real Gross Domestic Product(GDP) year 1km Figshare website ( https://doi.org/10.6084/m9.figshare.17004523.v1 )(2000-2015)/GitHub website( https://github.com/thestarlab/ChinaGDP)(2016-2020 ) Zhao, Liu.(2017) X5 Population density year 100m WorldPop website( https://www.worldpop.org/ ) X6 Nighttime lights year 1km HARVARD Dataverse ( https://dataverse.harvard.edu/ ) Wu, Yizhen.(2022) A25-A37 Urban statistics year National Bureau of Statistics of China ( https://www.stats.gov.cn/ ) Topographic data X7 elevation year 30m Geospatial Data Cloud( https://www.gscloud.cn/ ) ASTER GDEM 30M Land cover data X8 Land cover type year 30m Resource and Environmental Science Data Platform ( https://www.resdc.cn/ ) Xinliang Xu,(2018) Note: average temperature (A1), average maximum temperature (A2), average minimum temperature (A3), extreme maximum temperature (A4), extreme minimum temperature (A5), precipitation (A6), average pressure (A7), average relative humidity (A8), sunshine hours (A9), maximum wind speed and direction (A10), average surface temperature (A11), average maximum surface temperature (A12), average minimum surface temperature (A13), extreme maximum ground temperature (A14), extreme minimum ground temperature (A15), average ground temperature at 5 cm depth (A16), average ground temperature at 10 cm depth (A17), average ground temperature at 15cm depth (A18), average ground temperature at 20cm depth (A19), average ground temperature at 40cm depth (A20), average ground temperature at 0.8m depth (A21), average ground temperature at 1.6m depth (A22), average ground temperature at 3.2m depth (A23), E601 large-scale evaporation (A24), GDP (A25), value added of primary industry (A26), value added of secondary industry (A27), value added of tertiary industry (A28), value added of the primary industry as a proportion of GDP (A29), value added of secondary industry as a proportion of GDP (A30), value added of tertiary industry as a proportion of GDP (A31), employee number in the primary industry (A32), employee number in the secondary industry (A33), employee number in the tertiary industry(A34),employee proportion in the primary industry (A35), employee proportion in the secondary industry (A36), employee proportion in the tertiary industry (A37). 2.3.Research methods 2.3.1.Theil-Sen Median and Mann-Kendall trend analysis Theil-Sen belongs to non parametric calculation methods (Wang et al. 2009 ), Which is efficient, robust and apposite for trend analysis in long time series. Time series do not need to satisfy assumptions such as autocorrelation or normal distribution. It can effectively handle outliers and missing values. The Theil-Sen slope represents the positive or negative relationship between data points in a time series, with median slope representing the whole timely trendency (Adichie et al. 1967). The computing formula is: $$\:{\beta\:}=\text{M}\text{e}\text{d}\text{i}\text{a}\text{n}\left(\frac{{\text{x}}_{\text{j}}-{\text{x}}_{\text{i}}}{\text{j}-\text{i}}\right)\:\:\:\:\forall\:\text{j}>\text{i}$$ 1 Above it β represents the slope for determining the trend of NDVI changes in Nanchang City; Median represents the median function; j and i are time indices; \(\:{x}_{j}\) and \(\:{x}_{i}\) stand for the actual NDVI for the j th and i th years, respectively. At the time β > 0 , NDVI is upwarding in the period; at the time β = 0 , NDVI remains constant in the period; when β < 0 , NDVI shows a decreasing trend during this time period. Mann-Kendall analysis is non-parametric which the World Meteorological Organization recommended and be widely adopted, and can effectively discriminate natural process is natural fluctuation or definite changing trend (Gadedjisso et al. 2021). There is no need that the test samples have to meet a certain distribution pattern rules, while outliers will not interfere with the test results (Khaled et al. 2008). Calculate the Theil-Sen slope and then use MK analysis to check the changing trend. When the significance level is α , if the normal distribution statistic(|Z C |) satisfies |Z C |>Z 1-α/2 , the time series change is statistically significant. When the |Z C |≥1.28, it passes the 90% confidence level significance test; when the statistic satisfies |Z C |≥1.64, it passes 95% confidence interval significance test (Table 2 ). Table 2 Mann-Kendall test method significance statistics Significance test Range of Z C value Change trend |Z C | ≥1.64 1.64 Significant increase 1.64≥ |Z C | ≥1.28 −1.64～−1.28 Not significant reduction 1.28～1.64 Not significant increase |Z C | ≤1.28 −1.28～1.28 Unchanged 2.3.2.Hurst Exponent The study of the Hurst exponent(H) based on rescaled range analysis(R/S) was proposed by the British hydrologist Hurst (Tabares et al. 2021), which can be a index that judge time series data is random walk or biased random walk process. The Hurst exponent can describe the persistence or anti-persistence of NDVI changes (Zhang et al. 2023 ). Therefore, this study calculates the Hurst in R/S analysis method to forecast the vegetation NDVI evolution trend in Nanchang City. Time series ( t ), t = 1,2,……, where N DVI ( t ) = B( t )-B( t-1 ), B( t ) being the observed value at moment t , for any positive integer τ greater than or equal to 1 (Chen et al. 2024 ), define the sequence of mean values as $$\:{\text{N}}_{\text{D}\text{V}\text{I}\tau\:}=\frac{1}{\tau\:}\sum\:_{t=1}^{\tau\:}{\text{N}}_{\text{D}\text{V}\text{I}}\left(t\right),\tau\:=\text{1,2},\dots\:\dots\:$$ 2 Cumulative deviation series X( t , \(\:\tau\:\) ) $$\:\sum\:_{t=1}^{\tau\:}\left(\xi\:\right(t)-{\xi\:}_{\tau\:}),1\le\:\text{t}\le\:\tau\:$$ 3 The polar sequence R( \(\:\tau\:\) ) $$\:\text{R}\left(\tau\:\right)=\underset{1\le\:\text{t}\le\:\tau\:}{\text{max}}X(t,\tau\:)-\underset{1\le\:\text{t}\le\:\tau\:}{\text{min}}X(t,\tau\:),\tau\:=\text{1,2},\dots\:\dots\:$$ 4 Standard deviation sequence S( \(\:\tau\:\) ) $$\:\text{S}\left(\tau\:\right)={\left[\frac{1}{\tau\:}\sum\:_{\text{t}-1}^{\tau\:}{\left({\text{N}}_{\text{D}\text{V}\text{I}}\right(\text{t})-{\text{N}}_{\text{D}\text{V}\text{I}\tau\:})}^{2}\right]}^{\frac{1}{2}},\tau\:=\text{1,2},\dots\:\dots\:$$ 5 Hurst exponent calculation: $$\:\frac{\text{R}\left(\tau\:\right)}{\text{S}\left(\tau\:\right)}≌\frac{R}{S}\propto\:\tau\:H$$ 6 If Eq. 6 holds, it implies the existence of the Hurst phenomenon with time series N DVI ( t ),t = 1,2, \(\:\dots\:\dots\:\) . The Hurst shows as H with range between 0–1. At that time 0 < H < 0.5, it is inverse sustainability that future trend of N DVI is; when H = 0.5, it shows that there is no correlation between future and past trends in the N DVI time series; when 0.5 < H < 1, this states the future trend will keep pace with the past (Alvarez et al. 2008). 2.3.3.Machine learning feature selection algorithms (1)The ReliefF algorithm The ReliefF algorithm was proposed as an extension of the original Relief algorithm. Based on traditional filter-based feature selection method, it can reinforce learning algorithm by selecting the feature subset with the highest information content from the original feature set (Ryan et al. 2018). The ReliefF algorithm assigns different weights to driving factors by calculating the discriminative ability of features for each nearby sample. Due to its efficiency and effectiveness, it has been widely used in the field of feature selection. However, some of its variants and improved algorithms also focus on dealing with specific issues, such as handling multi-label data or applications on high-dimensional small sample data (Spolaôr et al. 2014). (2)The MIC algorithm The MIC algorithm can estimate the correlation between two variables and identify linear and it can non-linear relationships (Li et al. 2022). The MIC algorithm can not only quantify the strength of the correlation between variables, but also identify patterns of correlation between variables. This algorithm provides a normalization method that allows for quantifiable comparison of results and it is insensitive to the discretization of data (Li et al. 2024 ). When dealing with large datasets, the MIC algorithm can provide similar coefficients for different types of relationships with similar noise level, selecting the driving factor with the highest degree of correlation with the target variable (Ge et al. 2016 ). (3)The RF algorithm RF model integrates decision trees with Bagging algorithm for improving the generalization ability, with in a common use at machine learning (Breiman et al. 2001). The random forest algorithm which is easily come ture, as an ensemble classifier, performs well in regression, classification, and other problems (Iratxe et al. 2021). 2.3.4.Geographical detector Geographical detector is a primary tool in spatial stratified heterogeneity analysis. The heterogeneity of a single variable and the distribution coupling of two variables can be gauged by geographical detector in the meantime in case to find out the possible relationships and explore numerical and qualitative data between them (Song et al. 2020 ). (1)Factor detector As one of the geographical detector Model, critical analysis in factor detector focuses on: at a specific spatial location, the geographic entities and spatial variability in environmental factors are always exist (Wu et al. 2016 ). In spatial, if the alters in geographic entities at a certain point are consistent with the changes in environmental factors, this environmental factor is prominent in the alteration of geographic entities. Q-value is an indicator to measure the spatial variability of Y or the degree of the independent variable X on the attribute Y , shown as: $$\:q=1-\frac{\sum\:_{ℎ=1}^{L}{N}_{ℎ}{\sigma\:}_{ℎ}^{2}}{N{\sigma\:}^{2}}=1-\frac{SSW}{SST}$$ 7 $$\:SSW=\sum\:_{ℎ=1}^{L}{N}_{ℎ}{\sigma\:}_{ℎ}^{2},SST=N{\sigma\:}^{2}$$ 8 where h = 1, ……, L is the lamination of the variable Y or factor X ; N h and N are the number of cells in layer h and the whole area; \(\:{{\sigma\:}}_{\text{ℎ}}^{2}\) and \(\:{\sigma\:}^{2}\) are the variances of the Y -values for layer h and the whole region, respectively. SSW and SST are the within-stratum sum of squares and the district-wide total sum of squares, respectively (Wang et al. 2018 ). (2)Interaction detector The general method for interaction detector is adding up the product of regression model and statistical significance (Ren et al. 2014). We can be sure if it exist an interaction effect between factors, even the direction, strength, linearity, with analyzing q-values of each factor and multiple factors (Li et al. 2023). 2.3.5.Correlation analysis (1)Single correlation analysis The one-way correlation analysis in image measurement scale NDVI was performed separately with the factors that were significantly correlated to the NDVI as calculated by the geoprobe with the following formula: $$\:{R}_{xy}=\frac{\sum\:_{i=1}^{n}\left[\left({x}_{i}-\stackrel{-}{x}\right)\left({y}_{i}-\stackrel{-}{y}\right)\right]}{\sqrt{\sum\:_{i}^{n}{\left({x}_{i}-\stackrel{-}{x}\right)}^{2}}\sqrt{\sum\:_{i=1}^{n}{\left({y}_{i}-\stackrel{-}{y}\right)}^{2}}}$$ 9 where \(\:{R}_{xy}\) denotes the correlation coefficient of x and y ; x i denotes NDVI value in i year; y i denotes the value of the driver in i year; \(\:\stackrel{-}{x}\:\) and \(\:\:\stackrel{-}{y}\) denote the mean value of NDVI and driver in n years, respectively. (2)Partial correlation analysis Introducing partial correlation coefficient to represent the correlation between two factors, studying the impact of driving factors on NDVI while excluding interference from other factors. Expression is following (Dror et al. 2014 ): $$\:{R}_{xy,z}=\frac{{R}_{xy}-{R}_{xz}{R}_{yz}}{\sqrt{(1-{R}_{xz}^{2})(1-{R}_{yz}^{2})}}$$ 10 In the formula, \(\:{R}_{xy,z}\) stands for the partial correlation between NDVI and factor 1 when factor 2 is held constant, excluding the influence of factor 2 in analyzing the correlation between NDVI and factor 1; \(\:{R}_{xy}\) , \(\:{R}_{xz}\) and \(\:{R}_{yz}\) each represent the correlation coefficients between NDVI and factor 1, NDVI and factor 2 ,and factor 1 and factor 2. This method can be tested the significance of partial correlation coefficients. Expression is following (Lakens et al. 2017): $$\:\text{t}=\frac{{R}_{xy,z}}{\sqrt{1-{R}_{xy,z}^{2}}}\sqrt{\text{n}-\text{m}-1}$$ 11 In the formula: \(\:{R}_{xy,z}\) represents the partial correlation coefficient; n represents the sample size; m represents the degrees of freedom. (3)Multiple Correlation analysis Taking into account the compositive effects in multiple factors, the introduction in multiple correlation is used to calculate the correlation extent between multiple factors (Qiu et al. 2014 ). Expression is following: $$\:{R}_{x,yz}=\sqrt{1-(1-{R}_{xy}^{2})(1-{R}_{xz,y}^{2})}$$ 12 In the formula: \(\:{R}_{x,yz}\) is the multiple correlation analysis between NDVI and factors 1,2; \(\:{R}_{xy}\) is the correlation between NDVI and factor 1; \(\:{R}_{xz,y}\) is the partial correlation between NDVI and factor 2 when factor 1 is held constant. Here, the F-tests method has been inspected the significance of partial correlation coefficient (Quirk et al. 2020). Expression is following: $$\:\text{F}=\frac{{R}_{xy,z}^{2}}{1-{R}_{x,yz}^{2}}\times\:\frac{\text{n}-\text{k}-1}{\text{k}}$$ 13 In the formula: \(\:{R}_{x,yz}\) represents the multiple correlation analysis; n represents the sample size; k represents the number of independent variables. (4)Pearson’s correlation coefficient The degree of linear relationship between two variables can be assessed by it (Kirch et al. 2008). The range of it is -1 to 1,-1 is in perfect negative correlation, while − 1 is contrariwise, and 0 is irrelevant. 3. Results 3.1.Spatial-temporal evolution characteristics of NDVI in Nanchang City 3.1.1.Temporal Evolution Characteristics of NDVI The study introduces a concept of image average to quantitatively characterize the vegetation evolution in Nanchang over a period of time. The month-by-month NDVI values from 2000 to 2022 in Nanchang City were extracted for integrated analysis. From May to October,the vegetation will grow in Nanchang, and the monthly average of NDVI fluctuated from 0.254 to 0.656 during the 23-year period. The annual average NDVI values fluctuated in the range of 0.412 to 0.487, which 2013 and 2015 have the min. and max. annual valus. The multi-year average NDVI value is 0.453 (dashed line in Fig. 2 ), and years with NDVI values exceeding the multi-year average are generally concentrated after 2014. From Fig. 2 , it can be observed that the overall trend of NDVI change shows an upward trend, but the rate of increase is gradually slowing down. 3.1.2.Temporal Evolution For studying the spatial changes of NDVI in Nanchang City, Theil-Sen Median based on MK trend analysis is conducted in pixels, which obtain spatial distribution of the change rate in NDVI and spatial distribution of the Mann-Kendall test statistic. Combining spatial distribution of Hurst index in Nanchang calculated pixel by pixel, the relationship between the future direction and the past tendency of vegetation cover change in Nanchang is obtained. As shown in Fig. 3 , from 2000 to 2010, the H-index in Nanchang ranged from 0.165 to 0.988, with a mean of 0.542. During this period, the overall vegetation change showed a continuous development trend. From 2011 to 2022, the H-index was from 0.149 to 0.978, with a mean of 0.522. Although the overall change still showed a consistent development trend, the regions with reverse development significantly increased (Fig. 4 ). For the city of Nanchang from 2000 to 2022, the overall Hurst index ranges from 0.124 to 0.9, with a mean of 0.444. 26.068% of the city has the NDVI Hurst is more than 0.5, while 73.932% zone has NDVI Hurst less than 0.5 (Fig. 5 ). The results indicate that over the 23 years, the overall vegetation change in Nanchang City exhibits characteristics of reverse sustained development. The closer the Hurst index is to 0.5, the less correlation there is between the persistence of future changes and past trends. Therefore, based on the direction and strength of persistence, this study classifies the Hurst index into 4 categories: strong reverse persistence(0 < H < 0. 35)、weak reverse persistence(0. 35 < H < 0. 50)、weak co-directional persistence(0. 50 < H < 0. 65)、strong persistence in the same direction(0. 65 < H < 1). With superimposing the sustainability index(H)in spatial, the future development tendency of NDVI can be acquired (Table 3 ). From 2000 to 2022, the NDVI trend in the first 11 years developed in a positive direction (improvement direction accounting for 53.4%), while in the following 12 years it developed in a negative direction (degradation direction accounting for 54.6%), with a similar proportion of improvement and degradation. Based on Fig. 5 and Table 3 , the analysis of the overall changes for 23 years indicates that the future development trend of Nanchang City may move towards a negative direction (degradation direction accounting for 64.9%), with areas of continuous degradation accounting for 12% (strongly continuous degradation accounting for 1.8%, weakly continuous degradation accounting for 10.2%). Areas that have improved in the past but are projected to degrade in the future account for 52.9% (strongly reversing improvement accounting for 10.8%, weakly reversing improvement accounting for 42.1%); areas of continuous improvement account for 15.9% (weakly continuous improvement and strongly continuous improvement accounting for 15.1% and 0.8% respectively), and areas that degraded in the past but are expected to show improvement in the future account for 18.9% (strongly reversing degradation and weakly reversing degradation accounting for 15.7% and 3.2% respectively). In Nanchang City, the areas with continuous amelioration are scattered in central and eastern district, with the continuously improved vegetation cover type being farmland. The agricultural vegetation in Nanchang City is well protected and will be sustained growth. The areas with continuous degradation are unevenly distributed, with regions of strong continuous degradation concentrated near constructive land, and regions of weak continuous degradation clustered near urban land and rivers, indicating that human activities have had adverse effects on vegetation cover. However, the predicted proportion areas with continuous degradation is small, mainly concentrated within urban areas where human activities are well controlled and have not affected natural ecological areas such as forests and grasslands.Regions that have degraded in the past but show a trend of improvement in the future do not have a clear spatial distribution pattern, mainly existing on both sides of rivers, around lakes and sporadically on the outskirts of cities, indicating that human activities have restricted urban expansion and have focused on protecting and rebuilding ecological red zones. In the future, vegetation cover around lakes, riverbanks and in suburban areas is likely to increase. Regions that have improved in the past but are expected to degrade latter are mainly situated in the southeastern and western parts of Nanchang City, including the counties of Jinxian, Anyi, and Xinjian. Jinxian County, with its low hills in south and lakeshore in north, has a severe ecological and environmental pollution situation because it is a supply base for high-quality agricultural and sideline products in Jiangxi Province; Anyi County is polluted by brick kilns, and Xinjian County is polluted by mining enterprises, leading to significant vegetation degradation. Table 3 Future change trend of vegetation NDVI in Nanchang City Change directions Future change trend 2000–2010 Percentage 2011–2022 Percentage 2000–2022 Percentage Continuous degradation Strong persistent degradation 4.4% 12.2% 1.8% Weak persistent degeneration 7.1% 25.6% 10.2% The past has improved, but the future is a degradation trend Continuous improvement of anti-strong forces 3.7% 1.9% 10.8% Continuous improvement of anti-weakness 31% 14.2% 42.1% Degradation in the past but improvement in the future Anti-weakness and persistent degradation 5.1% 16.7% 15.7% Persistent degradation of anti-strength 0.5% 1.9% 3.2% Continuous improvement Weak sustained improvement 36% 19.7% 15.1% Strong continuous improvement 11.8% 7.3% 0.8% Basically unchanged Basically unchanged 0 0 0 Note: In the table, the outliers for each region percentage have been removed. 3.2.NDVI Temporal-Driven Analysis 3.2.1.Selection of meteorological factors This study uses the image averaging method to extract the monthly average NDVI values of Nanchang City from 2000 to 2022. Regression models are established by ReliefF algorithm, MIC algorithm and RF algorithm to calculate the impact weights of various meteorological factors on NDVI. Through comprehensive analysis of the fitting degree in training set and test set, the ReliefF algorithm is found to build a regression model with the closest R 2 to 1 and the smallest error, which is more suitable in calculating the driving weights of meteorological factors on NDVI among the three algorithms. (Table 4 ). Table 4 Evaluation factors of three machine learning models ReliefF MIC RF Training set R 2 0.963 0.946 0.956 MAE 0.017 0.021 0.019 MBE −0.001 −0.005 0.003 RMSE 0.023 0.027 0.025 Test set R 2 0.934 0.918 0.914 MAE 0.023 0.026 0.027 MBE −0.003 −0.005 0.003 RMSE 0.029 0.033 0.033 The regression model by the ReliefF algorithm evaluates the correlation between meteorological factors A1-A24 (see Table 1 ) and NDVI, and obtaines the importance of each meteorological factor (Fig. 6 ). Screening out the top ten meteorological factors with the highest proportion of NDVI driving weights, they are A23,A20,A22,A3,A18,A19,A21,A13,A17 and A16. Among them,underground temperature has a particularly prominent driving effect on NDVl. The underground temperature, i.e. soil temperature, can promote the absorption of nutrients by vegetation rhizomes and affect the accumulation of vegetation biomass, it affects soil evaporation and vegetation transpiration, which in turn affects the water supply and growth of vegetation. This temperature influences the metabolic activities of soil microorganisms, providing necessary nutrients for vegetation. At the same time, the underground temperature will affects not only the growth processes of vegetation such as photosynthesis and respiration, but also the growth cycle of vegetation, including stages such as germination, flowering and fruiting. 3.2.2.Comparison of the importance of meteorological factors and anthropogenic factors Conduct a correlation analysis between the ten meteorological factors screened above and anthropogenic factors, and perform ReliefF feature selection analysis. It is obvious that anthropogenic factors are significantly more important to NDVI than meteorological factors (Fig. 7 ). With social and economic developments, the increase in population density, acceleration of urbanization and expansion of farmland lead to a decrease in vegetation coverage, while changes in policy orientation and people's environmental awareness also affect vegetation inversion. Although meteorological factors can also affect vegetation coverage, the human activities on meteorological factors are more apparent than the past, thereby creating complex effects on vegetation growth. Climate change is a long-term cumulative natural process, with changes in meteorological factors such as rainfall and temperature usually slow and periodic, but the impact of anthropogenic factors is rapid and severe. So as time goes on, the impact of anthropogenic factors are more significant than meteorological factors. 3.3.NDVI Spatial-Driven Analysis Temporal-Driven Analysis 3.3.1.Driving analysis of NDVI derived from geographical detector (1)Factor detection This manuscript uses geographic detector to reveal the interpretive force of driving factors in Nanchang spatially differentiation in NDVI(q). As shown in Fig. 8 , factors X1-X8 (see Table 1 ) do not all significantly affect the evolution of NDVI. The blue frame bar indicates that the q value of the driver is significant (p X6 > X8. Similar to the results driven by time, the human factors are significant while the meteorological factors are not significant. Considering other driving factors, land cover types are more significant than elevation. Among the anthropogenic factors, the influence of X5 and X6 factors, spatial distribution in vegetation cover is more than 30%, with the explanatory power of X5 being over 90%. Considering other driving factors, land cover types are more significant than elevation. Among the anthropogenic factors, two driving factors, X5 and X6, have an impact of over 30% on spatial distribution of vegetation cover, with the explanatory power of X5 being over 90%. X8 are a secondary driving factor, with an average explanatory power over 20%. Among the eight driving factors, the factors that are not significantly explanatory for NDVI are: X1, X2, X3, X4 and X7. More than 75% of the study area has the following characteristics: an elevation of 30 meters, a slope of less than 4°, which is being an area with intensive human production activities. The main reason for vegetation evolution is the resource development and utilization led by human activities and land cover types’ variation. With the increase in population and the development of the economy, there is a growing demand for food and cash crops, leading to the conversion in some forests to farmland. Urban construction and expansion also occupy farmland, forests and grasslands, leading to large-scale deforestation. These variations in land use patterns result in the degradation of NDVI. Deforestation and farming activities had led to soil erosion and degradation, exacerbating the phenomenon of soil erosion. Biodiversity and ecosystem diversity have been damaged, even abandoned farmland, mainly in the low hills and mountains of Anyi, Jinxian, and Xinjian County. Once over-exploited land is abandoned, such as land that has been deforested or reclaimed from lakes, the natural vegetation can not recover in a short time, leaving the surface exposed and increasing the risk of soil sanding and desertification. This is mainly distributed in northern Nanchang County, the northern and southern parts of Xinjian County, and the low hill areas of Jinxian County. (2)Interaction detection Interaction detection can mainly evaluate the combined effects which investigate different driving factors in NDVI. There were only two interactions, dual-factor enhancement and non-linear enhancement, for factors X1-X8 in 2000, 2005, 2010, 2015, and 2020. Most factor interactions were nonlinear enhancements, among which X1∩X2, X1∩X7, X2∩X7, and X3∩X7 represent the interaction effects of meteorological factors and topographical factors, which go beyond a simple linear superposition of their respective independent effects (Fig. 9 ). The most important factors in vegetation growth are appropriate precipitation and temperature, and synergistic effect can greatly promote vegetation growth and increase NDVI value; as the altitude increases, temperatures usually decrease, forming the different climatic vertical zones, which have different temperature ranges that affect the types and distribution of vegetation; topographical factors, together with rainfall, influence the distribution of water, soil formation and nutrient retention, thereby affecting the structure and function of vegetation; topographical features affect water evaporation, transpiration and recycling, for example, high-altitude mountain ranges can promote the formation of clouds and fog, increasing air humidity, and indirectly increase water supply to vegetation. With the passage of time, most driving factors have evolved into dual-factor enhanced interaction modes, such as X3∩X6, X4∩X8, X6∩X8. Human activities may have different impacts on environmental factors at different time points, leading to stronger synergistic effects between two factors that originally had a non-linear relationship. The interactions between different factors become more closely intertwined and long-term environmental trends may also mask or enhance short-term fluctuation effects, resulting in different interaction patterns observed at different time points. 3.3.2.The impact of human factors on NDVI in Nanchang City (1)Single correlation analysis between NDVI and anthropogenic factors In base pixel scale, a single correlation analysis had been managed on the NDVI, nighttime lights, and population density from 2000 to 2022, as shown in Fig. 10 . Through statistical analysis, there are 56.9% and 50.4% of the areas that vegetation cover is positively correlated with nighttime lighting and population density respectively, so with an overall positive relationship. The regions where NDVI is negatively correlated with nighttime lights are scattered in main urban area of Nanchang City, Xinjian County and Jinxian County; and the negatively correlate regions in population density are sporadically distributed in main urban area of Nanchang City and Jinxian County. Regions which NDVI is positive correlation with population density and night lights are often found near rivers and lakes. (2)Partial correlation analysis of NDVI and anthropogenic factors According to pixel scale, a partial correlation analysis had been controlled in NDVI and anthropogenic factors(Fig. 11 ). A two-tailed t-test at a significance level(α = 0.05) indicates that the areas where NDVI and nighttime lights have the partial correlation coefficient and pass significance test(t ≥ t 0.05 ), mainly located at heart of Nanchang City. Regions showing a negative correlation are mainly in the main urban area and Jinxian County, accounting for 43.2% of the significant cassociation areas. The biased correlation area between NDVI and population density is dominated by negative correlation, which is scattered in main urban area and Jinxian County, with a total area of 911.097 km 2 , making for 51.9% of the significant correlation areas. (3)Complex correlation analysis between NDVI and anthropogenic factors Complex correlation in spatial pattern between NDVI and anthropogenic factors in Nanchang City is shown in Fig. 12 . The complex correlation coefficient ranges from 0 to 0.999, and the spatial mean is 0.066. In most areas downtown of Nanchang City and Jinxian County, complex correlation between NDVI and anthropogenic factors is significant, while areas with weaker correlation are located at the boundaries of various districts and counties. The F-test at a significance level of α = 0.05 indicates that the regions passing the significance test (F > F 0.05 ) are mainly scattered in the downtown of Nanchang City, with a total area is 1755.48 km 2 where accounts 24.4% of the total area. Due to the rapid economic development, areas where NDVI and population density have negative correlation are characterized by high reflectance and heat absorption of buildings and roads, leading to urban heat island effect which adversely affects vegetation growth in the process of urbanization. At the same time, nighttime lights may impact the growth cycle and physiological processes of fern plants. Areas where NDVI and population density have negative correlation is owing to the high density with population in the main urban areas, resulting in air, soil and water pollution. In Jinxian County, as a high-quality agricultural and sideline product supply base in Jiangxi Province, agricultural pollution is generated, which may have a negative impact on vegetation health, reducing NDVI values. The regions where NDVI is positively correlated with nighttime lighting and population density are all located around rivers and lakes. For protecting ecological system of lakes and enhancing landscapes, ecological protection measures and urban greening projects are implemented around the lakes. These measures have helped to increase vegetation coverage, while also improved lighting facilities in these areas for nighttime protection and monitoring. Due to its beautiful natural scenery and recreational value, lakes often become hotspots of tourism and leisure activities. This has led to the construction of more tourist facilities and leisure areas near the lakes, increasing human activities and resulting in higher brightness in nighttime lights data. 3.3.3.The influence of land use change on spatiotemporal variation of NDVI This study utilized land cover data of 2000, 2005, 2010, 2015 and 2020 to investigate the spatiotemporal changes in NDVI for different land cover types. The charges in area proportion and NDVI are shown in Fig. 13 . The order with area size in different land cover types is: farmland > other types > forest > urban and rural land > grassland; the order of the average annual NDVI for different land cover types is: forest (0.746) > grassland (0.712) > farmland (0.708) > other types (0.593) > urban and rural land (0.552). The NDVI values of different land cover types do not show consistent correlation with their respective area proportions. Forest is almost 16.2% of the total study area, equivalent to 29.9% of the total farmland, but the average annual NDVI of forest is about 0.038 higher than that of farmland. During the research period, the area of farmland is decreasing, while the corresponding NDVI is increasing. This is closely related with the policy that return farmland to forests in Nanchang City. Although the area of farmland has decreased, the increased protection and management of existing arable land has improved the quality of the arable land, thus increasing the NDVI value. Due to the prominent performance of this land use pattern, the NDVI values of farmland are on the rise; the fluctuation range of forest and grassland areas is relatively small, and they have also shown an increasing trend over the years. 4. Discussion 4.1 Spatial-temporal evolution characteristics of NDVI in Nanchang City The terrain of Nanchang City is mainly plain, where the overall topography of northwest is higher than southeast, developing into low hills, hillock, and plains in sequence, which shows a layered landform characteristic. The vegetation coverage is influenced with natural factors in meteorological and anthropogenic. From 2000 to 2022, the average NDVI of vegetation was 0.453, showing an overall increasing trend, but the growth rate gradually slowed over time. This view is consistent with Chen's study of the evolution of NDVI in Nanchang during the periods 1995 to 2005 and 2013 to 2017, which showed that although NDVI showed an increasing trend during these two periods, the value added was decreasing (Chen et al. 2022). According to the "Overall Land Use Plan of Nanchang City (2006–2020)", the farmland in areas like Donghu District and Xihu District significantly decreased, while the construction land is significantly increased. Continuous progress of social economy and the improvement of urbanization level have increased the demand for construction land, to some extent, it affects the upward trend of NDVI. In ZHAO's study, it was also shown that the construction and expansion of cities led to the occupation of farmland, forest and grassland, as well as large areas of abandonment and deforestation, and that these land-use changes became the main cause of the degradation of vegetation growth (Zhao et al. 2016). In 2008, 2010, 2016, 2019 and 2022, the vegetation coverage significantly decreased. In 2008, the continuous economic development of Nanchang City led to more land being used for industrial construction, resulting the reduction in vegetation coverage and affecting the NDVI value. In 2010, persistent heavy rainfall in Nanchang City caused severe flooding disasters. According to the 2016 government work report, the city made significant progress in urban planning and construction, including old city renovation and new city construction, which changed land use and subsequently affected vegetation coverage and NDVI values. In 2019, the provincial government issued the "Opinions on Comprehensive Promotion of Comprehensive Tourism Development", making the tourism industry to a priority in deepening the construction of a strong tourism province, thereby affecting the NDVI. With the gradual easing of the COVID-19 pandemic, people resumed orderly outdoor activities and daily life in 2022, increasing human activities that impact vegetation. Spatially, the vegetation coverage in Nanchang City is mainly showing a trend of degradation, with areas of strong persistent degradation concentrated near construction land, areas of weak persistent degeneration clustered near urban land and rivers, and regions that improved in the past but are now trending towards degradation mainly scattered in southeast and west, including Jinxian County, Anyi County, and Xinjian County. Due to human activities, agricultural pollution in Jinxian County, brick kiln pollution in Anyi County, and mining pollution in Xinjian County have led to vegetation degradation. In addition, human activities may indirectly affect meteorological factors, further influencing the changes in NDVI and indicating a trend of ecological degradation. 4.2. Analysis of NDVI Driving Factors in Nanchang City This study compared and analyzed the impacts of meteorological factors and anthropogenic factors on NDVI. The anthropogenic driving effect on NDVI is more notable. ZHAO similarly identified that human-led activities, resource exploitation and land reclamation, were the main cause of vegetation cover change (Zhao et al. 2016). This discuss confirmed that among meteorological drivers soil temperature mainly affected vegetation growth, while among anthropogenic factors the most significant factors affecting vegetation were population density and night lighting. This was different from Zhang's (2022) study where air temperature and rainfall had the most significant effect on vegetation cover. We cannot consider the effect of air temperature on photosynthesis and transpiration of plants alone, the thermal properties ofsoil and the response of plant root system to temperature should also be considered. Changes in land cover types also significantly impact the evolution of NDVI, showing no difference with Wei's (2010) study elaborating that soils are critical to the ecosystem. Although we did not directly conclude that economic development has a significant effect on NDVI, it still influences vegetation growth condition by affecting drivers indirectly such as nighttime lighting, which comes to an agreement with Li's view (2024). Interannual variation trend of urban meteorological disasters is becoming increasingly significant. For example, urban heavy rainfall events exhibit periodic fluctuations. Although the annual average number of heavy rainfall and thunderstorm days has decreased in the past decade, the intensity and frequency in short-term intense precipitation events have increased (Wang et al. 2014 ). Additionally, over the past 55 years, Nanchang City has shown an increasing trend in the annual average number of hazy days, especially in the autumn and winter seasons (Chen et al. 2016). In recent years, Jiangxi Province has also been facing increasingly severe ozone pollution issues, with factors such as prolonged sunshine duration, extreme temperature changes and abnormal decreases in precipitation collectively leading to a significant increase in near-surface ozone concentrations in autumn (Qian et al. 2021 ). These phenomena highlight the key role of climate change factors in monitoring the ecological development of cities. Given the global warming trend and the extreme weather events, it is expected that future climate will more and more variable. Meteorological combined effects will be significant than it is now. Therefore, further in-depth research on this issue is still needed. Human activities like population density and nighttime lights are predominantly affected vegetation cover and ecosystem health, and they need to be managed properly to ensure the ecosystem have a sustainable development. The government should formulate strict ecological protection policies to limit overgrazing, illegal logging and other activities that destroy natural ecosystems. Only rational planning of land use and optimization of land resource allocation, especially in ecologically sensitive areas, can reduce the erosion of natural ecosystems. Moreover, ecological restoration projects such as afforestation, wetland restoration and desertification control should be strengthened to ameliorate the ecological environment. The monitoring and evaluation system of vegetation cover and ecosystem health has been continuously established and improved, and the monitoring and evaluation work of vegetation cover has been carried out regularly. It is necessary to develop ecological tourism, ensure the sustainable development of the ecological environment, rationally plan the flow of tourists and avoid the destruction of vegetation and ecological pressure caused by excessive tourism. 5. Conclusions In manuscript, the NDVI of Nanchang City were used to analyze the evolution and drivers of spatio-temporal separation from 2000 to 2022 through trend analysis and machine learning feature selection and so on . Throughout the study period in Nanchang City,the average NDVI was 0.453, showing an overall upward trend. However, as time passed, the rate of increase gradually slowed down.Spatially, the average Hurst index of Nanchang City was 0.444, and the overall vegetation change exhibited characteristics of reverse sustained development. The degraded areas of NDVI accounted for 64.9%, showing a trend of “improvement around rivers and lakes” and “large-scale degradation of urban land”.In time-driven analyses,soil temperature had a prominent effect on vegetation cover, but overall the impact of human factors was more significant than the meteorological factors.The order of factors with significant influence on vegetation cover and evolution in Nanchang by geographical detector are as following: population density > nighttime lights > land cover types. The interactions touching the evolution of NDVI are characterized by both dual-factor enhancement and nonlinear enhancement. Spatial correlation indicated that the interannual variation levels of NDVI which negatively correlated with population density and nighttime lights, mainly scattered in heart of Nanchang City and Jinxian County; while the positively correlated areas are distributed around rivers and lakes. There is no consistency between the area proportion of land cover types and vegetation coverage. Economic level cannot directly influence NDVI,but it still influences vegetation growth condition by affecting drivers indirectly such as nighttime lighting.And among the meteorological drivers, the effect of soil temperature on NDVI is not negligible.Qualitatively and quantitatively exploration were been progressed in the evolution and driving of NDVI by spatial and temporal separation, to provide scientific basis for dynamic monitoring, management and protection of vegetation in Nanchang City.In future work, we plan to conduct multi-driver response analysis for different vegetation types and synthesize the driving factors to study the multi-factor driving mechanism of vegetation ecology, so as to provide scientific references for the dynamic changes of the ecological environment. Declarations Author Contributions: Li Jiatong: Conceptualization, methodology, formal analysis, writing—original draft preparation, writing—review and editing, visualization. Zhu Jiaqi: Conceptualization, methodology, writing—review and editing. Guo Qiyun: Conceptualization, formal analysis, writing—review and editing. Xu Yue: formal analysis, writing—original draft preparation. Li Huishan: data curation. Liu Sihang: visualization. Wu Hua: supervision. Funding: Project support: Construction of Talent Innovation Team and Laboratory Platform of Tibet University, Project No. 2022ZDTD10. Data Availability: The data presented in this study are available on request from the corresponding author. Ethics approval and consent to participate: I would like to declare on behalf of my co-authors thatthe work described was an original research that has not been publishedpreviously and not under consideration for publication elsewhere, inwhole or in part. Consent for publication: The participant has consented to the submission of the article research to the joumal. Conflict of interes: The authors declare no confict of interest. Publisher's Note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. References Ahmad A.,Zhang JH,Bashir B.,et al. Exploring vegetation trends and restoration possibilities in Pakistan by using Hurst exponent. Environ Sci Pollut Res 2023 , 30:91915–91928. https://doi.org/10.1007/s11356-023-28822-0. Adichie, J. N. Estimates of regression parameters based on rank tests. Annals of Mathematical Statistics, 1967,38: 894–904. Breiman, L. Random Forests. Machine Learning 2001 ,45:5–32. https://doi.org/10.1023/A:1010933404324. Cao, R., Jiang, W., Yuan, L. et al. Inter-annual variations in vegetation and their response to climatic factors in the upper catchments of the Yellow River from 2000 to 2010. J. Geogr. Sci. 2014 , 24:963–979.https://doi.org/10.1007/s11442-014-1131-1. Chen,F.,Wu,F..Spatial and temporal evolution analysis of vegetation cover in Nanchang based on NDVI. Yangtze River Information and Communication 2022 ,35(11):120-122.(in chinese) Chen,P.;Pan,H.;Xu,Y.;He,W.;Yao,H.Spatial-Temporal Evolution Characteristics and Driving Force Analysis of NDVI in Hubei Province, China, from 2000 to 2022. Forests 2024 , 15:719. https://doi.org/10.3390/f15040719. Chen X.g,Hu L.,Peng W.,et al.Characteristics of hazy weather meteorological elements and major air pollutants changes in Nanchang City from 1960 to 2014. Journal of Meteorology and Environment 2016 ,32(5):114-121.(in chinese) Das,U.,Behera,B.. Geospatial assessment of ecological vulnerability of fragile Eastern Duars Forest integrating GIS-based AHP, CRITIC and AHP-TOPSIS models. Geomatics, Natural Hazards and Risk 2024 , 15(1). https://doi.org/10.1080/19475705.2024.2330529. Didan, K..MODIS/Terra Vegetation Indices Monthly L3 Global 1km SIN Grid V006, NASA EOSDIS Land Processes DAAC . 2015 https://doi.org/10.5067/MODIS/MOD13A3.006. Dror Y. Kenett,Huang XQ,Irena Vodenska,et al.Partial correlation analysis:applications for financial markets. Quantitative Finance 2014 ,15:569-578.https://api.semanticscholar.org/CorpusID:14156969. Gadedjisso-Tossou, A.; Adjegan, K.I.; Kablan, A.K.M. Rainfall and Temperature Trend Analysis by Mann–Kendall Test and Significance for Rainfed Cereal Yields in Northern Togo. Sci 2021 , 3,17. https://doi.org/10.3390/sci3010017. Ge, R., Zhou, M., Luo, Y. et al. McTwo: a two-step feature selection algorithm based on maximal information coefficient. BMC Bioinformatics 2016 , 17:142. https://doi.org/10.1186/s12859-016-0990-0. Iratxe Niño-Adan,Diana Manjarres,Itziar Landa-Torres,et al.Feature weighting methods: A review. Expert Systems with Applications . 2021 ,184:115424.https://www.sciencedirect.com/science/article/pii/S0957417421008423. Jiangxi Bureau of Statistics, 2021 , Jiangxi Statistical Yearbook 2017[J]. Beijing: China Statistics Press: 259-298.(in chinese) Jin,HY.,Chen,XH.,Wang,YM.,et al.Spatio-temporal distribution of NDVI and its influencing factors in China. Journal of Hydrology 2021 ,603:127129.https://doi.org/10.1016/j.jhydrol.2021.127129. Jose Alvarez-Ramirez,Juan C. Echeverria,Eduardo Rodriguez.Performance of a high-dimensional R/S method for Hurst exponent estimation. Physica A: Statistical Mechanics and its Applications 2008 ,26:6452-6462. Khaled H. Hamed.Trend detection in hydrologic data: The Mann–Kendall trend test under the scaling hypothesis. Journal of Hydrology 2008 ,348:350-363. Kirch, W.(eds) Pearson’s Correlation Coefficient. Encyclopedia of Public Health. Springer, Dordrecht 2008 . https://doi.org/10.1007/978-1-4020-5614-7_2569. Kong,C.,Huang,J.,Du,S., Li,G.. Exploring the Driving Forces of Vegetation Greening on the Loess Plateau at the County Scale. Forests. 2024 ,15(3):486. https://doi.org/10.3390/f15030486 Kumari N, Srivastava A, Dumka UC.A Long-Term Spatiotemporal Analysis of Vegetation Greenness over the Himalayan Region Using Google Earth Engine. Climate 2021 ,9(7):109. https://doi.org/10.3390/cli9070109. Lakens D. Equivalence Tests: A Practical Primer for t Tests, Correlations, and Meta-Analyses. Soc Psychol Personal Sci. 2017 ,8(4):355-362. doi: 10.1177/1948550617697177. Li,P.,Wang,J.,Liu,MM.,et al.Spatio-temporal variation characteristics of NDVI and its response to climate on the Loess Plateau from 1985 to 2015. CATENA 2021 ,203:105331.https://doi.org/10.1016/j.catena.2021.105331. Li,Q, Gao,X, Li,J, Yan,A, Chang,S, Song,X, Lo,K. Nonlinear time effects of vegetation response to climate change: Evidence from Qilian Mountain National Park in China. Sci Total Environ. 2024 ,933:173149. Li, T., Dong, Y., Wei, X. et al. The rapid prosperity of China’s Pearl River Delta from the perspective of social–ecological coupling: implications for sustainable management. Sci Rep 2024 ,14:19914. Li SX,Feng C,Yang L.Spatiotemporal patterns of evapotranspiration in the Xiangjiang River Basin from 2000 to 2020. Hydrology Research 2023 ,54.doi = {10.2166/nh.2023.143}. Li XQ,Jia R. MICQ-IPSO: An effective two-stage hybrid feature selection algorithm for high-dimensional data. Neurocomputing 2022 ,501:328-342. Liu,Y.; Huang,C.;Zhang,L. The Spatio-Temporal Patterns and Driving Forces of Land Use in the Context of Urbanization in China: Evidence from Nanchang City. Int. J. Environ. Res. Public Health 2023 , 20:2330. https://doi.org/10.3390/ijerph2003233. Luo, H., Wang, L., Fang, J., Li, Y., Li, H., Dai, S. NDVI, Temperature and Precipitation Variables and Their Relationships in Hainan Island from 2001 to 2014 Based on MODIS NDVI. In: Bian, F., Xie, Y. (eds) Geo-Informatics in Resource Management and Sustainable Ecosystem. GRMSE 2015 2015. Communications in Computer and Information Science , 2016 ,569. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-49155-3_34. Novillo CJ, Arrogante-Funes P, Romero-Calcerrada R. Recent NDVI Trends in Mainland Spain: Land-Cover and Phytoclimatic-Type Implications. ISPRS International Journal of Geo-Information. 2019 ,8(1):43. https://doi.org/10.3390/ijgi8010043. Peng, S. 1-km monthly mean temperature dataset for china (1901-2022). National Tibetan Plateau / Third Pole Environment Data Center . 2019 . https://doi.org/10.11888/Meteoro.tpdc.270961. https://cstr.cn/18406.11.Meteoro.tpdc.270961. Peng, S. 1-km monthly precipitation dataset for China (1901-2022). National Tibetan Plateau / Third Pole Environment Data Center . 2020 . https://doi.org/10.5281/zenodo.3185722. Qian,Y.,Xu,B.,Xia,LJ.,et al.Characterization of ozone pollution in Jiangxi Province from 2016 to 2019 and analysis of the influence of meteorological factors[J]. Environmental Science 2021 ,42(5):2190-2201.(in chinese) Qiu HB,Luo J,Zhang JJ.Robustness of F-tests in singular linear models.Acta Mathematica Sinica, English Series 2014 ,30:872-880.https://api.semanticscholar.org/CorpusID:120336999. Quirk, T.J., Palmer-Schuyler, J. Multiple Correlation and Multiple Regression. In: Excel 2019 for Human Resource Management Statistics. Excel for Statistics. Springer, Cham. 2020. https://doi.org/10.1007/978-3-030-58001-8_7 Qu,S.,Wang,LC.,Lin,AW.et al.Distinguishing the impacts of climate change and anthropogenic factors on vegetation dynamics in the Yangtze River Basin,China. Ecological Indicators . 2020 ,108:105724.https://doi.org/10.1016/j.ecolind.2019.105724. Ren, Y., Deng, L., Zuo, S. et al. Geographical modeling of spatial interaction between human activity and forest connectivity in an urban landscape of southeast China. Landscape Ecol 2014 ,29:1741–1758. https://doi.org/10.1007/s10980-014-0094-z. Rivas-Tabares, D.A.; Saa-Requejo, A.; Martín-Sotoca, J.J.; Tarquis, A.M. Multiscaling NDVI Series Analysis of Rainfed Cereal in Central Spain. Remote Sens. 2021 , 13:568. https://doi.org/10.3390/rs13040568. Ryan J. Urbanowicz,Melissa Meeker,William La Cava,et al.Relief-based feature selection: Introduction and review. Journal of Biomedical Informatics 2018 ,85:189-203. Song YZ,Wang JF,Ge Yong,et al.An optimal parameters-based geographical detector model enhances geographic characteristics of explanatory variables for spatial heterogeneity analysis: cases with different types of spatial data. GIScience & Remote Sensing 2020 ,57(5):593-610.https://api.semanticscholar.org/CorpusID:219418482. Spolaôr N,Monard MC.Evaluating ReliefF-Based Multi-Label Feature Selection Algorithm. Ibero-American Conference on AI 2014 .https://api.semanticscholar.org/CorpusID:23555972. Sun,R., Chen,S., Su,H.,Hao,G.. "Spatiotemporal Variation of Vegetation Coverage and its Response to Climate Change Before and After Implementation of Grain for Green Project in the Loess Plateau," IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium, Yokohama, Japan, 2019 :9546-9549, doi: 10.1109/IGARSS.2019.8898716. Tang,X.; Wang,J.; Lu, J.; Liu, G.; Chen, J. Improving Bearing Fault Diagnosis Using Maximum Information Coefficient Based Feature Selection. Appl. Sci. 2018 , 8:2143. https://doi.org/10.3390/app8112143 Teixeira, A.L., Leal, J.P. , Falcao, A.O. Random forests for feature selection in QSPR Models - an application for predicting standard enthalpy of formation of hydrocarbons. J Cheminform 5 , 2013 . https://doi.org/10.1186/1758-2946-5-9. Terefe T.,Feyera S.,Moges K..The impact of land use/land cover change on ecosystem services in the central highlands of Ethiopia. Ecosystem Services 2017 ,(23):47-54.https://doi.org/10.1016/j.ecoser.2016.11.010 Tian,XR.,Tao,Z.,Xie,Y.,et al.Spatiotemporal Evolution and Driving Mechanism of Fractional Vegetation Coverage in the Yangtze River Delta. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. 2024 ,17,10979-10997.https://api.semanticscholar.org/CorpusID:270177260. Tong,S.,Zhang,J.,Bao,Y. et al. Analyzing vegetation dynamic trend on the Mongolian Plateau based on the Hurst exponent and influencing factors from 1982–2013. J. Geogr. Sci. 2018 , 28:595–610. https://doi.org/10.1007/s11442-018-1493-x. Wang,J.,Zhao,J.,Zhou,P.,Li K,Cao Z,Zhang H,Han Y,Luo Y, Yuan X.Study on the Spatial and Temporal Evolution of NDVI and Its Driving Mechanism Based on Geodetector and Hurst Indexes: A Case Study of the Tibet Autonomous Region. Sustainability. 2023 ,15(7):5981. https://doi.org/10.3390/su15075981. Wang JX,Hu MG,Zhang FS,et al.Influential factors detection for surface water quality with geographical detectors in China. Stochastic Environmental Research and Risk Assessment 2018 ,32:2633-2645. Wang,RL.,Wu,F.,Liu,ZP.,et al. Characterization of urban meteorological disaster changes in Nanchang over the past 60 years. Research on Meteorology and Disaster Reduction 2014 ,37(2):44-49.(in chinese) Wang XQ,Dang X,Peng HX,et al.The Theil-Sen Estimators in a Multiple Linear Regression Model. Manuscript , 2009 . Wang,Z.,Liu,Z.,Huang,M.NDVI joint process-based models drive a learning ensemble model for accurately estimating cropland net primary productivity (NPP). Front. Environ. Sci . 2024 ,11.https://doi.org/10.3389/fenvs.2023.1304400. Wei,LN.,Zheng,L.,Zhang,JW.,et al. Analysis of land use/cover change and its human factor drivers in Nanchang. Anhui Agricultural Science 2010 ,38(21):11263-11268.DOI:10.13989/j.cnki.0517-6611.2010.21.193.(in chinese) Wu RN,Zhang JQ,Bao YH,et al.Geographical Detector Model for Influencing Factors of Industrial Sector Carbon Dioxide Emissions in Inner Mongolia, China. Sustainability 2016 ,8:149.https://api.semanticscholar.org/CorpusID:18103433. Wu,YZ.; Shi,KF.; Chen,ZQ.; Liu,SR.; Chang,ZJ. An improved time-series DMSP-OLS-like data (1992-2022) in China by integrating DMSP-OLS and SNPP-VIIRS. IEEE Transactions on Geoscience and Remote Sensing , 202 2 ,60. DOI:10.1109/TGRS.2021.3135333. Xu,XL.,Liu,JY.,Zhang,SW.,Li,RD.,Yan,CZ.,Wu,SX.. China multi-period land use remote sensing monitoring dataset (CNLUCC). Resource and Environmental Science Data Registration and Publication System 2018 .DOI:10.12078/2018070201.(in chinese) Xu,Y.,Dai,QY.,Zou,B.,et al.Tracing climatic and human disturbance in diverse vegetation zones in China: Over 20 years of NDVI observations. Ecological Indicators 2023 ,156:105331.https://doi.org/10.1016/j.ecolind.2023.111170. Yang,Y.,Wang,S.,Bai,X.,Tan,Q.,Li,Q.,Wu,L.,Tian,S.,Hu,Z.,Li,C.,Deng,Y.. Factors Affecting Long-Term Trends in Global NDVI. Forests. 2019 ,10(5):372. https://doi.org/10.3390/f10050372. Yang,ZW.,Gao,X.,Lei,JQ.,et al.Analysis of spatiotemporal changes and driving factors of desertification in the Africa Sahel. CATENA 2022 ,213:106213.https://doi.org/10.1016/j.catena.2022.106213. Zhang HY,Feng ZQ,Feng SY,et al.A Survey of Methods for Estimating Hurst Exponent of Time Sequence. ArXiv 2023 .https://api.semanticscholar.org/CorpusID:264828969. Zhang, J., Chen, M., Zhao, S., Hu, S., Shi, Z., Cao, Y. ReliefF-based EEG sensor selection methods for emotion recognition. Sensors (Switzerland) 2016 ,16(10):1558. https://doi.org/10.3390/s16101558. Zhang,J.; Jiao,G.; Ye,Q.; Gu,X. The Impact of Urban Expansion on the Urban Thermal Environment: A Case Study in Nanchang, Jiangxi, China. Sustainability 2022 , 14:16531. https://doi.org/10.3390/su142416531. Zhang, L., Cong, Z., Zhang, D., & Li, Q. Response of vegetation dynamics to climatic variables across a precipitation gradient in the Northeast China Transect. Hydrological Sciences Journal 2017 , 62(10):1517–1531. https://doi.org/10.1080/02626667.2017.1337274 Zhang,L.,Zhang,Y.,Wang,J. et al. Spatiotemporal evolution characteristics and driving forces of vegetation cover variations in the Chengdu-Chongqing region of China under the background of rapid urbanization. Environ Sci Pollut Res 2024 ,31:22976–22993. https://doi.org/10.1007/s11356-024-32645-y. Zhang,X.,Han,L.,Li,L.,and Bai, Z..Analysis of desertification and the driving factors over the Loess Plateau. Geocarto International , 2023 ,38(1). https://doi.org/10.1080/10106049.2023.2290175. Zhang XM. Characterization of spatial and temporal variations of NDVI in Jiangxi Province and its correlation analysis with climate factors[J]. Yangtze River Information and Communication , 2022 ,35(06):7-12.(in chinese) Zhang, Y., Zhang, L., Wang, J.,et al. Quantitative analysis of NDVI driving factors based on the geographical detector model in the Chengdu-Chongqing region, China. Ecological Indicators . 2023 ,155:110978. Zhao,LH.,Wang,P.,Ouyang,XZ.,Wu,ZW.. Spatio-temporal evolution of vegetation cover and its response to non-climatic factors in Nanchang. Journal of Ecology 2016 ,36(12):3723-3733.(in chinese) Zhao NZ,Liu Y,Cao GF,et al.Forecasting China’s GDP at the pixel level using nighttime lights time series and population images. GIScience & Remote Sensing 2017 ,(1):1-19. Zhao,W.,Wang,H.,Zhang,H. et al. Precipitation and anthropogenic activities regulate the changes of NDVI in Zhegucuo Valley on the southern Tibetan Plateau. J. Mt. Sci. 2024 , 21:607–618. https://doi.org/10.1007/s11629-023-8299-8. Zhong, H.; Wang, H. Temporal and spatial variation of normalized vegetation index in Hubei Province from 2007 to 2016. J. Cent.China Norm. Univ. (Nat. Sci.) 2018 , 52(04), 582–588. Zhu,LM.,Zhu,KX.,Zeng,XJ..Evolution of landscape pattern and response of ecosystem service value in international wetland cities: A case study of Nanchang City. Ecological Indicators 2023 ,155:110987. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5366943","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":378961465,"identity":"e58952ab-e73c-425e-82d7-efd5edd2ebd1","order_by":0,"name":"Jiatong Li","email":"","orcid":"","institution":"Tibet University","correspondingAuthor":false,"prefix":"","firstName":"Jiatong","middleName":"","lastName":"Li","suffix":""},{"id":378961466,"identity":"61e66260-8bb1-446e-9dee-b006afd64ea8","order_by":1,"name":"Hua Wu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1ElEQVRIiWNgGAWjYBACNvb+D4f/VEjI8bM3HyBOCx/PAcMHPGcsjCV7jiUQp0VOIsHYgLetItHgRo4BkQ5jSEiTkGyTSJBsyPl44w2DnZxuA0EtB45JGJyTyONnOLvZcg5DsrHZAUJaGBuBVpRJFEs29m6T5mE4kLiNoBZmZjaJA2wSiRsO8zwjUgsbG7NhQxtQyzEeNiK18PAwPmY4IwEMZDZjyzkGRPhFfv4bhsMMFXVy/PKPH954U2EnR1ALCpDgITJqkLWQqmMUjIJRMApGBAAAMdM+5Zj4nqMAAAAASUVORK5CYII=","orcid":"","institution":"Tibet University","correspondingAuthor":true,"prefix":"","firstName":"Hua","middleName":"","lastName":"Wu","suffix":""},{"id":378961467,"identity":"e3cf4182-d78b-4c42-bdc3-acd362e4a48d","order_by":2,"name":"Qiyun Guo","email":"","orcid":"","institution":"Tibet University","correspondingAuthor":false,"prefix":"","firstName":"Qiyun","middleName":"","lastName":"Guo","suffix":""},{"id":378961468,"identity":"55c03df1-3ebd-4cda-b073-e68541bc54f7","order_by":3,"name":"Yue Xu","email":"","orcid":"","institution":"Tibet University","correspondingAuthor":false,"prefix":"","firstName":"Yue","middleName":"","lastName":"Xu","suffix":""},{"id":378961469,"identity":"e0a7b8ca-9831-4f44-ae9c-8fc21bf09453","order_by":4,"name":"Huishan Li","email":"","orcid":"","institution":"Nanchang Meteorological Bureau","correspondingAuthor":false,"prefix":"","firstName":"Huishan","middleName":"","lastName":"Li","suffix":""},{"id":378961470,"identity":"6158e14c-2be8-47d2-a18a-86419f81b2bf","order_by":5,"name":"Sihang Liu","email":"","orcid":"","institution":"Northeast Agricultural University","correspondingAuthor":false,"prefix":"","firstName":"Sihang","middleName":"","lastName":"Liu","suffix":""},{"id":378961471,"identity":"d0b70668-2e02-46a4-9d7b-9305b092dd27","order_by":6,"name":"Jiaqi Zhu","email":"","orcid":"","institution":"Tibet University","correspondingAuthor":false,"prefix":"","firstName":"Jiaqi","middleName":"","lastName":"Zhu","suffix":""}],"badges":[],"createdAt":"2024-10-31 10:53:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5366943/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5366943/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":71685987,"identity":"7bb63472-f08e-4777-a350-a8a98e60b676","added_by":"auto","created_at":"2024-12-17 17:21:40","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":195053,"visible":true,"origin":"","legend":"\u003cp\u003eOverview of the study area,Nanchang City\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/2ed9dc6cb24b10a19c764b65.png"},{"id":71686710,"identity":"9c8453d5-e61e-4b75-a5d6-ce3dbb29a79d","added_by":"auto","created_at":"2024-12-17 17:29:40","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":67320,"visible":true,"origin":"","legend":"\u003cp\u003eInter-annual changes of NDVI in Nanchang City from 2000 to 2022\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/0263e61f2c9672f9773f62dd.png"},{"id":71685979,"identity":"039192e8-875e-4cc4-856d-e8883e339ab8","added_by":"auto","created_at":"2024-12-17 17:21:35","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":304808,"visible":true,"origin":"","legend":"\u003cp\u003eHurst index of NDVI in Nanchang from 2000 to 2010 and future trends\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/255fbc8cc9f89eb5fe89a04f.png"},{"id":71686707,"identity":"644d7528-0c92-4e56-be69-a2a6d53f9c83","added_by":"auto","created_at":"2024-12-17 17:29:35","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":237374,"visible":true,"origin":"","legend":"\u003cp\u003eHurst index of NDVI in Nanchang from 2011 to 2022 and future trends\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/66bb97696cbeb0f49a21fb72.png"},{"id":71685986,"identity":"e1b5d79b-2230-476e-9744-f7f529a88de5","added_by":"auto","created_at":"2024-12-17 17:21:40","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":277446,"visible":true,"origin":"","legend":"\u003cp\u003eHurst index of NDVI in Nanchang from 2000 to 2022 and future trends\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/8466df08562652a4784af890.png"},{"id":71685976,"identity":"d6373038-fb43-420a-a76c-9d301047c4eb","added_by":"auto","created_at":"2024-12-17 17:21:34","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":77626,"visible":true,"origin":"","legend":"\u003cp\u003eImportance of meteorological factors in ReliefF algorithm\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/367bef30e4a21ea941250e16.png"},{"id":71685988,"identity":"febf5708-b5a0-41a4-8406-472bcc568d5d","added_by":"auto","created_at":"2024-12-17 17:21:40","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":340766,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of the importance of meteorological and anthropogenic factors (a)Pearson’s correlation analysis (b)ReliefF algorithm\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/5cb72c63a31344a78d2780ee.png"},{"id":71685977,"identity":"3f3d7a0d-badf-4306-8b35-26b37423bad2","added_by":"auto","created_at":"2024-12-17 17:21:34","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":107218,"visible":true,"origin":"","legend":"\u003cp\u003eImpact on driving factors for NDVI in Nanchang City\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/4b5c54481c3b65691d137f79.png"},{"id":71686708,"identity":"9547db8f-6db6-45eb-befe-c7a4ddf0ad29","added_by":"auto","created_at":"2024-12-17 17:29:35","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":455840,"visible":true,"origin":"","legend":"\u003cp\u003eInteraction of driving factors in Nanchang City\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/ffffe9b4c3a1cafa1ce54802.png"},{"id":71685981,"identity":"526ff4ec-ade8-4b99-9ddb-1109f096eeb8","added_by":"auto","created_at":"2024-12-17 17:21:35","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":196127,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution between NDVI and air temperature and precipitation in Nanchang City during 2000-2022\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/70428f032e47639dce6f9ee3.png"},{"id":71685982,"identity":"e9c3d112-6307-4dda-b256-b15fc5737775","added_by":"auto","created_at":"2024-12-17 17:21:35","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":191672,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of partial correlation between NDVI and air temperature and precipitation in Nanchang City during 2000-2022\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/ad95ca3f47f36c7d962c2c6a.png"},{"id":71685978,"identity":"80c84a99-a947-4142-9224-847405784636","added_by":"auto","created_at":"2024-12-17 17:21:34","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":155862,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of multiple correlation between NDVI and anthropogenic factors in Nanchang City during 2000-2022\u003c/p\u003e","description":"","filename":"12.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/de835cfd429584162cf31c6f.png"},{"id":71685984,"identity":"0d212de3-99c3-4652-8fbf-175836c2c8f2","added_by":"auto","created_at":"2024-12-17 17:21:35","extension":"png","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":352407,"visible":true,"origin":"","legend":"\u003cp\u003eAreal and NDVI changes of different land cover types in Nanchang City, 2000-2022\u003c/p\u003e","description":"","filename":"floatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/89157dd204575307dae4d0c1.png"},{"id":83341872,"identity":"496a0429-f05e-4d95-8876-aaeccdf6aa7c","added_by":"auto","created_at":"2025-05-23 10:53:32","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4343473,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5366943/v1/b9e8b258-b0b2-4a70-895b-7686851ffe0d.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Spatio-temporal separating analysis of NDVI evolution and driving factors: a case study in Nanchang, China","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eVegetation has vital relevance in climate regulation, material cycling, information transmission and other aspects, which is dominating component in terrestical ecosystems (Terefe et al. \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Jin et al. \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The vegetation coverage in secular change reflexes the eco-environment not only the alterations, but also the influences in natural factors(such as weather) or human activities indirectly (Zhao et al. \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Therefore, monitoring vegetation dynamics which is in domain research has become one of the hot topics. Normalized Difference Vegetation Index(NDVI) is a key variable for monitoring climate change, studying ecological balance, and exploring regional phenological patterns. It has been widely used for estimating vegetation productivity and assessing ecological vulnerability (Wang et al. 2024; Das et al. 2024).\u003c/p\u003e \u003cp\u003eVegetation coverage monitor is crucial for ecological recover and governance. Many scholars have conducted extensive researches in spatio-temporal vegetation variations and its driving factors on the study area using different scales of NDVI. Previous scholars have used methods such as Sen\u0026thinsp;+\u0026thinsp;Mann-Kendall trend analysis, residual analysis, correlation analysis or partial correlation analysis, Hurst index, geographical detector, etc., to study the regional changes and driving factors of NDVI in different regions globally, including Africa (Yang et al. 2022), Europe (Novillo et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), China (Xu et al. \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), Loess Plateau (Li et al. 2021; Kong et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), Qinghai-Tibet Plateau (Wang et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), Yangtze River Delta (Tian et al. \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) and so on. Inter-annual vegetation prediction is discussed by Theil-Sen Median and MK-tendency analysis in the Yellow River Basin and acquiring preferably results (Cao et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Yang et al.(2019) investigated the evolution trend of vegetation, revealing the global NDVI trend from 1982 to 2015. The Hurst index can detect whether there is super-long periodicity in long time sequences and can simulate future vegetation growth trends (Ahmad et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Tong et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). For example, Qu et al.(2020) forecasted the future vegetation growth trends by Hearst index over the Yangtze River Basin, and Zhang et al.(2024) calculated the future growth trends of vegetation in Chengdu-Chongqing region with the Hurst exponent of high-dimensional fractals. However, so far, it has not been used for vegetation trend prediction in Nanchang City.\u003c/p\u003e \u003cp\u003eResearch on the driving forces of surface vegetation, it can be roughly divided into meteorological factors, topographic factors, human factors and other factors. Numerous scholars have used a range of analytical tools to explore the causes of vegetation change, including correlation analysis (Zhang et al. \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), linear and nonlinear regression models (Li et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), and geographical detector (Zhang et al. \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). And the same applies to machine learning in the screening of drivers.Teixeira et al.(2013) applied the RF algorithm to the application of predicting the standard enthalpy of formation of hydrocarbons by downscaling the high-dimensional feature data selectively; Tang et al.(2018) utilized MIC algorithm for linear and nonlinear correlation acquisition for bearing fault diagnosis in feature selection algorithm; Zhang et al.(2016) performed electroencephalogram sensor sentiment recognition based on ReliefF feature selection algorithm, which quantified the importance of each sensor channel by obtaining all the feature weights of the analyzed channels through the algorithm.\u003c/p\u003e \u003cp\u003ePrevious studies had often failed to adequately consider the role of soil temperature when considering the effects of meteorological factors on vegetation growth (Luo et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), and had also neglected the potential effects in economic level of population (Zhang et al. \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), among others, on NDVI.Most of these studies used qualitative methods to analyze spatial and temporal changes together (Zhong et al. 2018; Sun et al. 2019), and lack in-depth exploration of how much the climate change and human activities quantitatively affect vegetation changes.In addition, there was a relative lack of research on the sustained development and future vegetation growth trend in Nanchang City.\u003c/p\u003e \u003cp\u003eIn manuscript, we temporally and spatially divide up the NDVI evolution and driver analysis of Nanchang City from 2000 to 2022 systematically. Hurst Exponent, Sen\u0026thinsp;+\u0026thinsp;Mann-Kendall trend analysis and other methods are used to quest its evolutionary characteristics and trends of NDVI; ReliefF, MIC, RF algorithms are used for the selection of meteorological driving factors of NDVI, and geographical detector and complex correlation analysis can quantitatively study the impact of anthropogenic and meteorological factors to NDVI. The central objective of this study is to divide the analysis of vegetation evolution and its drivers into two independent dimensions, temporal and spatial, and to explore each dimension in detail.In the temporal dimension, we focus on the dynamic changes of NDVI over time, and reveal its long-term evolution patterns through trend analysis. In the spatial dimension, we examine vegetation conditions in different geographic locations to explore the effects of spatial heterogeneity on vegetation distribution and pattern. Through this detailed spatial and temporal segmentation method, we can more accurately grasp the pattern of vegetation change, and more precisely identify the key factors affecting the evolution of vegetation. It can bring a scientific foundation in ambulatory monitoring and vegetation protecting in Nanchang City, as well as theoretical guidance for the formulation and implementation of sustainable advancement plans.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1.Study area\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eNanchang, locating at north-central of Jiangxi Province, East China (115\u0026deg;27'E-116\u0026deg;35'E, 28\u0026deg;09'N-29\u0026deg;11'N), is the core city at Poyang Lake ecological economic zone (Zhu et al. 2023), with be in Ganjiang River, Fuhe River downstream (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). In the mean time, it is the only provincial capital designated with low-carbon economic development pilot city, which is important comprehensive demonstration area for ecological cities (Liu et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Nanchang City is a subtropical monsoon climate zone, characterized by both rain and heat in the same season, with four distinct seasons and abundant rainfall, making it one of the rainy areas in China (Zhu et al. 2023). There are notable differences in the distribution of rainfall, which can easily lead to droughts or flood disasters. Nanchang has abundant resources which is about ecological, including a part of Poyang Lake. For the past few years,the number of occurrences in extreme weather events continue to increase because of the ecological destruction (Zhang et al. \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), exacerbating the impact on NDVI evolution. In the preliminary statistics from the Nanchang City Disaster Reduction Committee, the 3-hour rainfall of the two extremely heavy rainstorms in 2012 exceeded 100mm, more than the overall rainfall of the August in previous years. 298.5 hectares of crops area had been affected, with pecuniary losses exceeding tens of millions (Jiangxi Bureau of Statistics 2021).\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2.Data sources\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe research data mainly contain MODIS13 NDVI, meteorological, terrain, land cover type, and anthropogenic data (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). ArcGIS 10.8 was used to perform spatial transformating, masking and clipping, grid creation, and multi-value extraction to points for modeling and detecting each year's data; when performing data resampling, we standardized the resolution to 1km; Matlab R2018b was used to standardize and normalize modeling of long time series data, and to perform machine learning model-driven analysis.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eData source and data preprocessing.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eData Type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCode\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eData Content\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTime resolution\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSpatial resolution\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.Data Source\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eProduct\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRemote sensing data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eY\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMODIS NDVI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1km/500m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNASA website(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://search.earthdata.nasa.gov/search\u003c/span\u003e\u003cspan address=\"https://search.earthdata.nasa.gov/search\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDidan, K.(2015)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eMeteorological data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eX1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAnnual mean temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNational Tibetan Plateau / Third Pole Environment Data Center.(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.tpdc.ac.cn/zh-hans/data/\u003c/span\u003e\u003cspan address=\"https://data.tpdc.ac.cn/zh-hans/data/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePeng, S.(2019)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eX2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTotal annual precipitation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eThird Pole Environment Data Center.\u0026nbsp;(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.tpdc.ac.cn/\u003c/span\u003e\u003cspan address=\"https://data.tpdc.ac.cn/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u0026nbsp;)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePeng, S.(2020)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eX3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEvaporation and transpiration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNASAwebsite(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://ladsweb.modaps.eosdis.nasa.gov/\u003c/span\u003e\u003cspan address=\"https://ladsweb.modaps.eosdis.nasa.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMODIS16A2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA1-A24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMeteorological data from weather stations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003emonth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eChina Meteorological Data Service Centre(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.cma.cn/\u003c/span\u003e\u003cspan address=\"https://data.cma.cn/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eData from the National Climatic Observatory, Nanchang\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eAnthropogenic data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eX4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReal Gross Domestic Product(GDP)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFigshare website\u003c/p\u003e \u003cp\u003e(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.6084/m9.figshare.17004523.v1\u003c/span\u003e\u003cspan address=\"10.6084/m9.figshare.17004523.v1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u0026nbsp;)(2000-2015)/GitHub website(\u0026nbsp;\u0026nbsp;\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/thestarlab/ChinaGDP)(2016-2020\u003c/span\u003e\u003cspan address=\"https://github.com/thestarlab/ChinaGDP)(2016-2020\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eZhao, Liu.(2017)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eX5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePopulation density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e100m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eWorldPop website(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.worldpop.org/\u003c/span\u003e\u003cspan address=\"https://www.worldpop.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eX6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNighttime lights\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1km\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eHARVARD Dataverse (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://dataverse.harvard.edu/\u003c/span\u003e\u003cspan address=\"https://dataverse.harvard.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eWu, Yizhen.(2022)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eA25-A37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUrban statistics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNational Bureau of Statistics of China\u003c/p\u003e \u003cp\u003e(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.stats.gov.cn/\u003c/span\u003e\u003cspan address=\"https://www.stats.gov.cn/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTopographic data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eX7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eelevation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGeospatial Data Cloud(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gscloud.cn/\u003c/span\u003e\u003cspan address=\"https://www.gscloud.cn/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eASTER GDEM 30M\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLand cover data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eX8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLand cover type\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eResource and Environmental Science Data Platform (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.resdc.cn/\u003c/span\u003e\u003cspan address=\"https://www.resdc.cn/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eXinliang Xu,(2018)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eNote: average temperature (A1), average maximum temperature (A2), average minimum temperature (A3), extreme maximum temperature (A4), extreme minimum temperature (A5), precipitation (A6), average pressure (A7), average relative humidity (A8), sunshine hours (A9), maximum wind speed and direction (A10), average surface temperature (A11), average maximum surface temperature (A12), average minimum surface temperature (A13), extreme maximum ground temperature (A14), extreme minimum ground temperature (A15), average ground temperature at 5 cm depth (A16), average ground temperature at 10 cm depth (A17), average ground temperature at 15cm depth (A18), average ground temperature at 20cm depth (A19), average ground temperature at 40cm depth (A20), average ground temperature at 0.8m depth (A21), average ground temperature at 1.6m depth (A22), average ground temperature at 3.2m depth (A23), E601 large-scale evaporation (A24), GDP (A25), value added of primary industry (A26), value added of secondary industry (A27), value added of tertiary industry (A28), value added of the primary industry as a proportion of GDP (A29), value added of secondary industry as a proportion of GDP (A30), value added of tertiary industry as a proportion of GDP (A31), employee number in the primary industry (A32), employee number in the secondary industry (A33), employee number in the tertiary industry(A34),employee proportion in the primary industry (A35), employee proportion in the secondary industry (A36), employee proportion in the tertiary industry (A37).\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3.Research methods\u003c/h2\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e2.3.1.Theil-Sen Median and Mann-Kendall trend analysis\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eTheil-Sen belongs to non parametric calculation methods (Wang et al. \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2009\u003c/span\u003e), Which is efficient, robust and apposite for trend analysis in long time series. Time series do not need to satisfy assumptions such as autocorrelation or normal distribution. It can effectively handle outliers and missing values. The Theil-Sen slope represents the positive or negative relationship between data points in a time series, with median slope representing the whole timely trendency (Adichie et al. 1967). The computing formula is:\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ1\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:{\\beta\\:}=\\text{M}\\text{e}\\text{d}\\text{i}\\text{a}\\text{n}\\left(\\frac{{\\text{x}}_{\\text{j}}-{\\text{x}}_{\\text{i}}}{\\text{j}-\\text{i}}\\right)\\:\\:\\:\\:\\forall\\:\\text{j}\u0026gt;\\text{i}$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eAbove it \u003cem\u003eβ\u003c/em\u003e represents the slope for determining the trend of NDVI changes in Nanchang City; Median represents the median function; \u003cem\u003ej\u003c/em\u003e and \u003cem\u003ei\u003c/em\u003e are time indices;\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{x}_{j}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{x}_{i}\\)\u003c/span\u003e\u003c/span\u003e stand for the actual NDVI for the \u003cem\u003ej\u003c/em\u003eth and \u003cem\u003ei\u003c/em\u003eth years, respectively. At the time \u003cem\u003eβ\u0026thinsp;\u0026gt;\u0026thinsp;0\u003c/em\u003e, NDVI is upwarding in the period; at the time \u003cem\u003eβ\u0026thinsp;=\u0026thinsp;0\u003c/em\u003e, NDVI remains constant in the period; when \u003cem\u003eβ\u0026thinsp;\u0026lt;\u0026thinsp;0\u003c/em\u003e, NDVI shows a decreasing trend during this time period.\u003c/p\u003e \u003cp\u003eMann-Kendall analysis is non-parametric which the World Meteorological Organization recommended and be widely adopted, and can effectively discriminate natural process is natural fluctuation or definite changing trend (Gadedjisso et al. 2021). There is no need that the test samples have to meet a certain distribution pattern rules, while outliers will not interfere with the test results (Khaled et al. 2008). Calculate the Theil-Sen slope and then use MK analysis to check the changing trend. When the significance level is \u003cem\u003eα\u003c/em\u003e, if the normal distribution statistic(|Z\u003csub\u003eC\u003c/sub\u003e|) satisfies |Z\u003csub\u003eC\u003c/sub\u003e|\u0026gt;Z\u003csub\u003e1-α/2\u003c/sub\u003e, the time series change is statistically significant. When the |Z\u003csub\u003eC\u003c/sub\u003e|\u0026ge;1.28, it passes the 90% confidence level significance test; when the statistic satisfies |Z\u003csub\u003eC\u003c/sub\u003e|\u0026ge;1.64, it passes 95% confidence interval significance test (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMann-Kendall test method significance statistics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSignificance test\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRange of \u003cem\u003eZ\u003c/em\u003e\u003csub\u003e\u003cem\u003eC\u003c/em\u003e\u003c/sub\u003e value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eChange trend\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cem\u003e|Z\u003c/em\u003e\u003csub\u003e\u003cem\u003eC\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e|\u003c/em\u003e\u0026ge;1.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026minus;1.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSignificant reduction\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;1.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSignificant increase\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e1.64\u0026ge;\u003cem\u003e|Z\u003c/em\u003e\u003csub\u003e\u003cem\u003eC\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e|\u003c/em\u003e\u0026ge;1.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;1.64～\u0026minus;1.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNot significant reduction\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.28～1.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNot significant increase\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003e|Z\u003c/em\u003e\u003csub\u003e\u003cem\u003eC\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e|\u003c/em\u003e\u0026le;1.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;1.28～1.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUnchanged\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e2.3.2.Hurst Exponent\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe study of the Hurst exponent(H) based on rescaled range analysis(R/S) was proposed by the British hydrologist Hurst (Tabares et al. 2021), which can be a index that judge time series data is random walk or biased random walk process. The Hurst exponent can describe the persistence or anti-persistence of NDVI changes (Zhang et al. \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Therefore, this study calculates the Hurst in R/S analysis method to forecast the vegetation NDVI evolution trend in Nanchang City. Time series (\u003cem\u003et\u003c/em\u003e),\u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1,2,\u0026hellip;\u0026hellip;, where N\u003csub\u003eDVI\u003c/sub\u003e(\u003cem\u003et\u003c/em\u003e)\u0026thinsp;=\u0026thinsp;B(\u003cem\u003et\u003c/em\u003e)-B(\u003cem\u003et-1\u003c/em\u003e), B(\u003cem\u003et\u003c/em\u003e) being the observed value at moment \u003cem\u003et\u003c/em\u003e, for any positive integer τ greater than or equal to 1 (Chen et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), define the sequence of mean values as\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ2\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$\\:{\\text{N}}_{\\text{D}\\text{V}\\text{I}\\tau\\:}=\\frac{1}{\\tau\\:}\\sum\\:_{t=1}^{\\tau\\:}{\\text{N}}_{\\text{D}\\text{V}\\text{I}}\\left(t\\right),\\tau\\:=\\text{1,2},\\dots\\:\\dots\\:$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eCumulative deviation series X(\u003cem\u003et\u003c/em\u003e,\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\tau\\:\\)\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ3\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ3\" name=\"EquationSource\"\u003e\n$$\\:\\sum\\:_{t=1}^{\\tau\\:}\\left(\\xi\\:\\right(t)-{\\xi\\:}_{\\tau\\:}),1\\le\\:\\text{t}\\le\\:\\tau\\:$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e3\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe polar sequence R(\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\tau\\:\\)\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ4\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ4\" name=\"EquationSource\"\u003e\n$$\\:\\text{R}\\left(\\tau\\:\\right)=\\underset{1\\le\\:\\text{t}\\le\\:\\tau\\:}{\\text{max}}X(t,\\tau\\:)-\\underset{1\\le\\:\\text{t}\\le\\:\\tau\\:}{\\text{min}}X(t,\\tau\\:),\\tau\\:=\\text{1,2},\\dots\\:\\dots\\:$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e4\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eStandard deviation sequence S(\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\tau\\:\\)\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ5\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ5\" name=\"EquationSource\"\u003e\n$$\\:\\text{S}\\left(\\tau\\:\\right)={\\left[\\frac{1}{\\tau\\:}\\sum\\:_{\\text{t}-1}^{\\tau\\:}{\\left({\\text{N}}_{\\text{D}\\text{V}\\text{I}}\\right(\\text{t})-{\\text{N}}_{\\text{D}\\text{V}\\text{I}\\tau\\:})}^{2}\\right]}^{\\frac{1}{2}},\\tau\\:=\\text{1,2},\\dots\\:\\dots\\:$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e5\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eHurst exponent calculation:\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ6\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ6\" name=\"EquationSource\"\u003e\n$$\\:\\frac{\\text{R}\\left(\\tau\\:\\right)}{\\text{S}\\left(\\tau\\:\\right)}≌\\frac{R}{S}\\propto\\:\\tau\\:H$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e6\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIf Eq.\u0026nbsp;\u003cspan refid=\"Equ6\" class=\"InternalRef\"\u003e6\u003c/span\u003e holds, it implies the existence of the Hurst phenomenon with time series N\u003csub\u003eDVI\u003c/sub\u003e(\u003cem\u003et\u003c/em\u003e),t\u0026thinsp;=\u0026thinsp;1,2,\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\dots\\:\\dots\\:\\)\u003c/span\u003e\u003c/span\u003e. The Hurst shows as H with range between 0\u0026ndash;1. At that time 0\u0026thinsp;\u0026lt;\u0026thinsp;H\u0026thinsp;\u0026lt;\u0026thinsp;0.5, it is inverse sustainability that future trend of N\u003csub\u003eDVI\u003c/sub\u003e is; when H\u0026thinsp;=\u0026thinsp;0.5, it shows that there is no correlation between future and past trends in the N\u003csub\u003eDVI\u003c/sub\u003e time series; when 0.5\u0026thinsp;\u0026lt;\u0026thinsp;H\u0026thinsp;\u0026lt;\u0026thinsp;1, this states the future trend will keep pace with the past (Alvarez et al. 2008).\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003e2.3.3.Machine learning feature selection algorithms\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003e(1)The ReliefF algorithm\u003c/p\u003e \u003cp\u003eThe ReliefF algorithm was proposed as an extension of the original Relief algorithm. Based on traditional filter-based feature selection method, it can reinforce learning algorithm by selecting the feature subset with the highest information content from the original feature set (Ryan et al. 2018). The ReliefF algorithm assigns different weights to driving factors by calculating the discriminative ability of features for each nearby sample. Due to its efficiency and effectiveness, it has been widely used in the field of feature selection. However, some of its variants and improved algorithms also focus on dealing with specific issues, such as handling multi-label data or applications on high-dimensional small sample data (Spola\u0026ocirc;r et al. 2014).\u003c/p\u003e \u003cp\u003e(2)The MIC algorithm\u003c/p\u003e \u003cp\u003eThe MIC algorithm can estimate the correlation between two variables and identify linear and it can non-linear relationships (Li et al. 2022). The MIC algorithm can not only quantify the strength of the correlation between variables, but also identify patterns of correlation between variables. This algorithm provides a normalization method that allows for quantifiable comparison of results and it is insensitive to the discretization of data (Li et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). When dealing with large datasets, the MIC algorithm can provide similar coefficients for different types of relationships with similar noise level, selecting the driving factor with the highest degree of correlation with the target variable (Ge et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e(3)The RF algorithm\u003c/p\u003e \u003cp\u003eRF model integrates decision trees with Bagging algorithm for improving the generalization ability, with in a common use at machine learning (Breiman et al. 2001). The random forest algorithm which is easily come ture, as an ensemble classifier, performs well in regression, classification, and other problems (Iratxe et al. 2021).\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e2.3.4.Geographical detector\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eGeographical detector is a primary tool in spatial stratified heterogeneity analysis. The heterogeneity of a single variable and the distribution coupling of two variables can be gauged by geographical detector in the meantime in case to find out the possible relationships and explore numerical and qualitative data between them (Song et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e(1)Factor detector\u003c/p\u003e \u003cp\u003eAs one of the geographical detector Model, critical analysis in factor detector focuses on: at a specific spatial location, the geographic entities and spatial variability in environmental factors are always exist (Wu et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). In spatial, if the alters in geographic entities at a certain point are consistent with the changes in environmental factors, this environmental factor is prominent in the alteration of geographic entities.\u003c/p\u003e \u003cp\u003eQ-value is an indicator to measure the spatial variability of \u003cem\u003eY\u003c/em\u003e or the degree of the independent variable \u003cem\u003eX\u003c/em\u003e on the attribute \u003cem\u003eY\u003c/em\u003e, shown as:\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ7\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ7\" name=\"EquationSource\"\u003e\n$$\\:q=1-\\frac{\\sum\\:_{ℎ=1}^{L}{N}_{ℎ}{\\sigma\\:}_{ℎ}^{2}}{N{\\sigma\\:}^{2}}=1-\\frac{SSW}{SST}$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e7\u003c/div\u003e\u003c/div\u003e \u003cdiv id=\"Equ8\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ8\" name=\"EquationSource\"\u003e\n$$\\:SSW=\\sum\\:_{ℎ=1}^{L}{N}_{ℎ}{\\sigma\\:}_{ℎ}^{2},SST=N{\\sigma\\:}^{2}$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e8\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003ewhere \u003cem\u003eh\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1, \u0026hellip;\u0026hellip;, \u003cem\u003eL\u003c/em\u003e is the lamination of the variable \u003cem\u003eY\u003c/em\u003e or factor \u003cem\u003eX\u003c/em\u003e; \u003cem\u003eN\u003c/em\u003e\u003csub\u003eh\u003c/sub\u003e and \u003cem\u003eN\u003c/em\u003e are the number of cells in layer \u003cem\u003eh\u003c/em\u003e and the whole area; \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{{\\sigma\\:}}_{\\text{ℎ}}^{2}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\sigma\\:}^{2}\\)\u003c/span\u003e\u003c/span\u003eare the variances of the \u003cem\u003eY\u003c/em\u003e-values for layer \u003cem\u003eh\u003c/em\u003e and the whole region, respectively. SSW and SST are the within-stratum sum of squares and the district-wide total sum of squares, respectively (Wang et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e(2)Interaction detector\u003c/p\u003e \u003cp\u003eThe general method for interaction detector is adding up the product of regression model and statistical significance (Ren et al. 2014). We can be sure if it exist an interaction effect between factors, even the direction, strength, linearity, with analyzing q-values of each factor and multiple factors (Li et al. 2023).\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e2.3.5.Correlation analysis\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003e(1)Single correlation analysis\u003c/p\u003e \u003cp\u003eThe one-way correlation analysis in image measurement scale NDVI was performed separately with the factors that were significantly correlated to the NDVI as calculated by the geoprobe with the following formula:\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ9\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ9\" name=\"EquationSource\"\u003e\n$$\\:{R}_{xy}=\\frac{\\sum\\:_{i=1}^{n}\\left[\\left({x}_{i}-\\stackrel{-}{x}\\right)\\left({y}_{i}-\\stackrel{-}{y}\\right)\\right]}{\\sqrt{\\sum\\:_{i}^{n}{\\left({x}_{i}-\\stackrel{-}{x}\\right)}^{2}}\\sqrt{\\sum\\:_{i=1}^{n}{\\left({y}_{i}-\\stackrel{-}{y}\\right)}^{2}}}$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e9\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{xy}\\)\u003c/span\u003e\u003c/span\u003e denotes the correlation coefficient of \u003cem\u003ex\u003c/em\u003e and \u003cem\u003ey\u003c/em\u003e; \u003cem\u003ex\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e denotes NDVI value in \u003cem\u003ei\u003c/em\u003e year; \u003cem\u003ey\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e denotes the value of the driver in \u003cem\u003ei\u003c/em\u003e year; \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\stackrel{-}{x}\\:\\)\u003c/span\u003e\u003c/span\u003eand\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\:\\stackrel{-}{y}\\)\u003c/span\u003e\u003c/span\u003e denote the mean value of NDVI and driver in \u003cem\u003en\u003c/em\u003e years, respectively.\u003c/p\u003e \u003cp\u003e(2)Partial correlation analysis\u003c/p\u003e \u003cp\u003eIntroducing partial correlation coefficient to represent the correlation between two factors, studying the impact of driving factors on NDVI while excluding interference from other factors. Expression is following (Dror et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2014\u003c/span\u003e):\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ10\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ10\" name=\"EquationSource\"\u003e\n$$\\:{R}_{xy,z}=\\frac{{R}_{xy}-{R}_{xz}{R}_{yz}}{\\sqrt{(1-{R}_{xz}^{2})(1-{R}_{yz}^{2})}}$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e10\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn the formula, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{xy,z}\\)\u003c/span\u003e\u003c/span\u003e stands for the partial correlation between NDVI and factor 1 when factor 2 is held constant, excluding the influence of factor 2 in analyzing the correlation between NDVI and factor 1; \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{xy}\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{xz}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{yz}\\)\u003c/span\u003e\u003c/span\u003e each represent the correlation coefficients between NDVI and factor 1, NDVI and factor 2 ,and factor 1 and factor 2. This method can be tested the significance of partial correlation coefficients. Expression is following (Lakens et al. 2017):\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ11\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ11\" name=\"EquationSource\"\u003e\n$$\\:\\text{t}=\\frac{{R}_{xy,z}}{\\sqrt{1-{R}_{xy,z}^{2}}}\\sqrt{\\text{n}-\\text{m}-1}$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e11\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn the formula: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{xy,z}\\)\u003c/span\u003e\u003c/span\u003e represents the partial correlation coefficient; \u003cem\u003en\u003c/em\u003e represents the sample size; \u003cem\u003em\u003c/em\u003e represents the degrees of freedom.\u003c/p\u003e \u003cp\u003e(3)Multiple Correlation analysis\u003c/p\u003e \u003cp\u003eTaking into account the compositive effects in multiple factors, the introduction in multiple correlation is used to calculate the correlation extent between multiple factors (Qiu et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Expression is following:\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ12\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ12\" name=\"EquationSource\"\u003e\n$$\\:{R}_{x,yz}=\\sqrt{1-(1-{R}_{xy}^{2})(1-{R}_{xz,y}^{2})}$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e12\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn the formula: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{x,yz}\\)\u003c/span\u003e\u003c/span\u003e is the multiple correlation analysis between NDVI and factors 1,2; \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{xy}\\)\u003c/span\u003e\u003c/span\u003e is the correlation between NDVI and factor 1; \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{xz,y}\\)\u003c/span\u003e\u003c/span\u003e is the partial correlation between NDVI and factor 2 when factor 1 is held constant. Here, the F-tests method has been inspected the significance of partial correlation coefficient (Quirk et al. 2020). Expression is following:\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Equ13\" class=\"Equation\"\u003e \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ13\" name=\"EquationSource\"\u003e\n$$\\:\\text{F}=\\frac{{R}_{xy,z}^{2}}{1-{R}_{x,yz}^{2}}\\times\\:\\frac{\\text{n}-\\text{k}-1}{\\text{k}}$$\u003c/div\u003e \u003cdiv class=\"EquationNumber\"\u003e13\u003c/div\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn the formula: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}_{x,yz}\\)\u003c/span\u003e\u003c/span\u003e represents the multiple correlation analysis; \u003cem\u003en\u003c/em\u003e represents the sample size; \u003cem\u003ek\u003c/em\u003e represents the number of independent variables.\u003c/p\u003e \u003cp\u003e(4)Pearson\u0026rsquo;s correlation coefficient\u003c/p\u003e \u003cp\u003eThe degree of linear relationship between two variables can be assessed by it (Kirch et al. 2008). The range of it is -1 to 1,-1 is in perfect negative correlation, while \u0026minus;\u0026thinsp;1 is contrariwise, and 0 is irrelevant.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.1.Spatial-temporal evolution characteristics of NDVI in Nanchang City\u003c/h2\u003e \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e \u003ch2\u003e3.1.1.Temporal Evolution Characteristics of NDVI\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe study introduces a concept of image average to quantitatively characterize the vegetation evolution in Nanchang over a period of time. The month-by-month NDVI values from 2000 to 2022 in Nanchang City were extracted for integrated analysis. From May to October,the vegetation will grow in Nanchang, and the monthly average of NDVI fluctuated from 0.254 to 0.656 during the 23-year period. The annual average NDVI values fluctuated in the range of 0.412 to 0.487, which 2013 and 2015 have the min. and max. annual valus. The multi-year average NDVI value is 0.453 (dashed line in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), and years with NDVI values exceeding the multi-year average are generally concentrated after 2014. From Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, it can be observed that the overall trend of NDVI change shows an upward trend, but the rate of increase is gradually slowing down.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e \u003ch2\u003e3.1.2.Temporal Evolution\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eFor studying the spatial changes of NDVI in Nanchang City, Theil-Sen Median based on MK trend analysis is conducted in pixels, which obtain spatial distribution of the change rate in NDVI and spatial distribution of the Mann-Kendall test statistic. Combining spatial distribution of Hurst index in Nanchang calculated pixel by pixel, the relationship between the future direction and the past tendency of vegetation cover change in Nanchang is obtained. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, from 2000 to 2010, the H-index in Nanchang ranged from 0.165 to 0.988, with a mean of 0.542. During this period, the overall vegetation change showed a continuous development trend. From 2011 to 2022, the H-index was from 0.149 to 0.978, with a mean of 0.522. Although the overall change still showed a consistent development trend, the regions with reverse development significantly increased (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). For the city of Nanchang from 2000 to 2022, the overall Hurst index ranges from 0.124 to 0.9, with a mean of 0.444. 26.068% of the city has the NDVI Hurst is more than 0.5, while 73.932% zone has NDVI Hurst less than 0.5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). The results indicate that over the 23 years, the overall vegetation change in Nanchang City exhibits characteristics of reverse sustained development. The closer the Hurst index is to 0.5, the less correlation there is between the persistence of future changes and past trends. Therefore, based on the direction and strength of persistence, this study classifies the Hurst index into 4 categories: strong reverse persistence(0\u0026thinsp;\u0026lt;\u0026thinsp;H\u0026thinsp;\u0026lt;\u0026thinsp;0. 35)、weak reverse persistence(0. 35\u0026thinsp;\u0026lt;\u0026thinsp;H\u0026thinsp;\u0026lt;\u0026thinsp;0. 50)、weak co-directional persistence(0. 50\u0026thinsp;\u0026lt;\u0026thinsp;H\u0026thinsp;\u0026lt;\u0026thinsp;0. 65)、strong persistence in the same direction(0. 65\u0026thinsp;\u0026lt;\u0026thinsp;H\u0026thinsp;\u0026lt;\u0026thinsp;1). With superimposing the sustainability index(H)in spatial, the future development tendency of NDVI can be acquired (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eFrom 2000 to 2022, the NDVI trend in the first 11 years developed in a positive direction (improvement direction accounting for 53.4%), while in the following 12 years it developed in a negative direction (degradation direction accounting for 54.6%), with a similar proportion of improvement and degradation. Based on Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, the analysis of the overall changes for 23 years indicates that the future development trend of Nanchang City may move towards a negative direction (degradation direction accounting for 64.9%), with areas of continuous degradation accounting for 12% (strongly continuous degradation accounting for 1.8%, weakly continuous degradation accounting for 10.2%). Areas that have improved in the past but are projected to degrade in the future account for 52.9% (strongly reversing improvement accounting for 10.8%, weakly reversing improvement accounting for 42.1%); areas of continuous improvement account for 15.9% (weakly continuous improvement and strongly continuous improvement accounting for 15.1% and 0.8% respectively), and areas that degraded in the past but are expected to show improvement in the future account for 18.9% (strongly reversing degradation and weakly reversing degradation accounting for 15.7% and 3.2% respectively). In Nanchang City, the areas with continuous amelioration are scattered in central and eastern district, with the continuously improved vegetation cover type being farmland. The agricultural vegetation in Nanchang City is well protected and will be sustained growth. The areas with continuous degradation are unevenly distributed, with regions of strong continuous degradation concentrated near constructive land, and regions of weak continuous degradation clustered near urban land and rivers, indicating that human activities have had adverse effects on vegetation cover. However, the predicted proportion areas with continuous degradation is small, mainly concentrated within urban areas where human activities are well controlled and have not affected natural ecological areas such as forests and grasslands.Regions that have degraded in the past but show a trend of improvement in the future do not have a clear spatial distribution pattern, mainly existing on both sides of rivers, around lakes and sporadically on the outskirts of cities, indicating that human activities have restricted urban expansion and have focused on protecting and rebuilding ecological red zones. In the future, vegetation cover around lakes, riverbanks and in suburban areas is likely to increase. Regions that have improved in the past but are expected to degrade latter are mainly situated in the southeastern and western parts of Nanchang City, including the counties of Jinxian, Anyi, and Xinjian. Jinxian County, with its low hills in south and lakeshore in north, has a severe ecological and environmental pollution situation because it is a supply base for high-quality agricultural and sideline products in Jiangxi Province; Anyi County is polluted by brick kilns, and Xinjian County is polluted by mining enterprises, leading to significant vegetation degradation.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eFuture change trend of vegetation NDVI in Nanchang City\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChange directions\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFuture change trend\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2000\u0026ndash;2010 Percentage\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2011\u0026ndash;2022 Percentage\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2000\u0026ndash;2022 Percentage\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eContinuous degradation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eStrong persistent degradation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e12.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.8%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWeak persistent degeneration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e25.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eThe past has improved, but the future is a degradation trend\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eContinuous improvement of anti-strong forces\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10.8%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eContinuous improvement of anti-weakness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e31%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e42.1%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eDegradation in the past but improvement in the future\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAnti-weakness and persistent degradation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e15.7%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePersistent degradation of anti-strength\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.5%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eContinuous improvement\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWeak sustained improvement\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e19.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e15.1%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eStrong continuous improvement\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.8%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBasically unchanged\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBasically unchanged\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eNote: In the table, the outliers for each region percentage have been removed.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.2.NDVI Temporal-Driven Analysis\u003c/h2\u003e \u003cdiv id=\"Sec16\" class=\"Section3\"\u003e \u003ch2\u003e3.2.1.Selection of meteorological factors\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThis study uses the image averaging method to extract the monthly average NDVI values of Nanchang City from 2000 to 2022. Regression models are established by ReliefF algorithm, MIC algorithm and RF algorithm to calculate the impact weights of various meteorological factors on NDVI. Through comprehensive analysis of the fitting degree in training set and test set, the ReliefF algorithm is found to build a regression model with the closest R\u003csup\u003e2\u003c/sup\u003e to 1 and the smallest error, which is more suitable in calculating the driving weights of meteorological factors on NDVI among the three algorithms. (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eEvaluation factors of three machine learning models\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReliefF\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMIC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRF\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eTraining set\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eR\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.963\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.946\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.956\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMAE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.017\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.019\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMBE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;0.005\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRMSE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.027\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.025\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eTest set\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eR\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.934\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.918\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.914\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMAE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.026\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.027\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMBE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;0.005\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRMSE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.029\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.033\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.033\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe regression model by the ReliefF algorithm evaluates the correlation between meteorological factors A1-A24 (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and NDVI, and obtaines the importance of each meteorological factor (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). Screening out the top ten meteorological factors with the highest proportion of NDVI driving weights, they are A23,A20,A22,A3,A18,A19,A21,A13,A17 and A16. Among them,underground temperature has a particularly prominent driving effect on NDVl. The underground temperature, i.e. soil temperature, can promote the absorption of nutrients by vegetation rhizomes and affect the accumulation of vegetation biomass, it affects soil evaporation and vegetation transpiration, which in turn affects the water supply and growth of vegetation. This temperature influences the metabolic activities of soil microorganisms, providing necessary nutrients for vegetation. At the same time, the underground temperature will affects not only the growth processes of vegetation such as photosynthesis and respiration, but also the growth cycle of vegetation, including stages such as germination, flowering and fruiting.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section3\"\u003e \u003ch2\u003e3.2.2.Comparison of the importance of meteorological factors and anthropogenic factors\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eConduct a correlation analysis between the ten meteorological factors screened above and anthropogenic factors, and perform ReliefF feature selection analysis. It is obvious that anthropogenic factors are significantly more important to NDVI than meteorological factors (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e). With social and economic developments, the increase in population density, acceleration of urbanization and expansion of farmland lead to a decrease in vegetation coverage, while changes in policy orientation and people's environmental awareness also affect vegetation inversion. Although meteorological factors can also affect vegetation coverage, the human activities on meteorological factors are more apparent than the past, thereby creating complex effects on vegetation growth. Climate change is a long-term cumulative natural process, with changes in meteorological factors such as rainfall and temperature usually slow and periodic, but the impact of anthropogenic factors is rapid and severe. So as time goes on, the impact of anthropogenic factors are more significant than meteorological factors.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.3.NDVI Spatial-Driven Analysis Temporal-Driven Analysis\u003c/h2\u003e \u003cdiv id=\"Sec19\" class=\"Section3\"\u003e \u003ch2\u003e3.3.1.Driving analysis of NDVI derived from geographical detector\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003e(1)Factor detection\u003c/p\u003e \u003cp\u003eThis manuscript uses geographic detector to reveal the interpretive force of driving factors in Nanchang spatially differentiation in NDVI(q). As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e, factors X1-X8 (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) do not all significantly affect the evolution of NDVI. The blue frame bar indicates that the q value of the driver is significant (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Under the condition of considering the significance in driving factors, the filtered factors are ranked in the following order of importance: X5\u0026thinsp;\u0026gt;\u0026thinsp;X6\u0026thinsp;\u0026gt;\u0026thinsp;X8. Similar to the results driven by time, the human factors are significant while the meteorological factors are not significant. Considering other driving factors, land cover types are more significant than elevation. Among the anthropogenic factors, the influence of X5 and X6 factors, spatial distribution in vegetation cover is more than 30%, with the explanatory power of X5 being over 90%.\u003c/p\u003e \u003cp\u003eConsidering other driving factors, land cover types are more significant than elevation. Among the anthropogenic factors, two driving factors, X5 and X6, have an impact of over 30% on spatial distribution of vegetation cover, with the explanatory power of X5 being over 90%. X8 are a secondary driving factor, with an average explanatory power over 20%. Among the eight driving factors, the factors that are not significantly explanatory for NDVI are: X1, X2, X3, X4 and X7. More than 75% of the study area has the following characteristics: an elevation of 30 meters, a slope of less than 4\u0026deg;, which is being an area with intensive human production activities. The main reason for vegetation evolution is the resource development and utilization led by human activities and land cover types\u0026rsquo; variation. With the increase in population and the development of the economy, there is a growing demand for food and cash crops, leading to the conversion in some forests to farmland. Urban construction and expansion also occupy farmland, forests and grasslands, leading to large-scale deforestation. These variations in land use patterns result in the degradation of NDVI. Deforestation and farming activities had led to soil erosion and degradation, exacerbating the phenomenon of soil erosion. Biodiversity and ecosystem diversity have been damaged, even abandoned farmland, mainly in the low hills and mountains of Anyi, Jinxian, and Xinjian County. Once over-exploited land is abandoned, such as land that has been deforested or reclaimed from lakes, the natural vegetation can not recover in a short time, leaving the surface exposed and increasing the risk of soil sanding and desertification. This is mainly distributed in northern Nanchang County, the northern and southern parts of Xinjian County, and the low hill areas of Jinxian County.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003e(2)Interaction detection\u003c/p\u003e \u003cp\u003eInteraction detection can mainly evaluate the combined effects which investigate different driving factors in NDVI. There were only two interactions, dual-factor enhancement and non-linear enhancement, for factors X1-X8 in 2000, 2005, 2010, 2015, and 2020. Most factor interactions were nonlinear enhancements, among which X1\u0026cap;X2, X1\u0026cap;X7, X2\u0026cap;X7, and X3\u0026cap;X7 represent the interaction effects of meteorological factors and topographical factors, which go beyond a simple linear superposition of their respective independent effects (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e). The most important factors in vegetation growth are appropriate precipitation and temperature, and synergistic effect can greatly promote vegetation growth and increase NDVI value; as the altitude increases, temperatures usually decrease, forming the different climatic vertical zones, which have different temperature ranges that affect the types and distribution of vegetation; topographical factors, together with rainfall, influence the distribution of water, soil formation and nutrient retention, thereby affecting the structure and function of vegetation; topographical features affect water evaporation, transpiration and recycling, for example, high-altitude mountain ranges can promote the formation of clouds and fog, increasing air humidity, and indirectly increase water supply to vegetation. With the passage of time, most driving factors have evolved into dual-factor enhanced interaction modes, such as X3\u0026cap;X6, X4\u0026cap;X8, X6\u0026cap;X8. Human activities may have different impacts on environmental factors at different time points, leading to stronger synergistic effects between two factors that originally had a non-linear relationship. The interactions between different factors become more closely intertwined and long-term environmental trends may also mask or enhance short-term fluctuation effects, resulting in different interaction patterns observed at different time points.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section3\"\u003e \u003ch2\u003e3.3.2.The impact of human factors on NDVI in Nanchang City\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003e(1)Single correlation analysis between NDVI and anthropogenic factors\u003c/p\u003e \u003cp\u003eIn base pixel scale, a single correlation analysis had been managed on the NDVI, nighttime lights, and population density from 2000 to 2022, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e. Through statistical analysis, there are 56.9% and 50.4% of the areas that vegetation cover is positively correlated with nighttime lighting and population density respectively, so with an overall positive relationship. The regions where NDVI is negatively correlated with nighttime lights are scattered in main urban area of Nanchang City, Xinjian County and Jinxian County; and the negatively correlate regions in population density are sporadically distributed in main urban area of Nanchang City and Jinxian County. Regions which NDVI is positive correlation with population density and night lights are often found near rivers and lakes.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003e(2)Partial correlation analysis of NDVI and anthropogenic factors\u003c/p\u003e \u003cp\u003eAccording to pixel scale, a partial correlation analysis had been controlled in NDVI and anthropogenic factors(Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003e). A two-tailed t-test at a significance level(α\u0026thinsp;=\u0026thinsp;0.05) indicates that the areas where NDVI and nighttime lights have the partial correlation coefficient and pass significance test(t\u0026thinsp;\u0026ge;\u0026thinsp;t\u003csub\u003e0.05\u003c/sub\u003e), mainly located at heart of Nanchang City. Regions showing a negative correlation are mainly in the main urban area and Jinxian County, accounting for 43.2% of the significant cassociation areas. The biased correlation area between NDVI and population density is dominated by negative correlation, which is scattered in main urban area and Jinxian County, with a total area of 911.097 km\u003csup\u003e2\u003c/sup\u003e, making for 51.9% of the significant correlation areas.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003e(3)Complex correlation analysis between NDVI and anthropogenic factors\u003c/p\u003e \u003cp\u003eComplex correlation in spatial pattern between NDVI and anthropogenic factors in Nanchang City is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003e. The complex correlation coefficient ranges from 0 to 0.999, and the spatial mean is 0.066. In most areas downtown of Nanchang City and Jinxian County, complex correlation between NDVI and anthropogenic factors is significant, while areas with weaker correlation are located at the boundaries of various districts and counties. The F-test at a significance level of α\u0026thinsp;=\u0026thinsp;0.05 indicates that the regions passing the significance test (F\u0026thinsp;\u0026gt;\u0026thinsp;F\u003csub\u003e0.05\u003c/sub\u003e) are mainly scattered in the downtown of Nanchang City, with a total area is 1755.48 km\u003csup\u003e2\u003c/sup\u003e where accounts 24.4% of the total area.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eDue to the rapid economic development, areas where NDVI and population density have negative correlation are characterized by high reflectance and heat absorption of buildings and roads, leading to urban heat island effect which adversely affects vegetation growth in the process of urbanization. At the same time, nighttime lights may impact the growth cycle and physiological processes of fern plants. Areas where NDVI and population density have negative correlation is owing to the high density with population in the main urban areas, resulting in air, soil and water pollution. In Jinxian County, as a high-quality agricultural and sideline product supply base in Jiangxi Province, agricultural pollution is generated, which may have a negative impact on vegetation health, reducing NDVI values. The regions where NDVI is positively correlated with nighttime lighting and population density are all located around rivers and lakes. For protecting ecological system of lakes and enhancing landscapes, ecological protection measures and urban greening projects are implemented around the lakes. These measures have helped to increase vegetation coverage, while also improved lighting facilities in these areas for nighttime protection and monitoring. Due to its beautiful natural scenery and recreational value, lakes often become hotspots of tourism and leisure activities. This has led to the construction of more tourist facilities and leisure areas near the lakes, increasing human activities and resulting in higher brightness in nighttime lights data.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section3\"\u003e \u003ch2\u003e3.3.3.The influence of land use change on spatiotemporal variation of NDVI\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThis study utilized land cover data of 2000, 2005, 2010, 2015 and 2020 to investigate the spatiotemporal changes in NDVI for different land cover types. The charges in area proportion and NDVI are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003e. The order with area size in different land cover types is: farmland\u0026thinsp;\u0026gt;\u0026thinsp;other types\u0026thinsp;\u0026gt;\u0026thinsp;forest\u0026thinsp;\u0026gt;\u0026thinsp;urban and rural land\u0026thinsp;\u0026gt;\u0026thinsp;grassland; the order of the average annual NDVI for different land cover types is: forest (0.746)\u0026thinsp;\u0026gt;\u0026thinsp;grassland (0.712)\u0026thinsp;\u0026gt;\u0026thinsp;farmland (0.708)\u0026thinsp;\u0026gt;\u0026thinsp;other types (0.593)\u0026thinsp;\u0026gt;\u0026thinsp;urban and rural land (0.552). The NDVI values of different land cover types do not show consistent correlation with their respective area proportions. Forest is almost 16.2% of the total study area, equivalent to 29.9% of the total farmland, but the average annual NDVI of forest is about 0.038 higher than that of farmland. During the research period, the area of farmland is decreasing, while the corresponding NDVI is increasing. This is closely related with the policy that return farmland to forests in Nanchang City. Although the area of farmland has decreased, the increased protection and management of existing arable land has improved the quality of the arable land, thus increasing the NDVI value. Due to the prominent performance of this land use pattern, the NDVI values of farmland are on the rise; the fluctuation range of forest and grassland areas is relatively small, and they have also shown an increasing trend over the years.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Spatial-temporal evolution characteristics of NDVI in Nanchang City\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe terrain of Nanchang City is mainly plain, where the overall topography of northwest is higher than southeast, developing into low hills, hillock, and plains in sequence, which shows a layered landform characteristic. The vegetation coverage is influenced with natural factors in meteorological and anthropogenic. From 2000 to 2022, the average NDVI of vegetation was 0.453, showing an overall increasing trend, but the growth rate gradually slowed over time. This view is consistent with Chen's study of the evolution of NDVI in Nanchang during the periods 1995 to 2005 and 2013 to 2017, which showed that although NDVI showed an increasing trend during these two periods, the value added was decreasing (Chen et al. 2022). According to the \"Overall Land Use Plan of Nanchang City (2006\u0026ndash;2020)\", the farmland in areas like Donghu District and Xihu District significantly decreased, while the construction land is significantly increased. Continuous progress of social economy and the improvement of urbanization level have increased the demand for construction land, to some extent, it affects the upward trend of NDVI. In ZHAO's study, it was also shown that the construction and expansion of cities led to the occupation of farmland, forest and grassland, as well as large areas of abandonment and deforestation, and that these land-use changes became the main cause of the degradation of vegetation growth (Zhao et al. 2016). In 2008, 2010, 2016, 2019 and 2022, the vegetation coverage significantly decreased. In 2008, the continuous economic development of Nanchang City led to more land being used for industrial construction, resulting the reduction in vegetation coverage and affecting the NDVI value. In 2010, persistent heavy rainfall in Nanchang City caused severe flooding disasters. According to the 2016 government work report, the city made significant progress in urban planning and construction, including old city renovation and new city construction, which changed land use and subsequently affected vegetation coverage and NDVI values. In 2019, the provincial government issued the \"Opinions on Comprehensive Promotion of Comprehensive Tourism Development\", making the tourism industry to a priority in deepening the construction of a strong tourism province, thereby affecting the NDVI. With the gradual easing of the COVID-19 pandemic, people resumed orderly outdoor activities and daily life in 2022, increasing human activities that impact vegetation. Spatially, the vegetation coverage in Nanchang City is mainly showing a trend of degradation, with areas of strong persistent degradation concentrated near construction land, areas of weak persistent degeneration clustered near urban land and rivers, and regions that improved in the past but are now trending towards degradation mainly scattered in southeast and west, including Jinxian County, Anyi County, and Xinjian County. Due to human activities, agricultural pollution in Jinxian County, brick kiln pollution in Anyi County, and mining pollution in Xinjian County have led to vegetation degradation. In addition, human activities may indirectly affect meteorological factors, further influencing the changes in NDVI and indicating a trend of ecological degradation.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e4.2. Analysis of NDVI Driving Factors in Nanchang City\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThis study compared and analyzed the impacts of meteorological factors and anthropogenic factors on NDVI. The anthropogenic driving effect on NDVI is more notable. ZHAO similarly identified that human-led activities, resource exploitation and land reclamation, were the main cause of vegetation cover change (Zhao et al. 2016). This discuss confirmed that among meteorological drivers soil temperature mainly affected vegetation growth, while among anthropogenic factors the most significant factors affecting vegetation were population density and night lighting. This was different from Zhang's (2022) study where air temperature and rainfall had the most significant effect on vegetation cover. We cannot consider the effect of air temperature on photosynthesis and transpiration of plants alone, the thermal properties ofsoil and the response of plant root system to temperature should also be considered. Changes in land cover types also significantly impact the evolution of NDVI, showing no difference with Wei's (2010) study elaborating that soils are critical to the ecosystem. Although we did not directly conclude that economic development has a significant effect on NDVI, it still influences vegetation growth condition by affecting drivers indirectly such as nighttime lighting, which comes to an agreement with Li's view (2024). Interannual variation trend of urban meteorological disasters is becoming increasingly significant. For example, urban heavy rainfall events exhibit periodic fluctuations. Although the annual average number of heavy rainfall and thunderstorm days has decreased in the past decade, the intensity and frequency in short-term intense precipitation events have increased (Wang et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Additionally, over the past 55 years, Nanchang City has shown an increasing trend in the annual average number of hazy days, especially in the autumn and winter seasons (Chen et al. 2016). In recent years, Jiangxi Province has also been facing increasingly severe ozone pollution issues, with factors such as prolonged sunshine duration, extreme temperature changes and abnormal decreases in precipitation collectively leading to a significant increase in near-surface ozone concentrations in autumn (Qian et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). These phenomena highlight the key role of climate change factors in monitoring the ecological development of cities. Given the global warming trend and the extreme weather events, it is expected that future climate will more and more variable. Meteorological combined effects will be significant than it is now. Therefore, further in-depth research on this issue is still needed. Human activities like population density and nighttime lights are predominantly affected vegetation cover and ecosystem health, and they need to be managed properly to ensure the ecosystem have a sustainable development. The government should formulate strict ecological protection policies to limit overgrazing, illegal logging and other activities that destroy natural ecosystems. Only rational planning of land use and optimization of land resource allocation, especially in ecologically sensitive areas, can reduce the erosion of natural ecosystems. Moreover, ecological restoration projects such as afforestation, wetland restoration and desertification control should be strengthened to ameliorate the ecological environment. The monitoring and evaluation system of vegetation cover and ecosystem health has been continuously established and improved, and the monitoring and evaluation work of vegetation cover has been carried out regularly. It is necessary to develop ecological tourism, ensure the sustainable development of the ecological environment, rationally plan the flow of tourists and avoid the destruction of vegetation and ecological pressure caused by excessive tourism.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"5. Conclusions","content":"\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn manuscript, the NDVI of Nanchang City were used to analyze the evolution and drivers of spatio-temporal separation from 2000 to 2022 through trend analysis and machine learning feature selection and so on .\u003c/p\u003e \u003cp\u003eThroughout the study period in Nanchang City,the average NDVI was 0.453, showing an overall upward trend. However, as time passed, the rate of increase gradually slowed down.Spatially, the average Hurst index of Nanchang City was 0.444, and the overall vegetation change exhibited characteristics of reverse sustained development. The degraded areas of NDVI accounted for 64.9%, showing a trend of \u0026ldquo;improvement around rivers and lakes\u0026rdquo; and \u0026ldquo;large-scale degradation of urban land\u0026rdquo;.In time-driven analyses,soil temperature had a prominent effect on vegetation cover, but overall the impact of human factors was more significant than the meteorological factors.The order of factors with significant influence on vegetation cover and evolution in Nanchang by geographical detector are as following: population density\u0026thinsp;\u0026gt;\u0026thinsp;nighttime lights\u0026thinsp;\u0026gt;\u0026thinsp;land cover types. The interactions touching the evolution of NDVI are characterized by both dual-factor enhancement and nonlinear enhancement. Spatial correlation indicated that the interannual variation levels of NDVI which negatively correlated with population density and nighttime lights, mainly scattered in heart of Nanchang City and Jinxian County; while the positively correlated areas are distributed around rivers and lakes. There is no consistency between the area proportion of land cover types and vegetation coverage.\u003c/p\u003e \u003cp\u003eEconomic level cannot directly influence NDVI,but it still influences vegetation growth condition by affecting drivers indirectly such as nighttime lighting.And among the meteorological drivers, the effect of soil temperature on NDVI is not negligible.Qualitatively and quantitatively exploration were been progressed in the evolution and driving of NDVI by spatial and temporal separation, to provide scientific basis for dynamic monitoring, management and protection of vegetation in Nanchang City.In future work, we plan to conduct multi-driver response analysis for different vegetation types and synthesize the driving factors to study the multi-factor driving mechanism of vegetation ecology, so as to provide scientific references for the dynamic changes of the ecological environment.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u0026nbsp;\u003c/strong\u003eLi Jiatong: Conceptualization, methodology, formal analysis, writing\u0026mdash;original draft preparation, writing\u0026mdash;review and editing, visualization. Zhu Jiaqi: Conceptualization, methodology, writing\u0026mdash;review and editing. Guo Qiyun: Conceptualization, formal analysis, writing\u0026mdash;review and editing. Xu Yue: formal analysis, writing\u0026mdash;original draft preparation. Li Huishan: data curation. Liu Sihang: visualization. Wu Hua: supervision.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eProject support: Construction of Talent Innovation Team and Laboratory Platform of Tibet University, Project No. 2022ZDTD10.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability:\u0026nbsp;\u003c/strong\u003eThe data presented in this study are available on request from the corresponding author.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u003c/strong\u003eI would like to declare on behalf of my co-authors thatthe work described was an original research that has not been publishedpreviously and not under consideration for publication elsewhere, inwhole or in part.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u003c/strong\u003eThe participant has consented to the submission of the article research to the joumal.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interes:\u003c/strong\u003eThe authors declare no confict of interest.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003ePublisher\u0026apos;s Note:\u003c/strong\u003eSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations.\u003c/p\u003e\n\u003cp\u003eSpringer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAhmad A.,Zhang JH,Bashir B.,et al. Exploring vegetation trends and restoration possibilities in Pakistan by using Hurst exponent. \u003cem\u003eEnviron Sci Pollut Res\u003c/em\u003e \u003cstrong\u003e2023\u003c/strong\u003e\u003cstrong\u003e,\u003c/strong\u003e30:91915\u0026ndash;91928. https://doi.org/10.1007/s11356-023-28822-0.\u003c/li\u003e\n\u003cli\u003eAdichie, J. N. Estimates of regression parameters based on rank tests. Annals of Mathematical Statistics, 1967,38: 894\u0026ndash;904.\u003c/li\u003e\n\u003cli\u003eBreiman, L. Random Forests. \u003cem\u003eMachine Learning\u003c/em\u003e \u003cstrong\u003e2001\u003c/strong\u003e,45:5\u0026ndash;32. https://doi.org/10.1023/A:1010933404324.\u003c/li\u003e\n\u003cli\u003eCao, R., Jiang, W., Yuan, L. et al. Inter-annual variations in vegetation and their response to climatic factors in the upper catchments of the Yellow River from 2000 to 2010. \u003cem\u003eJ. Geogr. Sci.\u003c/em\u003e \u003cstrong\u003e2014\u003c/strong\u003e\u003cstrong\u003e,\u003c/strong\u003e24:963\u0026ndash;979.https://doi.org/10.1007/s11442-014-1131-1.\u003c/li\u003e\n\u003cli\u003eChen,F.,Wu,F..Spatial and temporal evolution analysis of vegetation cover in Nanchang based on NDVI.\u003cem\u003eYangtze River Information and Communication\u003c/em\u003e \u003cstrong\u003e2022\u003c/strong\u003e,35(11):120-122.(in chinese)\u003c/li\u003e\n\u003cli\u003eChen,P.;Pan,H.;Xu,Y.;He,W.;Yao,H.Spatial-Temporal Evolution Characteristics and Driving Force Analysis of NDVI in Hubei Province, China, from 2000 to 2022. \u003cem\u003eForests\u003c/em\u003e \u003cstrong\u003e2024\u003c/strong\u003e, 15:719. https://doi.org/10.3390/f15040719.\u003c/li\u003e\n\u003cli\u003eChen X.g,Hu L.,Peng W.,et al.Characteristics of hazy weather meteorological elements and major air pollutants changes in Nanchang City from 1960 to 2014.\u003cem\u003eJournal of Meteorology and Environment\u003c/em\u003e \u003cstrong\u003e2016\u003c/strong\u003e,32(5):114-121.(in chinese)\u003c/li\u003e\n\u003cli\u003eDas,U.,Behera,B.. Geospatial assessment of ecological vulnerability of fragile Eastern Duars Forest integrating GIS-based AHP, CRITIC and AHP-TOPSIS models. \u003cem\u003eGeomatics, Natural Hazards and Risk\u003c/em\u003e \u003cstrong\u003e2024\u003c/strong\u003e\u003cstrong\u003e,\u003c/strong\u003e15(1). https://doi.org/10.1080/19475705.2024.2330529.\u003c/li\u003e\n\u003cli\u003eDidan, K..MODIS/Terra Vegetation Indices Monthly L3 Global 1km SIN Grid V006,\u003cem\u003eNASA EOSDIS Land Processes DAAC\u003c/em\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003e2015\u003c/strong\u003ehttps://doi.org/10.5067/MODIS/MOD13A3.006.\u003c/li\u003e\n\u003cli\u003eDror Y. Kenett,Huang XQ,Irena Vodenska,et al.Partial correlation analysis:applications for financial markets.\u003cem\u003eQuantitative Finance\u003c/em\u003e \u003cstrong\u003e2014\u003c/strong\u003e,15:569-578.https://api.semanticscholar.org/CorpusID:14156969.\u003c/li\u003e\n\u003cli\u003eGadedjisso-Tossou, A.; Adjegan, K.I.; Kablan, A.K.M. Rainfall and Temperature Trend Analysis by Mann\u0026ndash;Kendall Test and Significance for Rainfed Cereal Yields in Northern Togo.\u003cem\u003e Sci\u003c/em\u003e\u003cstrong\u003e 2021\u003c/strong\u003e, 3,17. https://doi.org/10.3390/sci3010017.\u003c/li\u003e\n\u003cli\u003eGe, R., Zhou, M., Luo, Y. et al. McTwo: a two-step feature selection algorithm based on maximal information coefficient. \u003cem\u003eBMC Bioinformatics\u003c/em\u003e \u003cstrong\u003e2016\u003c/strong\u003e\u003cstrong\u003e,\u003c/strong\u003e17:142. https://doi.org/10.1186/s12859-016-0990-0.\u003c/li\u003e\n\u003cli\u003eIratxe Ni\u0026ntilde;o-Adan,Diana Manjarres,Itziar Landa-Torres,et al.Feature weighting methods: A review.\u003cem\u003eExpert Systems with Applications\u003c/em\u003e.\u003cstrong\u003e2021\u003c/strong\u003e,184:115424.https://www.sciencedirect.com/science/article/pii/S0957417421008423.\u003c/li\u003e\n\u003cli\u003eJiangxi Bureau of Statistics,\u003cstrong\u003e2021\u003c/strong\u003e, Jiangxi Statistical Yearbook 2017[J]. \u003cem\u003eBeijing: China Statistics Press:\u003c/em\u003e 259-298.(in chinese)\u003c/li\u003e\n\u003cli\u003eJin,HY.,Chen,XH.,Wang,YM.,et al.Spatio-temporal distribution of NDVI and its influencing factors in China.\u003cem\u003eJournal of Hydrology\u003c/em\u003e \u003cstrong\u003e2021\u003c/strong\u003e,603:127129.https://doi.org/10.1016/j.jhydrol.2021.127129.\u003c/li\u003e\n\u003cli\u003eJose Alvarez-Ramirez,Juan C. Echeverria,Eduardo Rodriguez.Performance of a high-dimensional R/S method for Hurst exponent estimation.\u003cem\u003ePhysica A: Statistical Mechanics and its Applications\u003c/em\u003e \u003cstrong\u003e2008\u003c/strong\u003e,26:6452-6462.\u003c/li\u003e\n\u003cli\u003eKhaled H. Hamed.Trend detection in hydrologic data: The Mann\u0026ndash;Kendall trend test under the scaling hypothesis.\u003cem\u003eJournal of Hydrology\u003c/em\u003e \u003cstrong\u003e2008\u003c/strong\u003e,348:350-363.\u003c/li\u003e\n\u003cli\u003eKirch, W.(eds) Pearson\u0026rsquo;s Correlation Coefficient. Encyclopedia of Public Health. \u003cem\u003eSpringer, Dordrecht\u003c/em\u003e \u003cstrong\u003e2008\u003c/strong\u003e. https://doi.org/10.1007/978-1-4020-5614-7_2569.\u003c/li\u003e\n\u003cli\u003eKong,C.,Huang,J.,Du,S., Li,G.. Exploring the Driving Forces of Vegetation Greening on the Loess Plateau at the County Scale. \u003cem\u003eForests.\u003c/em\u003e \u003cstrong\u003e2024\u003c/strong\u003e,15(3):486. https://doi.org/10.3390/f15030486\u003c/li\u003e\n\u003cli\u003eKumari N, Srivastava A, Dumka UC.A Long-Term Spatiotemporal Analysis of Vegetation Greenness over the Himalayan Region Using Google Earth Engine. \u003cem\u003eClimate\u003c/em\u003e \u003cstrong\u003e2021\u003c/strong\u003e,9(7):109. https://doi.org/10.3390/cli9070109.\u003c/li\u003e\n\u003cli\u003eLakens D. Equivalence Tests: A Practical Primer for t Tests, Correlations, and Meta-Analyses. \u003cem\u003eSoc Psychol Personal Sci.\u003c/em\u003e \u003cstrong\u003e2017\u003c/strong\u003e,8(4):355-362. doi: 10.1177/1948550617697177.\u003c/li\u003e\n\u003cli\u003eLi,P.,Wang,J.,Liu,MM.,et al.Spatio-temporal variation characteristics of NDVI and its response to climate on the Loess Plateau from 1985 to 2015.\u003cem\u003eCATENA\u003c/em\u003e \u003cstrong\u003e2021\u003c/strong\u003e,203:105331.https://doi.org/10.1016/j.catena.2021.105331.\u003c/li\u003e\n\u003cli\u003eLi,Q, Gao,X, Li,J, Yan,A, Chang,S, Song,X, Lo,K. Nonlinear time effects of vegetation response to climate change: Evidence from Qilian Mountain National Park in China.\u003cem\u003e Sci Total Environ. \u003c/em\u003e\u003cstrong\u003e2024\u003c/strong\u003e,933:173149.\u003c/li\u003e\n\u003cli\u003eLi, T., Dong, Y., Wei, X. et al. The rapid prosperity of China\u0026rsquo;s Pearl River Delta from the perspective of social\u0026ndash;ecological coupling: implications for sustainable management. \u003cem\u003eSci Rep\u003c/em\u003e\u003cem\u003e \u003c/em\u003e\u003cstrong\u003e2024\u003c/strong\u003e,14:19914.\u003c/li\u003e\n\u003cli\u003eLi SX,Feng C,Yang L.Spatiotemporal patterns of evapotranspiration in the Xiangjiang River Basin from 2000 to 2020.\u003cem\u003eHydrology Research\u003c/em\u003e \u003cstrong\u003e2023\u003c/strong\u003e,54.doi = {10.2166/nh.2023.143}.\u003c/li\u003e\n\u003cli\u003eLi XQ,Jia R. MICQ-IPSO: An effective two-stage hybrid feature selection algorithm for high-dimensional data.\u003cem\u003eNeurocomputing\u003c/em\u003e \u003cstrong\u003e2022\u003c/strong\u003e,501:328-342.\u003c/li\u003e\n\u003cli\u003eLiu,Y.; Huang,C.;Zhang,L. The Spatio-Temporal Patterns and Driving Forces of Land Use in the Context of Urbanization in China: Evidence from Nanchang City. \u003cem\u003eInt. J. Environ. Res. Public Health\u003c/em\u003e \u003cstrong\u003e2023\u003c/strong\u003e, 20:2330. https://doi.org/10.3390/ijerph2003233.\u003c/li\u003e\n\u003cli\u003eLuo, H., Wang, L., Fang, J., Li, Y., Li, H., Dai, S. NDVI, Temperature and Precipitation Variables and Their Relationships in Hainan Island from 2001 to 2014 Based on MODIS NDVI. In: Bian, F., Xie, Y. (eds) Geo-Informatics in Resource Management and Sustainable Ecosystem. GRMSE 2015 2015. \u003cem\u003eCommunications in Computer and Information Science\u003c/em\u003e, \u003cstrong\u003e2016\u003c/strong\u003e,569. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-49155-3_34.\u003c/li\u003e\n\u003cli\u003eNovillo CJ, Arrogante-Funes P, Romero-Calcerrada R. Recent NDVI Trends in Mainland Spain: Land-Cover and Phytoclimatic-Type Implications. \u003cem\u003eISPRS International Journal of Geo-Information.\u003c/em\u003e \u003cstrong\u003e2019\u003c/strong\u003e,8(1):43. https://doi.org/10.3390/ijgi8010043.\u003c/li\u003e\n\u003cli\u003ePeng, S. 1-km monthly mean temperature dataset for china (1901-2022). \u003cem\u003eNational Tibetan Plateau / Third Pole Environment Data Center\u003c/em\u003e. \u003cstrong\u003e2019\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003ehttps://doi.org/10.11888/Meteoro.tpdc.270961. https://cstr.cn/18406.11.Meteoro.tpdc.270961.\u003c/li\u003e\n\u003cli\u003ePeng, S. 1-km monthly precipitation dataset for China (1901-2022). \u003cem\u003eNational Tibetan Plateau / Third Pole Environment Data Center\u003c/em\u003e.\u003cstrong\u003e2020\u003c/strong\u003e. https://doi.org/10.5281/zenodo.3185722.\u003c/li\u003e\n\u003cli\u003eQian,Y.,Xu,B.,Xia,LJ.,et al.Characterization of ozone pollution in Jiangxi Province from 2016 to 2019 and analysis of the influence of meteorological factors[J]. \u003cem\u003eEnvironmental Science\u003c/em\u003e \u003cstrong\u003e2021\u003c/strong\u003e,42(5):2190-2201.(in chinese)\u003c/li\u003e\n\u003cli\u003eQiu HB,Luo J,Zhang JJ.Robustness of F-tests in singular linear models.Acta Mathematica Sinica, \u003cem\u003eEnglish Series\u003c/em\u003e \u003cstrong\u003e2014\u003c/strong\u003e,30:872-880.https://api.semanticscholar.org/CorpusID:120336999.\u003c/li\u003e\n\u003cli\u003eQuirk, T.J., Palmer-Schuyler, J. Multiple Correlation and Multiple Regression. In: Excel 2019 for Human Resource Management Statistics. Excel for Statistics. \u003cem\u003eSpringer, Cham.\u003c/em\u003e \u003cstrong\u003e2020. \u003c/strong\u003ehttps://doi.org/10.1007/978-3-030-58001-8_7\u003c/li\u003e\n\u003cli\u003eQu,S.,Wang,LC.,Lin,AW.et al.Distinguishing the impacts of climate change and anthropogenic factors on vegetation dynamics in the Yangtze River Basin,China.\u003cem\u003eEcological Indicators\u003c/em\u003e. \u003cstrong\u003e2020\u003c/strong\u003e,108:105724.https://doi.org/10.1016/j.ecolind.2019.105724.\u003c/li\u003e\n\u003cli\u003eRen, Y., Deng, L., Zuo, S. et al. Geographical modeling of spatial interaction between human activity and forest connectivity in an urban landscape of southeast China. \u003cem\u003eLandscape Ecol\u003c/em\u003e\u003cem\u003e \u003c/em\u003e\u003cstrong\u003e2014\u003c/strong\u003e,29:1741\u0026ndash;1758. https://doi.org/10.1007/s10980-014-0094-z.\u003c/li\u003e\n\u003cli\u003eRivas-Tabares, D.A.; Saa-Requejo, A.; Mart\u0026iacute;n-Sotoca, J.J.; Tarquis, A.M. Multiscaling NDVI Series Analysis of Rainfed Cereal in Central Spain. \u003cem\u003eRemote Sens.\u003c/em\u003e \u003cstrong\u003e2021\u003c/strong\u003e, 13:568. https://doi.org/10.3390/rs13040568.\u003c/li\u003e\n\u003cli\u003eRyan J. Urbanowicz,Melissa Meeker,William La Cava,et al.Relief-based feature selection: Introduction and review.\u003cem\u003eJournal of Biomedical Informatics\u003c/em\u003e \u003cstrong\u003e2018\u003c/strong\u003e,85:189-203.\u003c/li\u003e\n\u003cli\u003eSong YZ,Wang JF,Ge Yong,et al.An optimal parameters-based geographical detector model enhances geographic characteristics of explanatory variables for spatial heterogeneity analysis: cases with different types of spatial data.\u003cem\u003eGIScience \u0026amp; Remote Sensing\u003c/em\u003e \u003cstrong\u003e2020\u003c/strong\u003e,57(5):593-610.https://api.semanticscholar.org/CorpusID:219418482.\u003c/li\u003e\n\u003cli\u003eSpola\u0026ocirc;r N,Monard MC.Evaluating ReliefF-Based Multi-Label Feature Selection Algorithm.\u003cem\u003eIbero-American Conference on AI\u003c/em\u003e \u003cstrong\u003e2014\u003c/strong\u003e.https://api.semanticscholar.org/CorpusID:23555972.\u003c/li\u003e\n\u003cli\u003eSun,R., Chen,S., Su,H.,Hao,G.. \u0026quot;Spatiotemporal Variation of Vegetation Coverage and its Response to Climate Change Before and After Implementation of Grain for Green Project in the Loess Plateau,\u0026quot; \u003cem\u003eIGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium,\u003c/em\u003e Yokohama, Japan, \u003cstrong\u003e2019\u003c/strong\u003e:9546-9549, doi: 10.1109/IGARSS.2019.8898716.\u003c/li\u003e\n\u003cli\u003eTang,X.; Wang,J.; Lu, J.; Liu, G.; Chen, J. Improving Bearing Fault Diagnosis Using Maximum Information Coefficient Based Feature Selection. \u003cem\u003eAppl. Sci.\u003c/em\u003e \u003cstrong\u003e2018\u003c/strong\u003e, 8:2143. https://doi.org/10.3390/app8112143\u003c/li\u003e\n\u003cli\u003eTeixeira, A.L., Leal, J.P. , Falcao, A.O. Random forests for feature selection in QSPR Models - an application for predicting standard enthalpy of formation of hydrocarbons.\u003cem\u003e J Cheminform 5\u003c/em\u003e, \u003cstrong\u003e2013\u003c/strong\u003e. https://doi.org/10.1186/1758-2946-5-9.\u003c/li\u003e\n\u003cli\u003eTerefe T.,Feyera S.,Moges K..The impact of land use/land cover change on ecosystem services in the central highlands of Ethiopia. \u003cem\u003eEcosystem Services\u003c/em\u003e \u003cstrong\u003e2017\u003c/strong\u003e,(23):47-54.https://doi.org/10.1016/j.ecoser.2016.11.010\u003c/li\u003e\n\u003cli\u003eTian,XR.,Tao,Z.,Xie,Y.,et al.Spatiotemporal Evolution and Driving Mechanism of Fractional Vegetation Coverage in the Yangtze River Delta.\u003cem\u003eIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.\u003c/em\u003e\u003cstrong\u003e2024\u003c/strong\u003e,17,10979-10997.https://api.semanticscholar.org/CorpusID:270177260.\u003c/li\u003e\n\u003cli\u003eTong,S.,Zhang,J.,Bao,Y. et al. Analyzing vegetation dynamic trend on the Mongolian Plateau based on the Hurst exponent and influencing factors from 1982\u0026ndash;2013.\u003cem\u003e J. Geogr. Sci.\u003c/em\u003e \u003cstrong\u003e2018\u003c/strong\u003e\u003cstrong\u003e,\u003c/strong\u003e28:595\u0026ndash;610. https://doi.org/10.1007/s11442-018-1493-x.\u003c/li\u003e\n\u003cli\u003eWang,J.,Zhao,J.,Zhou,P.,Li K,Cao Z,Zhang H,Han Y,Luo Y, Yuan X.Study on the Spatial and Temporal Evolution of NDVI and Its Driving Mechanism Based on Geodetector and Hurst Indexes: A Case Study of the Tibet Autonomous Region. \u003cem\u003eSustainability.\u003c/em\u003e \u003cstrong\u003e2023\u003c/strong\u003e,15(7):5981. https://doi.org/10.3390/su15075981.\u003c/li\u003e\n\u003cli\u003eWang JX,Hu MG,Zhang FS,et al.Influential factors detection for surface water quality with geographical detectors in China.\u003cem\u003eStochastic Environmental Research and Risk Assessment\u003c/em\u003e \u003cstrong\u003e2018\u003c/strong\u003e,32:2633-2645.\u003c/li\u003e\n\u003cli\u003eWang,RL.,Wu,F.,Liu,ZP.,et al. Characterization of urban meteorological disaster changes in Nanchang over the past 60 years. \u003cem\u003eResearch on Meteorology and Disaster Reduction\u003c/em\u003e \u003cstrong\u003e2014\u003c/strong\u003e,37(2):44-49.(in chinese)\u003c/li\u003e\n\u003cli\u003eWang XQ,Dang X,Peng HX,et al.The Theil-Sen Estimators in a Multiple Linear Regression Model.\u003cem\u003eManuscript\u003c/em\u003e, \u003cstrong\u003e2009\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eWang,Z.,Liu,Z.,Huang,M.NDVI joint process-based models drive a learning ensemble model for accurately estimating cropland net primary productivity (NPP). \u003cem\u003eFront. Environ. Sci\u003c/em\u003e.\u003cstrong\u003e2024\u003c/strong\u003e,11.https://doi.org/10.3389/fenvs.2023.1304400.\u003c/li\u003e\n\u003cli\u003eWei,LN.,Zheng,L.,Zhang,JW.,et al. Analysis of land use/cover change and its human factor drivers in Nanchang. \u003cem\u003eAnhui Agricultural Science\u003c/em\u003e \u003cstrong\u003e2010\u003c/strong\u003e,38(21):11263-11268.DOI:10.13989/j.cnki.0517-6611.2010.21.193.(in chinese)\u003c/li\u003e\n\u003cli\u003eWu RN,Zhang JQ,Bao YH,et al.Geographical Detector Model for Influencing Factors of Industrial Sector Carbon Dioxide Emissions in Inner Mongolia, China.\u003cem\u003eSustainability\u003c/em\u003e \u003cstrong\u003e2016\u003c/strong\u003e,8:149.https://api.semanticscholar.org/CorpusID:18103433.\u003c/li\u003e\n\u003cli\u003eWu,YZ.; Shi,KF.; Chen,ZQ.; Liu,SR.; Chang,ZJ. An improved time-series DMSP-OLS-like data (1992-2022) in China by integrating DMSP-OLS and SNPP-VIIRS.\u003cem\u003eIEEE Transactions on Geoscience and Remote Sensing\u003c/em\u003e, \u003cstrong\u003e202\u003c/strong\u003e\u003cstrong\u003e2\u003c/strong\u003e,60. DOI:10.1109/TGRS.2021.3135333.\u003c/li\u003e\n\u003cli\u003eXu,XL.,Liu,JY.,Zhang,SW.,Li,RD.,Yan,CZ.,Wu,SX.. China multi-period land use remote sensing monitoring dataset (CNLUCC). \u003cem\u003eResource and Environmental Science Data Registration and Publication System\u003c/em\u003e \u003cstrong\u003e2018\u003c/strong\u003e.DOI:10.12078/2018070201.(in chinese)\u003c/li\u003e\n\u003cli\u003eXu,Y.,Dai,QY.,Zou,B.,et al.Tracing climatic and human disturbance in diverse vegetation zones in China: Over 20 years of NDVI observations. \u003cem\u003eEcological Indicators\u003c/em\u003e \u003cstrong\u003e2023\u003c/strong\u003e,156:105331.https://doi.org/10.1016/j.ecolind.2023.111170.\u003c/li\u003e\n\u003cli\u003eYang,Y.,Wang,S.,Bai,X.,Tan,Q.,Li,Q.,Wu,L.,Tian,S.,Hu,Z.,Li,C.,Deng,Y.. Factors Affecting Long-Term Trends in Global NDVI.\u003cem\u003e Forests.\u003c/em\u003e \u003cstrong\u003e2019\u003c/strong\u003e,10(5):372. https://doi.org/10.3390/f10050372.\u003c/li\u003e\n\u003cli\u003eYang,ZW.,Gao,X.,Lei,JQ.,et al.Analysis of spatiotemporal changes and driving factors of desertification in the Africa Sahel.\u003cem\u003eCATENA\u003c/em\u003e \u003cstrong\u003e2022\u003c/strong\u003e,213:106213.https://doi.org/10.1016/j.catena.2022.106213.\u003c/li\u003e\n\u003cli\u003eZhang HY,Feng ZQ,Feng SY,et al.A Survey of Methods for Estimating Hurst Exponent of Time Sequence.\u003cem\u003eArXiv\u003c/em\u003e\u003cem\u003e \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e.https://api.semanticscholar.org/CorpusID:264828969.\u003c/li\u003e\n\u003cli\u003eZhang, J., Chen, M., Zhao, S., Hu, S., Shi, Z., Cao, Y. ReliefF-based EEG sensor selection methods for emotion recognition. \u003cem\u003eSensors (Switzerland) \u003c/em\u003e \u003cstrong\u003e2016\u003c/strong\u003e,16(10):1558. https://doi.org/10.3390/s16101558.\u003c/li\u003e\n\u003cli\u003eZhang,J.; Jiao,G.; Ye,Q.; Gu,X. The Impact of Urban Expansion on the Urban Thermal Environment: A Case Study in Nanchang, Jiangxi, China. \u003cem\u003eSustainability\u003c/em\u003e \u003cstrong\u003e2022\u003c/strong\u003e, 14:16531. https://doi.org/10.3390/su142416531.\u003c/li\u003e\n\u003cli\u003eZhang, L., Cong, Z., Zhang, D., \u0026amp; Li, Q. Response of vegetation dynamics to climatic variables across a precipitation gradient in the Northeast China Transect. \u003cem\u003eHydrological Sciences Journal\u003c/em\u003e \u003cstrong\u003e2017\u003c/strong\u003e, 62(10):1517\u0026ndash;1531. https://doi.org/10.1080/02626667.2017.1337274\u003c/li\u003e\n\u003cli\u003eZhang,L.,Zhang,Y.,Wang,J. et al. Spatiotemporal evolution characteristics and driving forces of vegetation cover variations in the Chengdu-Chongqing region of China under the background of rapid urbanization. \u003cem\u003eEnviron Sci Pollut Res\u003c/em\u003e \u003cstrong\u003e2024\u003c/strong\u003e,31:22976\u0026ndash;22993. https://doi.org/10.1007/s11356-024-32645-y.\u003c/li\u003e\n\u003cli\u003eZhang,X.,Han,L.,Li,L.,and Bai, Z..Analysis of desertification and the driving factors over the Loess Plateau. \u003cem\u003eGeocarto International\u003c/em\u003e, \u003cstrong\u003e2023\u003c/strong\u003e,38(1). https://doi.org/10.1080/10106049.2023.2290175.\u003c/li\u003e\n\u003cli\u003eZhang XM. Characterization of spatial and temporal variations of NDVI in Jiangxi Province and its correlation analysis with climate factors[J]. \u003cem\u003eYangtze River Information and Communication\u003c/em\u003e, \u003cstrong\u003e2022\u003c/strong\u003e,35(06):7-12.(in chinese)\u003c/li\u003e\n\u003cli\u003eZhang, Y., Zhang, L., Wang, J.,et al. Quantitative analysis of NDVI driving factors based on the geographical detector model in the Chengdu-Chongqing region, China. \u003cem\u003eEcological Indicators\u003c/em\u003e. \u003cstrong\u003e2023\u003c/strong\u003e,155:110978.\u003c/li\u003e\n\u003cli\u003eZhao,LH.,Wang,P.,Ouyang,XZ.,Wu,ZW.. Spatio-temporal evolution of vegetation cover and its response to non-climatic factors in Nanchang. \u003cem\u003eJournal of Ecology\u003c/em\u003e \u003cstrong\u003e2016\u003c/strong\u003e,36(12):3723-3733.(in chinese)\u003c/li\u003e\n\u003cli\u003eZhao NZ,Liu Y,Cao GF,et al.Forecasting China\u0026rsquo;s GDP at the pixel level using nighttime lights time series and population images.\u003cem\u003eGIScience \u0026amp; Remote Sensing\u003c/em\u003e \u003cstrong\u003e2017\u003c/strong\u003e,(1):1-19.\u003c/li\u003e\n\u003cli\u003eZhao,W.,Wang,H.,Zhang,H. et al. Precipitation and anthropogenic activities regulate the changes of NDVI in Zhegucuo Valley on the southern Tibetan Plateau.\u003cem\u003e J. Mt. Sci.\u003c/em\u003e \u003cstrong\u003e2024\u003c/strong\u003e, 21:607\u0026ndash;618. https://doi.org/10.1007/s11629-023-8299-8.\u003c/li\u003e\n\u003cli\u003eZhong, H.; Wang, H. Temporal and spatial variation of normalized vegetation index in Hubei Province from 2007 to 2016. \u003cem\u003eJ. Cent.China Norm. Univ. (Nat. Sci.)\u003c/em\u003e \u003cstrong\u003e2018\u003c/strong\u003e, 52(04), 582\u0026ndash;588.\u003c/li\u003e\n\u003cli\u003eZhu,LM.,Zhu,KX.,Zeng,XJ..Evolution of landscape pattern and response of ecosystem service value in international wetland cities: A case study of Nanchang City.\u003cem\u003eEcological Indicators\u003c/em\u003e \u003cstrong\u003e2023\u003c/strong\u003e,155:110987.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"climate change, driving analysis, NDVI, temporal dimension, spatial dimension","lastPublishedDoi":"10.21203/rs.3.rs-5366943/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5366943/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eInvestigating vegetation coverage and quantitatively evaluating environmental changes can serve as the science knowledge in ecological protection, resource management, and policy-making, promoting harmonious coexistence between human and nature. In this study, we had explored the separation in space and time of evolutionary characteristics and driving factors of NDVI in Nanchang City from 2000 to 2022 based on Hurst Exponent, ReliefF feature selection algorithm, Geographical detector and so on. The results are: (1) From temporal dimension, the average NDVI in Nanchang City was 0.453, showing an overall upward trend. Although the growth rate gradually slowed over time. (2) In terms of spatial changes, vegetation in Nanchang City overall exhibited a characteristic of reverse sustained development, showing trends of \"improvement around rivers and lakes\" and \"large-scale degradation of urban land.\" (3) The ReliefF proved to be more suitable among the three algorithms in the temporal dimension-driven analysis. Human factors are the dominant factors significantly influencing the changes in NDVI, while meteorological factors are not as significant. (4) The driver-analysis of geographical dector shows that population density, nighttime lights, and land cover types emerged as significant driving factors. Regions where NDVI and human factors have negative correlation are primarily centred in the heart of Nanchang City and Jinxian County; while the positive correlations are found around rivers and lakes. This study delves into the changing patterns of vegetation cover in Nanchang City, providing scientific guidance for the protection and regulation the regional ecological environment to bring about a sustainable development.\u003c/p\u003e","manuscriptTitle":"Spatio-temporal separating analysis of NDVI evolution and driving factors: a case study in Nanchang, China","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-12-17 17:21:29","doi":"10.21203/rs.3.rs-5366943/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"30d853b6-d226-4dff-a890-c3a688e60ff3","owner":[],"postedDate":"December 17th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-05-23T10:53:15+00:00","versionOfRecord":[],"versionCreatedAt":"2024-12-17 17:21:29","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5366943","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5366943","identity":"rs-5366943","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}