Spatiotemporal variation in carbon sequestration in the forest ecosystem of Hainan Island over a 30-year period and its driving factors

Spatiotemporal variation in carbon sequestration in the forest ecosystem of Hainan Island over a 30-year period and its driving factors | 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 Spatiotemporal variation in carbon sequestration in the forest ecosystem of Hainan Island over a 30-year period and its driving factors ZhiHao Pi, Xu Wang, Zhuo Zang, XiQang Song, GuangYi Zhou, Hao Guo, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4105908/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 This study examines long-term carbon sequestration in the forest ecosystems of Hainan Island from 1990 to 2020 using the InVEST model and a geographic detector technique. We analysed changes in land use and forest cover, observing an 85.78%, 87.55%, and 256.96% decrease in undeveloped, shrub-covered, and burned urbanised land, respectively. Urbanised land increased by 4.01% annually. Forested land decreased by 3.62%, agricultural land expanded by 5.27%, and aquatic bodies decreased by 2.05%. The forest ecosystems sequestered 335.09–372.80 TgC of carbon, showing an upward trend from 1991 to 1997, a decline from 1997 to 2004, an increase from 2004 to 2010, a decrease from 2010 to 2015, and overall stability from 2015 to 2020. Spatial clustering analysis revealed substantial clustering of carbon sequestration, with central mountainous regions exhibiting elevated levels, coastal areas having diminished levels, the east experiencing higher levels than the west, and the south showing escalated levels compared to the north. Geographical detector analysis identified NDVI, elevation, and slope as primary drivers of spatial variance in carbon sequestration. Forested area changes and government forestry policies played a pivotal role in enhancing carbon sequestration. The combined effect of NDVI and elevation normalisation on vegetation coverage had the most potent synergistic impact. Carbon storage forest ecosystem InVEST model geographic detector Hainan Island Influencing factors Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Tropical rainforests are crucial forest ecosystems worldwide, covering just 12% of Earth’s land area and hosting over 50% of the world’s species, accounting for 40% of the carbon reservoirs in terrestrial ecosystems (Phillips et al., 1998 ). They also represent 55% of global forest carbon stocks (Pan et al., 2011 ). Minor alterations in carbon sequestration in tropical rainforest ecosystems can substantially affect ecosystems and global climate change. Estimating forest ecosystem carbon storage, analysing spatiotemporal variation patterns, and exploring the mechanisms influencing carbon sequestration capacity in tropical forest ecosystems are important for carbon management in tropical forest ecosystems. Hainan Island hosts China’s largest expanse of tropical forests, forming a concentrated, well-preserved, and contiguous collection of ‘island-like’ tropical forests within the nation (Chai et al., 2022). It is an essential component of global tropical forest ecosystems (Kothandaraman et al., 2020 ; Hai et al., 2014). Since the establishment of Hainan Province in 1988, rapid socioeconomic development and urban expansion have occurred, resulting in notable changes in land use that have profoundly affected the structure and function of the island’s ecosystem. Several domestic and international scholars have conducted extensive research on forest carbon storage in Hainan. Cao et al. ( 2002 ) examined changes in carbon storage within Hainan Island forests from 1979 to 1999. Zhang et al. ( 2004 ) analysed the variations in carbon storage attributed to different forest vegetation classification systems using forest inventory data and diverse land use/cover classification systems. Ren et al. ( 2014 ) estimated the distribution pattern of carbon storage across forest ecosystems on Hainan Island over four survey periods from 1993 to 2008. The study incorporated additional estimates from the understory vegetation, litter layer, and soil layer. Liu et al. ( 2022 ) investigated the relationship between land use/cover and carbon storage on Hainan Island from 1992 to 2019. Gao et al. ( 2023 ) calculated the carbon storage and spatial distribution characteristics of forest ecosystems on Hainan Island in 2023, using inventory data and remote sensing images. They enhanced the carbon density data for various parts of the forest ecosystem in the Hainan Province through extensive field measurements, thus improving the accuracy of forest carbon storage estimates. Although resource inventory data, remote sensing images, or a combination of both have been used in previous studies to evaluate forest carbon storage on Hainan Island, there is still a gap in understanding the primary factors that influence the spatial differentiation of carbon storage. Geographical detectors offer distinct advantages for the study of geographical phenomena. They can consider the complex relationships among multiple factors and reveal their mechanisms of impact on spatial distribution patterns. However, the use of geographical detectors requires the discretisation of continuous variables, as the driving factors must be continuous (Li, 2020). Continuous factor data are spatial data that describe attributes and spatial features. Attribute features contain either interval or ratio data (Peng, 2001). Discretisation is the process of converting interval or ratio data into nominal or ordinal discrete factor data (Hanning et al., 2003). Discretising continuous variables can be challenging owing to the subjective and random nature involved. Geographical detectors commonly use methods such as equal intervals, equal frequencies, geometric breakpoints, natural breakpoints, K-means, quantiles, and standard deviation distance for discretisation (Li, 2020). In the present study, we used the one-standard deviation method to classify and calculate the decision value q for each driving factor. This method is based on the numerical characteristics and distribution patterns of the original data, considering the mean value of the data as the classification centre, thereby yielding in more objective results (Meng et al., 2021 ). In forest carbon storage research, ground survey data are preferred due to their high accuracy. However, adequately capturing the spatial heterogeneity of ground vegetation remains challenging because of the limited number of available survey points. Conversely, simulation methods that require less data and yield precise evaluation outcomes are adept at exploring alterations in regional carbon storage across diverse spatial and temporal scales (Wang et al., 2011 ; Fu et al., 2023 ). These approaches have garnered considerable attention (Zhang et al., 2016 ; Yang et al., 2020; Fan et al., 2023 ). The Integrated Valuation of Ecosystem Services and Trade-offs (InVEST) model is widely used in carbon storage assessments, primarily using land-use data (He et al., 2016 ; Piyathilake et al., 2022 ; Li et al., 2023 ). Previous studies have mainly focused on examining the impact of land use changes on carbon storage dynamics. For instance, Zhu et al. ( 2019 ) used this model to examine the effects of land-use alterations on ecosystem carbon storage in the Qihai River Basin of the Taihang Mountains. Zou et al. ( 2021 ) illustrated the capability of the InVEST model to simplify the modelling of yearly carbon fixation by forest trees and enhance carbon density approximations for forests. Zhu et al. ( 2021 ) used this model to investigate the effects of land use changes on carbon storage during land expansion in Guangzhou. In this study, we used the InVEST model to examine the total volume and spatial dynamics of forest carbon storage on Hainan Island from 1990 to 2020. We conducted a geographical detector analysis to examine the driving factors and interaction patterns that contribute to the spatial differentiation of carbon storage in the forest ecosystems on Hainan Island. This facilitated identification of the primary factors influencing changes in forest carbon storage within the region. This study provides a theoretical basis for carbon management strategies in Hainan Forest ecosystems and aids in evaluating the role of forests in achieving carbon neutrality. Study Area Overview Hainan Island (18.80°–20.10° N, 108.37°–111.03° E) is located at the northern extremity of the South China Sea continental shelf, facing the Leizhou Peninsula across the water. It spans approximately 34,000 km 2 , encompasses 18 cities and counties, and has a coastline extending 1,944.4 km. The region exhibits diverse landforms, with towering mountains at its core and gradually declining elevations towards the periphery. The prominent mountainous features in the area are Wuzhi Mountain, with an elevation of 1867 m, and Yingge Ridge, which stands at 1811 m. The terrain gradually transitions from mountainous to hills, plateaus, and plains. The climate is characterised by high temperatures, abundant rainfall, and lengthy summers without winters. Annual precipitation ranges from 1000 to 2500 mm, whereas the average annual temperature fluctuates between 23.8 and 26.2°C, indicating a tropical monsoon maritime climate. The land use on Hainan Island is dominated by forests and farmland, which together account for approximately 90% of the total land area. According to Lei et al. ( 2020 ), forest coverage has increased substantially from 38.2% in the early stages of provincial establishment to 62.1%. Research Methodology Data collection The land-use data for Hainan Island from 1990 to 2020 were obtained from the “30 m Annual Land Cover Data for China from 1990 to 2022” dataset by Yang et al. (2021). The dataset has a spatial resolution of 30 m × 30 m. Normalised difference vegetation index (NDVI) data were obtained from the spatial distribution dataset of the annual vegetation index in China, curated by Xu Xinliang. The dataset covers the period 1998–2020 and has a spatial resolution of 1 km. Annual average temperature and precipitation data for Hainan Island were obtained from the National Earth System Science Data Center. The data have a spatial resolution of 1 km and can be accessed through http://www.geodata.cn/ . The elevation data used in this study were obtained from Professor Tang Guo’an’s China Digital Elevation Model (DEM) on a space-time three-pole environmental big data platform. DEM data can capture local terrain features at a certain resolution, facilitating the extraction of abundant surface morphology information, including the hillshade, slope, and aspect, at a spatial resolution of 1 km. The population density grids were obtained from WorldPop ( https://www.worldpop.org/ ) at a spatial resolution of 1 km. Nighttime light data were sourced from the “DMSP-OLS-like” nighttime light remote sensing dataset for China from 1992 to 2019, published online by Wu Yizhen et al. (2021) in the IEEE Transactions on Geoscience and Remote Sensing journal, with a spatial resolution of 1 km. Per capita GDP, forestry policy information, and other relevant data were extracted from various editions of the “Hainan Statistical Yearbook”, “China Forestry Statistical Yearbook”, and “China Forestry and Grassland Statistical Yearbook”. These publications were compiled based on annual forestry statistical reports and other relevant materials submitted by the forestry management departments of provinces, autonomous regions, municipalities directly under the central government, and units directly under the National Forestry and Grassland Administration. The carbon density data used in the InVEST model for the carbon storage function module, with land-use types as evaluation units, were based on the latest research results on carbon density on Hainan Island by Gao et al. ( 2023 ). Carbon density values were assigned to raster objects based on land-use type, with forestland as the primary carbon density table, resulting in the calculation of carbon storage in the forest ecosystem of Hainan Island. The carbon densities are listed in Table 1 . Table 1 Carbon storage and density of different carbon pools in the forest ecosystems of Hainan Island. Index Aboveground biomass carbon pool Belowground biomass carbon pool Litterfall carbon pool Soil carbon pool Forest ecosystem Carbon storage/(TgC) 85.12 18.73 2.90 231.40 338.15 Carbon density (MgC/hm 2 ) 37.17 8.18 1.27 101.04 147.66 Percentage/% 25.17 5.54 0.86 68.43 100.00 Carbon storage calculation The study used InVEST 3.12.1 (Sharp et al., 2014 ) and its carbon-storage module to calculate the carbon storage for each year over a 20-year period within the study area. Land use, aboveground biomass, belowground biomass, soil, and dead organic matter data were used for this purpose. The carbon storage pattern was mapped every five years, encompassing four main components of terrestrial ecosystem carbon storage: aboveground biomass carbon ( \({C}_{above}\) ), belowground biomass carbon ( \({C}_{below}\) ), soil carbon ( \({C}_{soil}\) ), and dead organic carbon ( \({C}_{dead}\) ). The formula used was as follows: $${C}_{i}={C}_{i,above}+{C}_{i,below}+{C}_{i,soil}+{C}_{i,dead}$$ 1 $${C}_{total}=\sum _{i=1}^{n}{C}_{i}\times {S}_{i}$$ 2 where, \(i\) represents the \(i\) -th land-use type, \({C}_{i}\) denotes the total carbon density for the \(i\) -th land-use type (t \(\cdot\) hm −2 ), \({C}_{i,above}\) represents the aboveground carbon density for the \(i\) -th land-use type (t \(\cdot\) hm −2 ), \({C}_{i,below}\) represents the belowground carbon density for the \(i\) -th land-use type (t \(\cdot\) hm −2 ), \({C}_{i,dead}\) represents the dead organic matter carbon density for the \(i\) -th land-use type (t \(\cdot\) hm −2 ), \({C}_{i,soil}\) represents the soil carbon density for the \(i\) -th land-use type (t \(\cdot\) hm −2 ), \({S}_{i}\) denotes the area for the \(i\) -th land-use type (m 2 ), \({C}_{i,total}\) represents the total carbon storage for all land areas (t). Identification of carbon sequestration driving factors In this study, we used the factor and interaction detection modules of geographic detectors to investigate the driving factors and interactions that shape the spatial distribution of carbon storage in the forest ecosystems of Hainan Island. Based on relevant literature, we identified nine key factors influencing spatiotemporal differentiation in carbon storage (Wang C et al., 2023; Wang Z, 2021 ). NDVI, elevation, slope, nighttime light index, GDP per capita, population average density, annual cumulative precipitation, annual average temperature, and cumulative sunshine hours were selected as indicators because they partially reflect the environmental conditions for vegetation growth and the impact of human activities on ecosystems. Factor Detection: The main objective of this method is to identify spatial variations in Y and evaluate the extent to which factor x accounts for spatial variations in attribute Y. The factor detection results were measured using the q-value, which was calculated using the following formula: $$q=1-\frac{\sum _{i=1}^{1}{n}_{i}{\delta }_{i}^{2}}{n{\delta }^{2}}$$ 3 where, \(q\) is the strength of the explanation of factor \(x\) for \(Y\) , and ranges from 0 to 1, \(i\) Number of different carbon storage classes in the study area, \({n}_{i}\) and \(n\) are the sample sizes of the \(i\) -th class in the study area and the total sample size of the study area, respectively. \(\delta {\text{i}}^{2}{\text{i}}^{2}\) and \({\delta }^{2}\) . The variance of carbon storage in the \(i\) -th class of the study area and the total variance of carbon storage in the study area. The interaction detection module examines whether interactions exist among different driving factors and evaluates the strength of these interactions. It assesses whether the combined influence of factors X1 and X2 enhances or diminishes the explanatory power of the dependent variable Y, or whether the effects of these factors on Y are independent. It determines whether the combined correlation of different driving factors amplifies or diminishes their impact on the dependent variable Y. The interactions between the driving factors are listed in Table 2 . Table 2 Types of interactions The criteria for judging Type q(x1∩x2) < Min[q(x1), q(x2)] Nonlinear attenuation Min[q(x1),q(x2)] < q(x1∩x2) Max[q(x1), q(x2)] Enhancement by two factors q(x1∩x2) = q(x1) + q(x2) Independent q(x1∩x2) > q(x1) + q(x2) Nonlinear enhancement Spatial pattern analysis of carbon storage This study used the Anselin Local Moran’s I index to describe the global spatial autocorrelation characteristics of carbon storage, which can be calculated based on the neighbouring areas of each region to obtain the local clustering situation of each region. Using this index, we further explored the spatial distribution patterns of carbon storage, identified local hotspots and cold spots, and identified clustering patterns and possible outliers in each region, gaining a deeper understanding of the spatial characteristics of carbon storage in the Hainan Island forest ecosystem. The formula used to calculate this index is as follows: $$\text{Moran's I }=\frac{n\sum _{i=1}^{n} \sum _{j=1}^{n} {w}_{ij}\left({x}_{i}-\stackrel{-}{x}\right)\left({x}_{j}-\stackrel{-}{x}\right)}{\sum _{i=1}^{n} {\left({x}_{i}-\stackrel{-}{x}\right)}^{2}\left(\sum _{i} \sum _{j} {w}_{ij}\right)}$$ 4 where, \({x}_{i}\) represents the carbon storage value of the \(i\) -th region, \(\stackrel{-}{x}\) the global mean value, \({w}_{ij}\) denotes the spatial weight between the regions \(i\) and \(j\) , \(\delta {\text{i}}^{2}, {\text{i}}^{2}\) , and \({\delta }^{2}\) represent the variance of carbon storage in the \(i\) -th region and the overall variance of carbon storage in the study area, respectively. Results and Analysis Land use change in Hainan Island Using land cover data of 1990–2020, we created spatial distribution maps and land-use transition matrices at 5-year intervals for Hainan Island (Fig. 1– 2 ). The island displays substantial spatial heterogeneity in land-use types, with forestland being the most dominant, ranging from 63.9–70.9% of all land-use types, followed by arable land, which accounts for 24.7–31.9%. This study found that the built-up land area ranged from 0.6–1.8%, whereas unused land accounted for only 0.002–0.14%. The other land-use types remained relatively stable. Notable transitional changes were observed in each land type during the study period. Substantial transitions occurred in all land-use types, except for minor changes in water body areas, from 1990 to 2020. Between 1990 and 1995, the most remarkable change was the conversion of 1819.2 km 2 of arable land to forest land, which accounted for 62.15% of the total conversion area. Conversely, 831.1 km 2 of forestland was converted into arable land, accounting for 29.39% of the total conversion area. The built-up and shrub areas experienced relatively minimal changes during this period. Between 1995 and 2000, the largest area of forestland was converted to arable land, reaching 1634.0 km 2 , accounting for 62.17% of the total conversion area. Conversely, 753.2 km 2 of arable land was converted into forestland, accounting for 28.66% of the total conversion area. During this period, shrubs, grasslands, and unused land experienced substantial reductions, whereas conversion between forestland and arable land remained relatively high. Between 2000 and 2005, arable land expanded, and land types other than built-up land were converted into arable land. The largest conversion during this period was the transformation of forest land into arable land, which totalled 1845.1 km 2 , accounting for 66.05% of the total conversion area. In contrast, 692.9 km 2 of arable land was converted to forestland, comprising 24.80% of the total conversion area. Between 2005 and 2010, the forestland area increased substantially, with a growth of 2408.9 km 2 in 2010 compared to 2005, primarily attributed to conversion from arable land. Built-up land received inputs from all land types except shrublands. From 2010 to 2015, arable land received inputs from forest land, with 2242.2 km 2 of forest land being converted to arable land, whereas the area of built-up land continued to grow. From 2015 to 2020, the forestland and arable land areas remained almost stable, indicating a relatively stable trend. The area of built-up land continued to increase. Figure 1 Spatial Distribution of Land Use on Hainan Island from 1990 to 2020 Spatiotemporal characteristics of carbon storage in the forest ecosystem of Hainan Island Using land use classification remote sensing images of Hainan Island from 1990 to 2020 and the InVEST model, we computed the total carbon storage of the forest ecosystem over a 30-year period. The analysis revealed a 'double-peak' curve in the total carbon storage of the forest ecosystem of Hainan Island from 1990 to 2020, as shown in Fig. 3 . The peaks occurred in 1998 and 2011, with carbon storage reaching 369.94 and 372.80 Tg, respectively. In contrast, carbon storage dropped to 335.09 Tg in 2005, following a valley. Substantial fluctuations are observed in 1997, 2005, and 2013. Further analysis indicated that carbon storage in the forest ecosystem of Hainan Island showed an increasing trend (Fig. 4) from 1991 to 1997. A decreasing trend was observed from 1997 to 2004, followed by an increasing trend from 2004 to 2010. Following that, a decrease occurred from 2010 to 2015, then a stable period from 2015 to 2020 with a slight decline. Figure 4 Carbon storage of the forest ecosystem in Hainan Island from 1990 to 2020 At the global spatial scale, there were local variations in forest carbon storage from 1990 to 2020; however, the overall distribution did not change significantly. Carbon storage is mainly concentrated in the central and southern tropical rainforest areas of Hainan Island, including Qiongzhong, Baisha, and Wuzhishan. In contrast, low-value areas were dispersed in coastal regions, such as Lingao, Lingshui, and Haikou. The analysis revealed that central tropical rainforest areas, which are characterised by high forest quality, are the main hotspots of carbon storage in the forest ecosystem of Hainan Island. The integration of the local Moran’s I index and land-use spatial distribution provided valuable insights into the spatial distribution of carbon storage in the forest ecosystem of Hainan Island. In contrast, coastal areas primarily constitute cold spots, forming a circular distribution pattern across the island. When observing the distribution of carbon storage hot and cold spots from 1990 to 2020, the degree of aggregation of high- and low-carbon storage was more pronounced, indicating a more concentrated aggregation of forests. High–high clustering primarily occurred in forest and grassland areas, whereas high–low clustering primarily occurred in urban areas. High–low clustering can identify the boundary between urban construction and forests on Hainan Island. Low–low clustering indicates areas with less forest carbon storage and more artificial construction. From 1990 to 2020, patches of increase and decrease were dispersed in the southwestern part of Hainan Island (Fig. 5 ). Areas of decreased carbon storage were more numerous than those of increased storage around the built-up coastal areas of Hainan Island. By combining different periods and the local Moran’s I index, we found that the spatial distribution of carbon storage in the forest ecosystem of Hainan Island changed. Between 1990 and 2000, carbon storage in the northwestern region of Hainan Island, particularly in Dongfang City, decreased significantly, transitioning from high–high clustering to non-significance. Areas of low–low clustering increased in the northern and eastern coastal areas of Hainan Island. Between 2000 and 2010, carbon storage in Ledong Li Autonomous County on Hainan Island increased significantly, transitioning from non-significant to high–high clustering. From 1990 to 2020, the overall high–high clustering of carbon storage showed a transition from scattered centres to an overall central aggregation. During the study period, the carbon storage in the northern region of Hainan Island decreased gradually. High–low clustering of forest carbon storage was commonly found near low–low clustering areas along the coast of Hainan Island. The region exhibited significant spatial heterogeneity in land-use types. Driving factors of carbon sequestration differentiation An analysis of the spatial differentiation of the nine factors driving forest ecosystem spatial differentiation on Hainan Island was conducted at intervals of 10 years from 2000 onwards. The factors analysed were NDVI, elevation, slope, nighttime light index, per capita GDP at the city level, average population density at the city level, annual cumulative precipitation, annual average temperature, and cumulative sunshine. Before 2000, there was insufficient data on the relevant factors. Tables 3 – 5 shows the explanatory power index of the geographical detector for the detecting factors. The normalised vegetation index (NDVI) ranged from 0.521 to 0.530, elevation from 0.511 to 0.529, slope from 0.314 to 0.339, nighttime light index from 0.086 to 0.130, and per capita GDP from 0.089 to 0.164. The spatial differentiation of carbon storage was influenced by several factors, including average population density (0.080–0.100), annual cumulative precipitation (0.066–0.144), annual average temperature (0.040–0.124), and cumulative sunshine (0.022–0.130). In different years, NDVI, elevation, and slope, which are the three influencing factors, had explanatory powers exceeding 30%, indicating that they were the dominant factors affecting carbon storage. This analysis is objective and based solely on the data presented. These three factors exhibited fluctuations of less than 2.5% in their explanatory power for carbon storage in forest ecosystems over 20 years, indicating a relatively stable trend of their long-term influence on forest carbon storage. The correlation between NDVI and carbon storage suggests that areas with good vegetation cover usually have higher carbon storage in forests. Similarly, the slope significantly affects forest carbon storage. The latter terrain promotes better vegetation growth and carbon sequestration, whereas steep terrain may lead to soil erosion and organic matter loss, hindering carbon sequestration by plants. The nighttime light index, per capita GDP, population density, annual cumulative precipitation, annual average temperature, and cumulative sunshine showed medium-to-low explanatory power for carbon storage in forest ecosystems. The explanatory power of all factors was consistently below 15% across all periods. Per capita GDP and population density had an impact of approximately 10%, indicating that regional development quality and population concentration affected forest carbon storage. The explanatory power remained stable. The impact of climatic factors on the spatial and temporal differentiation of forest carbon storage on Hainan Island was evident, particularly in years of extreme climatic events. The explanatory power of the annual cumulative precipitation, annual average temperature, and cumulative sunshine varied significantly between 0.4% and 10.8% across all periods. There are dual-factor and nonlinear enhancement interactions between the driving factors, indicating that any combination of factors enhances the spatial differentiation of carbon storage in the forest ecosystem of Hainan Island. In different years, the combined action of elevation and NDVI had the greatest explanatory power for spatial differentiation of forest carbon storage in the region, ranging from 0.645 to 0.655. This text suggests that areas where elevation and NDVI interact may have rich forest carbon storage resources. The interaction between annual average temperature and population density mostly showed a nonlinear enhancement with other factors in different years. This indicates that the synergistic effects of annual average temperature and population density with other factors are notably superior to their individual effects. Therefore, to understand the factors that influence carbon storage in the forest ecosystem of Hainan Island, it is crucial to consider the combined effects of various driving factors. Table 3 Interaction detection results of driving factors in 2000 Vegetation Coverage Index Elevation Slope Nighttime Light Index GDP Population Density Annual Cumulative Precipitation Annual Average Temperature Cumulative Sunshine Duration Vegetation Coverage Index 0.529 Elevation 0.648 0.511 Slope 0.577 0.534 0.339 Nighttime Light Index 0.575 0.546 0.404 0.130 GDP 0.551 0.539 0.373 0.155 0.089 Population Density 0.548 0.528 0.371 0.195 0.157 0.080 Annual Cumulative Precipitation 0.557 0.566 0.415 0.178 0.139 0.151 0.070 Annual Average Temperature 0.539 0.547 0.385 0.187 0.144 0.123 0.082 0.040 Cumulative Sunshine Duration 0.547 0.565 0.409 0.145 0.115 0.110 0.091 0.099 0.022 The blue area represents nonlinear enhancement, and the green area represents factor enhancement. Table 4 Interaction detection results of driving factors in 2010 Vegetation Coverage Index Elevation Slope Nighttime Light Index GDP Population Density Annual Cumulative Precipitation Annual Average Temperature Cumulative Sunshine Duration Vegetation Coverage Index 0.530 Elevation 0.645 0.518 Slope 0.582 0.537 0.314 Nighttime Light Index 0.556 0.549 0.384 0.086 GDP 0.578 0.537 0.367 0.184 0.114 Population Density 0.547 0.544 0.361 0.178 0.215 0.100 Annual Cumulative Precipitation 0.558 0.586 0.399 0.120 0.193 0.167 0.066 Annual Average Temperature 0.557 0.550 0.376 0.163 0.174 0.214 0.153 0.124 Cumulative Sunshine Duration 0.555 0.582 0.387 0.202 0.168 0.168 0.116 0.156 0.065 Table 5 Interaction detection results of driving factors in 2020 Vegetation Coverage Index Elevation Slope Nighttime Light Index GDP Population Density Annual Cumulative Precipitation Annual Average Temperature Cumulative Sunshine Duration Vegetation Coverage Index 0.521 Elevation 0.655 0.529 Slope 0.591 0.552 0.334 Nighttime Light Index 0.554 0.564 0.418 0.101 GDP 0.580 0.547 0.402 0.254 0.164 Population Density 0.531 0.549 0.372 0.164 0.229 0.080 Annual Cumulative Precipitation 0.581 0.614 0.440 0.198 0.247 0.212 0.144 Annual Average Temperature 0.542 0.599 0.452 0.185 0.243 0.149 0.190 0.073 Cumulative Sunshine Duration 0.585 0.614 0.436 0.183 0.211 0.203 0.161 0.167 0.130 Discussion Feasibility of using the inVEST model for evaluating carbon storage in forests of Hainan Island The InVEST model was used in this study to calculate the forest carbon storage on Hainan Island. The model relies on a carbon density table derived from field surveys of different vegetation types on the Hainan Island. The data used in the table include information from the 2021 forest resources survey of Hainan Province and observations from national forest ecosystem observation stations, such as Wenchang and Jianfengling. Compared to previous studies, such as Duan Xuan et al. (2022), who derived carbon density tables by correcting carbon density in the Pearl River Delta region based on average rainfall, temperature, and elevation data from Guangdong and Hainan, the results obtained in this study are more accurate. This is true for forest areas, where both aboveground and belowground biomass carbon densities are higher. In addition, the table includes carbon pools from litter, with a soil carbon density that is four times higher. The study by Liu et al. adheres to conventional academic structures and maintains regular author and institution formatting. (2022) used a comprehensive dataset to determine the ecological carbon density of terrestrial areas in China in the 2010s. They selected data points from the profiles of Hainan Island and calculated the corrected carbon density of the region. This study found that the total forest carbon density was like the data measured and compiled by Xuan et al., although there were discrepancies in the aboveground and soil carbon pools. Thus, the carbon density results obtained by Xuan et al. can be used to examine carbon storage in the forest ecosystem of Hainan Island. Obtaining clear satellite images within the same period is challenging because of the tropical and subtropical climate of Hainan Island. Furthermore, discrepancies in satellite image stitching, land cover type delineation, and calibration have been reported by different researchers, particularly at different resolutions. Large-scale remote sensing studies have been conducted on carbon sinks on Hainan Island using satellite data, such as Sentinel-2A from the European Space Agency (ESA), with a spatial resolution of 300 m and vegetation cover data from MODIS from 2000 to 2019. Han et al. ( 2022 ) used this data. However, comprehensively validating the classifications obtained is challenging in previous studies analysing land cover data from satellite remote sensing, due to the limited number of training samples and computational capabilities. This study uses the ‘China 30 m Annual Land Cover Data from 1990 to 2022’ by Yang Jie et al. (2021) from Wuhan University, which is based on 5,463 visual interpretation samples. Furthermore, it is worth noting that Liu et al. ( 2022 ) found that the rate of change of forest area on Hainan Island from 1992 to 2019 was 1.3%. Incorporating this rate of change into the results obtained in this study can effectively account for the variation in forest carbon storage with changes in the forest area. According to the results of the third forest resource survey of Hainan Province conducted in 2021, the total forest area in the province was 2,283,242.74 ha, which closely aligns with the forest area of 2,212,257.33 ha in 2020 used in this study, indicating the reliability of the remote sensing classification data used in this study. Impact of policies on forest carbon storage in Hainan Island Forest carbon storage policies have a crucial impact on the development of industries in Hainan Island. Over the past 30 years, the forest carbon storage on the island has undergone five stages of change. During Stage 1 (1991–1997), the forest ecosystem carbon storage increased from 350.92 to 368.95 Tg. Two significant forestry policies have played pivotal roles in China. In 1987, the ‘Directive on Strengthening Forest Resource Management in Southern Collective Forest Areas to Strictly Prohibit Illegal Logging’ was implemented. This directive reduced deforestation and prevented a decrease in forestland area. The second most significant event was the nationwide cessation of natural forest logging in 1994. This has led to the implementation of various systems primarily based on joint production and contract responsibility, resulting in the fastest afforestation period in the history of Hainan. The land development and utilisation were strictly controlled. During Stage 2 (1997–2004), the carbon storage decreased from 369.95 to 338.73 Tg. The arable land area increased from 884,496.51 to 983,509.74 hm 2 , possibly due to increased demand and higher prices in the tropical fruit and vegetable markets, driving an increase in the planting area for these crops. However, this increase in arable land came at the cost of forest areas, which were encroached upon or converted to farmland, leading to a decrease in the forestland area. During Stage 3 (2004–2010), carbon storage increased from 338.73 to 372.80 Tg. The increase in forest area could be credited to policy implementation, like the “Opinions on Further Improving the Measures for Returning Farmland to Forests”. This policy restored arable land with forest attributes to forest land, increasing forest area. In addition, a reform pilot of the collective-forestland tenure system was initiated in Hainan Province in 2007, which played a role in protecting and managing forest resources, increasing forest carbon storage. During Stage 4 (2010–2015), these policies and reforms continued to have a positive impact on forest conservation and management. Carbon storage decreased from 372.80 to 350.16 Tg, possibly due to policies such as the ‘Opinions on Promoting the Construction and Development of Hainan International Tourism Island’, which resulted in urban expansion and increased construction land area. During Stage 5 (2015–2020), the carbon storage remained relatively stable, with a slight downward trend. In 2017, the Hainan Provincial Government implemented the ‘Regulations on Balanced Management of Land Occupation and Compensation for Forest Land’ to ensure that the forest land area did not decrease. The Seventh Provincial Party Congress of the same year emphasised continued adherence to the policy of 'putting ecology first'. In 2018, the National Reserve Forest Construction Plan and Guiding Principles for National Ecological Civilisation Construction were implemented as policies. An analysis of almost 30 years of changes in forest ecosystem carbon storage on Hainan Island revealed that government forestry policies have played a crucial role in protecting and restoring the forestry ecological pattern in Hainan Province. Key factors influencing forest carbon storage on Hainan Island This study found that forest carbon storage on Hainan Island is primarily influenced by the NDVI, elevation, and slope. Other factors, including nighttime light index, per capita GDP, population density, annual cumulative precipitation, mean annual temperature, and cumulative sunshine, also contribute to the spatial differentiation of forest carbon storage within the region. These findings support the results of Gao et al. ( 2023 ), who used field experiments and remote sensing data to demonstrate that carbon storage in forest ecosystems on Hainan Island decreased with increasing elevation. The correlation between slope and forest carbon storage is consistent with the research conducted by Han et al. ( 2023 ) using geographic detectors, which highlights the influence of the mean annual NDVI on the spatial differentiation of carbon storage. Furthermore, although Wang et al. ( 2023 ) and Pereira et al. ( 2015 ) suggested that the soil type is a significant factor in the spatial distribution of forest carbon storage, future studies should explore and refine this relationship. In conclusion, it is crucial to understand the factors that influence forest ecosystem carbon storage on Hainan Island. Future studies should examine the interactive relationships between various factors, particularly the significant impacts of elevation and the vegetation coverage index on forest carbon storage. To gain a more comprehensive understanding of the spatial differentiation of forest ecosystem carbon storage on Hainan Island, it is necessary to further explore the synergistic effects of the different factors. Prospects The InVEST model was used to calculate forest ecosystem carbon storage data for Hainan Island over the past 30 years. The geographical detector method was subsequently used to quantitatively analyse the driving factors that influence the spatial distribution of carbon storage in the region. In contrast to previous large-scale studies on forest carbon storage on Hainan Island (Gao et al., 2023 ; Liu Q et al., 2022 ; Cao et al., 2002 ), which relied primarily on qualitative analysis, this study adopted a novel approach by adopting the geographical detector method. This method offers a new perspective for investigating ecosystem carbon storage on Hainan Island and provides valuable insights for future research on carbon sequestration issues in the region. It is important to note that the geographical detector method does not inherently address the directional effects of the factor importance. Future research could enhance the analysis by incorporating spatial statistical methods such as principal component analysis, analytic hierarchy process, system clustering analysis, and discriminant analysis. In further research on forest ecosystem carbon storage on Hainan Island, the use of geographic weighted regression (GWR) can provide a more comprehensive understanding of both direction and magnitude of the effects of driving factors on forest carbon storage. Guo et al. ( 2015 ) used this method to estimate regional forest carbon storage. To date, no study has explored the relationship between survey factors and carbon storage using models of forest ecosystem carbon storage on Hainan Island. However, it is crucial to understand the relationship and spatial correlation characteristics between forest carbon storage and survey factors to establish models for estimating regional forest carbon storage and its distribution. A factor estimation model for forest ecosystem carbon storage on Hainan Island can be developed using satellite bands, topographic, climate, and human activity data. This approach leads to more accurate results in the model estimation and prediction of forest ecosystem carbon storage based on forest inventories (Gao et al., 2023 ; Cao et al., 2002 ; Lin et al., 2000) or remote sensing image data (Liu et al., 2022 ). This will help achieve high-quality monitoring records year after year, advancing the early realisation of the 'double carbon' goal. Based on the information above, we now comprehensively discuss how to help Hainan Province achieve carbon neutrality through forest carbon sink in the future. As we move forward with carbon-neutral initiatives, we should focus on forest carbon sequestration and turn its potential and cost advantages into real benefits for carbon reduction. Focusing on Wuzhishan City, Qiongzhong City, Baisha City, Baoting City, and other places with high forest carbon storage, a forest carbon sequestration pilot base will be established. Regarding policy formulation, afforestation, quality improvement, quota cutting, mechanism innovation, etc., resources should be channelled towards increasing forest carbon sink capacity and improving carbon sequestration. In terms of specifically improving carbon sink capacity, there should a focus on supporting state-owned forest farms, collective economic cooperation organizations, enterprises and large forest farmers and other subjects. The application and demonstration of forestry carbon sequestration methodology and new technologies will be the main direction of the pilot base, and the first exploration of forestry carbon sequestration trading and carbon sequestration capacity potential improvement will be carried out at the same time. In addition, there must be a focus on coordinated promotion and pay attention to spillover effects. The terrain of Hainan Province is complex (with many high and low points), and there are great differences in forest resources, economic development level, energy consumption structure and carbon emission level among different regions. Therefore, the pace at which peak carbon emissions and carbon neutrality is achieved across different regions should be adapted to local conditions, rather than generalized. A global market for carbon emissions trading should be introduced, with plans to include forestry carbon sinks in the carbon market. Carbon sink trading between regions with poor economy but rich forest carbon reserves and regions with developed economy but relatively poor carbon reserves will become a reality. An accurate assessment of the ecological and economic value of forests in the region should contributes towards their sustainability. Research limitations This study utilized 30 years of consecutive land use remote sensing image data to classify land use types within urban areas were uniformly classified as forest land, without further subdivision of the classification units, which may have introduced certain errors into the experimental results. Given that Hainan Island has abundant original or contiguous forest resources, according to the Third National Land Survey of Hainan Province, urban land and town land encompassed 35,555.06 and 60,425.69 hectares in 2019, respectively. However, according to the forest resources inventory data of Hainan Province in 2018, the forest area was 1,944,900 hectares in 2018; the total land area only accounted for 4.93% of the forest area. Meanwhile, the remote sensing data used in this study showed that the area of construction land in Hainan Province in 2020 accounted for 2.05%, while forest land accounted for 64.55%. Therefore, treating forest land as a homogenous monolith, may impact experimental results, even though the impact would most likely be negligible. In future forest division studies, it will be necessary to focus on using a large amount of measured data or sufficient on-site monitoring data as the classification learning objects to improve the accuracy of regional ecosystem carbon storage assessment. In this study, the geographic detector was used to discretize the data, but only one discretization method was adopted. Specifically, results obtained via multiple discretization methods were not considered. However, the comprehensive analysis of multiple discretization methods may help to understand the relationship between data features and factors more comprehensively, to obtain more accurate results. In addition, although the geographic detector can show the important explanatory force of the factor, it cannot explain the direction of the explanatory force of the factor. Therefore, to comprehensively elucidate the role of various factors, the use of correlation factor analysis will be necessary. However, the value of correlation analysis results will lose some spatial explanatory power, so correlation analysis is mainly used to judge the direction of the influence factors. Overall, the shortcomings of the research are mainly in the selection of data processing and analysis methods, and the comprehensive utilization of various methods is not fully considered, which may lead to the limitation of the results. In the future, when using geo-detectors, it could be beneficial to combine more comprehensive auxiliary data analysis methods. For example, the use of geographic detectors and GWR or principal component analysis can provide a more comprehensive understanding of the relationship and degree that affect the spatial differentiation of forest carbon stocks, to ensure that the obtained data results contain both spatial interpretation and more accurate data interpretation. In this study, the geographic detector was used to discretize the data, but only one discretization method was adopted, and the comprehensive analysis of the results obtained by multiple discretization methods was not considered. The comprehensive analysis of multiple discretization methods may help to understand the relationship between data features and factors more comprehensively, to obtain more accurate results. In addition, although the geographic detector can show the important explanatory force of the factor, it cannot explain the direction of the explanatory force of the factor. Therefore, yield a more accurate understanding of factors, correlation factor analysis is also needed. However, the value of correlation analysis results will lose some spatial explanatory power, so correlation analysis would be mainly used to judge the direction of the influence factors. Overall, the shortcomings of the research are mainly in the selection of data processing and analysis methods, and the comprehensive utilization of various methods is not fully considered, which limit the rigour of the results. In the future, when using geo-detectors, it will be worth combining more comprehensive auxiliary data analysis methods such as: The use of geographic detectors and GWR or principal component analysis can provide a more comprehensive understanding of the relationship and degree that affect the spatial differentiation of forest carbon stocks. This approach would yield results that are not only accurate but that also have a spatial component. Innovation points In the past, ess attention was paid to the year-by-year changes of forest carbon stocks in Hainan Island. Here, we highlight the importance of studying the pattern of forest carbon stocks in Hainan Island in a long time series. The use of InVEST model to analyse forest carbon stocks provide a systematic and quantitative method for research. By analysing data over a long period of time, the dynamic change trend of forest carbon stocks can be more accurately understood, and era-specific changes in the area of forests can be easily explored. From the perspective of forestry policy, the causes and impact time of forest area change are viewed to provide scientific basis for the formulation of long-term forest management strategies. For the first time in Hainan Island, the correlation analysis of forest carbon storage factors was conducted using a geographic detector. The advantage of this method lies in the fact that it facilitates an exploration of nine spatial factors, including a measure of their spatial differentiation. In addition, this approach allows for an exploration of multiple independent variables and the degree of impact on forest carbon storage. Through the lens of space, we were able to explain whether the spatial distribution of important factors affecting forest carbon stocks is similar to the amount of forest carbon stocks. Simultaneously, by using the interpretation data of the third phase, we determined whether the interpretation strength of factors for forest carbon stocks has been improved, revealing the mechanism underlying the change of forest carbon stocks. Overall, this approach will allow for a better understanding of the carbon cycle process of forest ecosystem, providing a theoretical basis for forest management and protection in other areas in the future. In addition, this study has brought forth new research avenues in the context of analysing factors that disturb forest carbon storage in the context of large space and long-time series. Conclusions Between 1990 and 2020, there were substantial changes in land-use types on Hainan Island. The proportions, from largest to smallest, were forest land, cropland, water bodies, construction land, grassland, unused land, and shrubbery. The conversion between cropland and forestland had the highest proportion, and changes in forestland area were the primary direct cause of forest carbon storage variation. Between 1990 and 2020, the total carbon storage of the forest ecosystem on Hainan Island (excluding Sansha City) fluctuated between 335.09 and 372.80 teragrams of carbon (TgC). The lowest carbon storage was recorded in 2004 at 335.09 TgC, whereas the highest was in 2010 at 372.80 TgC. Between 1991 and 1997, forest ecosystem carbon storage showed an upward trend followed by a downward trend from 1997 to 2004. From 2004 to 2010, there was an upward trend followed by a downward trend from 2010 to 2015. Forest ecosystem carbon storage was stable from 2015 to 2020. Forest ecosystem carbon storage on Hainan Island exhibited spatial clustering ( p < 0.01), with higher values concentrated in central mountainous regions and lower values in coastal areas. The carbon storage levels on Hainan Island varied significantly across different regions. The eastern part of the island has higher carbon storage levels than the western part, whereas the southern regions exhibit higher values than the northern regions. The forest ecosystem carbon storage also showed significant variability with increasing elevation. Geographical detector analysis revealed that topography, climate, and anthropogenic factors were the main factors influencing the spatial differentiation of forest ecosystem carbon storage on Hainan Island. The distribution of forest carbon storage was significantly shaped by the NDVI (0.521–0.530), elevation (0.511–0.529), and slope (0.314–0.339), which emerged as the primary drivers. These factors had a greater impact than others. In addition, the interaction detection results highlighted the synergistic effects of various driving factors, surpassing their individual impacts on forest ecosystem carbon storage. The synergy between the NDVI and elevation exhibited the strongest nonlinear enhancement (0.645–0.655), which is noteworthy. Declarations Acknowledgments: We would like to thank Editage (www.editage.cn) for English language editing. Funding: This research is supported by the Special Fund for Basic Research Expenses of Central public welfare research institutes; the grant number is (CAFYBB2021ZE001). Xu Wang received financial support from the project Conflicts of interest/Competing interests: There are no conflicts of interest. Availability of data and material: Not Applicable Code availability: Not Applicable Authors' contributions: Pi. and Wang and Zang. wrote the main manuscript text; Song. and Wang. and Zhou. drafted the work; Pi. and Guo. and BaoYin. and Zhao. acquisition, analysis of data; Li. and Qiu. and Wu. interpretation of data; All authors reviewed the manuscript. References Cao J, Zhang Y, Liu Y (2002) Changes in forest biomass carbon storage in Hainan Island over the last 20 years. Geogr Res 21(5):551–560. Chai Y, Yu Y (2022) Research on innovative paths of Hainan tropical rainforest national park system. West For Sci 51(1):155–160. (In Chinese) Duan X, Gong W, Sun Y, Liu T, Qiu X, Zhang Y (2022) Land use change in the coastal zone of Hainan Island and its spatiotemporal evolutionary impact on carbon storage. Bull Soil Water Conserv 42(5):301–311. (In Chinese) Fan L, Cai T, Wen Q, Han J, Wang S, Wang J, Yin C (2023) Scenario simulation of land use change and carbon storage response in Henan Province. Ecol Indic 154(7):110660. Fu Y, Huang M, Gong D, Lin H, Fan Y, Du W (2023) Dynamic simulation and prediction of carbon storage based on land use/land cover change from 2000 to 2040: A case study of the Nanchang urban agglomeration. Remote Sens 15(19):4645. https://doi.org/10.3390/rs15194645 Gao S, Chen Y, Chen Z, Lei J, Wu T (2023) Carbon storage and spatial distribution characteristics of forest ecosystems on Hainan Island. Acta Ecol Sin 43(9):3558–3570. (In Chinese) Guan M, Xiong C (2022) The net spatio-temporal impact of the international tourism is-land strategy on the ecosystem service value of Hainan Island: A counterfactual analysis. Land 11(10):1694. https://doi.org/10.3390/land11101694 Guo H, Zhang M, Xu L, et al (2015) Estimation of regional forest carbon storage based on geographically weighted regression. J Zhejiang A&F University 32(4):497–508. Han J, Zhang G, Li W, et al (2022) Analysis of vegetation ecological quality change characteristics in Hainan Island over the past 20 years. Sci Ecol Sin 41(01):20–30. https://doi.org/10.14108/j.cnki.1008-8873.2022.01.003 Han Y, Ding S, Yang T (2023) Spatial and temporal distribution of carbon storage in the ecosystem of Southern Taihang Mountains in Southern Shanxi and its driving factors. Chin J Environ Sci 43(3):1298–1306. Hanning R (2003) Spatial data analysis: Theory and practice. Cambridge University Press He C, Zhang D, Huang Q, Zhao Y (2016) Assessing the potential impacts of urban expansion on regional carbon storage by linking the LUSD-urban and InVEST models. Environ Model Softw 75:44–58. https://doi.org/10.1016/j.envsoft.2015.09.015 Hou K, Li X, Wang JJ, Zhang J (2016). An analysis of the impact on land use and ecological vulnerability of the policy of returning farmland to forest in Yan’an, China. Environ Sci Pollut Res Int 23(5):4670–4680. Kang Hou et al (2016) An analysis of the impact on land use and ecological vulnerability of the policy of returning farmland to forest in Yan’an, China. Environmental Science and Pollution Research 23: 4670–4680. Kothandaraman S, Dar JA, Sundarapandian S, Dayanandan S, Khan ML (2020) Ecosystem-level carbon storage and its links to diversity, structural and environmental drivers in tropical forests of Western Ghats, India. Sci Rep 10(1):13444. https://doi.org/10.1038/s41598-020-70313-6 Lei JR, Chen ZZ, Chen XH, et al. (2020) Spatiotemporal changes in land use and ecosystem service value on Hainan Island from 1980 to 2018. Acta Ecol Sin 40(14):4760–4773. Li P, Chen J, Li Y, Wu W (2023) Using the InVEST-PLUS model to predict and analyze the pattern of ecosystem carbon storage in Liaoning Province, China. Remote Sens 15(16):4050. Li W, Chen J, Zhang ZJFE (2020). Forest quality-based assessment of the Returning Farmland to Forest Program at the community level in SW China Management 461:117938. Wenqing L, Chen J, Zhang Z (2020) Forest quality-based assessment of the Returning Farmland to Forest Program at the community level in SW China. Forest Ecology and Management 461: 117938. Lin MZ, Zhang YL Dynamic changes and sustainable development of forest resources in Hainan Island. Ecol Sci 2000(4):84–89. Liu Q, Yang D, Cao L, Anderson B (2022) Assessment and prediction of carbon storage based on land use/land cover dynamics in the tropics: A case study of Hainan Island, China. Land 11(2):244. https://doi.org/10.3390/land11020244 Meng Q, Wu ZT, Du ZQ, Zhang H (2021) Quantitative impact of regional vegetation cover based on geographic detector: A case study of the Beijing-Tianjin Sand source area. Chin J Environ Sci 41(2):826–836. Pan YD, Birdsey RA, Fang JY, et al. (2011) A large and persistent carbon sink in the world’s forests. Science 333(6045):988–993. https://doi.org/10.1126/science.1201609 Pereira OJR, Montes CR, Lucas Y, Santin RC, Melfi AJ (2015) A multi-sensor approach for mapping plant-derived carbon storage in Amazonian podzols. Int J Remote Sens 36(8):2076–2092. https://doi.org/10.1080/01431161.2015.1034896 Phillips OL, Malhi Y, Higuchi N, et al. (1998) Changes in the carbon balance of tropical forests: Evidence from long-term plots. Science 282(5388):439–442. https://doi.org/10.1126/science.282.5388.439 Piyathilake I, Udayakumara E, Ranaweera L, Gunatilake SJMES (2022) Modeling predictive assessment of carbon storage using InVEST model in Uva province, Sri Lanka. Environment 8:2213–2223. Ren H, Li L, Liu Q et al. (2014) Spatial and temporal patterns of carbon storage in forest ecosystems on Hainan Island, Southern China. PLoS One 9(9):e108163. https://doi.org/10.1371/journal.pone.0108163 Tallis H, Icketts T, Guerry A (2013) In: VEST User’s Guide: Integrated valuation of environmental services and tradeoffs. The Natural Capital Project, Stanford. Sharp, R., et al (2014) InVEST user’s guide: integrated valuation of environmental services and tradeoffs. The Natural Capital Project. In Stanford Woods Institute for the Environment. University of Minnesota's Institute on the Environment, the Nature Conservancy & WW Foundation Stanford. Tang G (2019) Digital elevation model of China (1KM). A big earth data platform for three poles. Wang CW, Luo JJ, Tang HH (2023) Analysis of spatiotemporal differentiation driving forces of ecosystem carbon storage in the Taihang Mountains area based on the InVEST model. Ecol Environ Sci 32(2):215–225. Wang JF, Xu CD (2017) Geographic detector: Principles and prospects. Acta Geogr Sin 72(1):116–134. Wang JJ Study on the dynamic characteristics and influencing factors of carbon storage of main tree species in forest arboreal layer in Anze County, Shanxi Province. Master’s thesis. Shanxi University of Finance & Economics. Wang Junjie (2022) Study on dynamic change characteristics and influencing factors of carbon storage of main tree species in forest tree layer in Anze County, Shanxi Province. Dissertation, Shanxi University of Finance and Economics Wang P (2010) Analysis of the development of tropical fruit industry in Hainan Province. Trop Agric Eng 34(2):67–70. Wang Z Study on the spatiotemporal changes and influencing factors of forest carbon storage in Hangzhou city based on the CASA model. Master’s thesis. Zhejiang A&F University. Wang Z (2021) Study on spatial-temporal changes and influencing factors of forest carbon storage in Hangzhou based on CASA model. Dissertation, Zhejiang A & F University Wang ZH, Liu HM, Guan QW, et al (2011) Carbon storage and density of urban forest ecosystems in Nanjing. J Nanjing For Univ 54(4):18. Wang ZH, Liu HM, Guan QW, et al (2011) Carbon storage and density of urban forest ecosystems in Nanjing. Journal of Nanjing Forestry University 54(4):18. Wei S, Yang Y, Lin Z, Zhang D (2014). A model of the new style of urbanization for Hainan province in the context of international tourism-island construction. Resources 35:14–18. Wei, S, Yang, Y, Lin, Z, Zhang, D. (2014) A model of the new style of urbanization for Hainan province in the context of international tourism-island construction. Shanghai Land and Resources, 35(1), 14–18. Wu Y, Shi K, Chen Z, Liu S, Chang Z (2021) Developing improved time-series of improved DMSP-OLS-like data (1992–2019) in China by integrating DMSP-OLS and SNPP-VIIRS. IEEE Trans Geosci Remote Sensing 60:1–14. https://doi.org/10.1109/TGRS.2021.3135333 Xu XL China annual vegetation index (NDVI) spatial distribution dataset. Data registration and publishing system of resource and environmental science data center. Chinese Academy of Sciences, 2018 https://doi.org/10.12078/2018060601 Yan LX Research on strategies to enhance the competitiveness of tropical fruit industry in Hainan Province. Master’s thesis. South China University of Tropical Agriculture. Yan Linxia (2005) Research on countermeasures to enhance the competitiveness of tropical fruits in Hainan Province. Dissertation, South China Tropical Agricultural University Yang J, Huang X (2021) The 30 m annual land cover dataset and its dynamics in China from 1990 to 2019. Earth Syst Sci Data 13(8):3907–3925. https://doi.org/10.5194/essd-13-3907-2021 Zhang L, Zhou G, Ji Y, Bai Y (2016) Spatiotemporal dynamic simulation of grassland carbon storage in China. Sci China Earth Sci 59(10):1946–1958. https://doi.org/10.1007/s11430-015-5599-4 Zhang Y, Zhang W, Ding M (2004) Differences in estimating carbon stocks based on land use/cover classification system: A case study of forests in Hainan Island. Prog Geogr 06:63–70. Zhu W, Zhang J, Cui Y, et al (2019) Assessment of ecosystem carbon stock based on scenarios of land use change: A case study of Qihe River Basin in Taihang Mountains. Acta Geogr Sin 74(3):446–459. Zhu Z, Ma X, Hu H (2021) Spatiotemporal evolution and prediction of ecosystem carbon stock in Guangzhou based on coupled FLUS-InVEST model. Bull Soil Water Conserv 41(2):222–229 + 239 Zou W, He Y, Ye B, et al (2021) Study on ecosystem carbon stock of Qianjiangyuan National Park based on InVEST model. J Cent S Univ For Technol 41(3):120–128. 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-4105908","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":281765516,"identity":"f4238f5f-9f7f-42ac-b94b-2a2264df5b95","order_by":0,"name":"ZhiHao Pi","email":"","orcid":"","institution":"Hainan University","correspondingAuthor":false,"prefix":"","firstName":"ZhiHao","middleName":"","lastName":"Pi","suffix":""},{"id":281765517,"identity":"b4599164-c759-4100-a943-cb38cf5bd037","order_by":1,"name":"Xu Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9UlEQVRIiWNgGAWjYFACxgcQmpmx4cAHBgZ+IrQwG0BodubGgzMYGCQbiNfCz958mIcYLfzth9kkfu6oTdxwmLHhsE2NnQQD++EHDD934NYicSaZTbL3zHFjA5CWnGPJEgw8aQaMvWdwazFgyD92g7ftmBxYS24Dcx0DQw4wKNrwaOF/zHbzb9sxHrAWy4Z6CQb+NwS0SCSz3eZtq4HYAkQSDBIEbJG48Zj9t2zbAWNJoPqDPceOS7BJPDM42ItHC39/MrPh27a6xL7zxx9/+FFTLcHPn/zwwU88WqDgMILJBsQHCGpgYKgjQs0oGAWjYBSMWAAA1QdQ4p94mRkAAAAASUVORK5CYII=","orcid":"","institution":"Chinese Academy of Forestry","correspondingAuthor":true,"prefix":"","firstName":"Xu","middleName":"","lastName":"Wang","suffix":""},{"id":281765518,"identity":"e996f5b2-2be1-49df-94ba-d71331696f27","order_by":2,"name":"Zhuo Zang","email":"","orcid":"","institution":"Central South University of Forestry \u0026 Technology","correspondingAuthor":false,"prefix":"","firstName":"Zhuo","middleName":"","lastName":"Zang","suffix":""},{"id":281765519,"identity":"3c07be1c-33ff-40f0-9cd2-89d034c5809d","order_by":3,"name":"XiQang Song","email":"","orcid":"","institution":"Hainan University","correspondingAuthor":false,"prefix":"","firstName":"XiQang","middleName":"","lastName":"Song","suffix":""},{"id":281765520,"identity":"570f746e-f7cd-4b18-b4ac-5ad40bcf1ea7","order_by":4,"name":"GuangYi Zhou","email":"","orcid":"","institution":"Chinese Academy of Forestry","correspondingAuthor":false,"prefix":"","firstName":"GuangYi","middleName":"","lastName":"Zhou","suffix":""},{"id":281765521,"identity":"27b4f14c-af85-4c12-af07-7879081453dd","order_by":5,"name":"Hao Guo","email":"","orcid":"","institution":"Hainan University","correspondingAuthor":false,"prefix":"","firstName":"Hao","middleName":"","lastName":"Guo","suffix":""},{"id":281765522,"identity":"95018858-634c-4269-a358-9aae9fb15886","order_by":6,"name":"ManDa BaoYin","email":"","orcid":"","institution":"Chinese Academy of Forestry","correspondingAuthor":false,"prefix":"","firstName":"ManDa","middleName":"","lastName":"BaoYin","suffix":""},{"id":281765523,"identity":"0afec740-3bbf-4fa0-99f0-13a0d8565055","order_by":7,"name":"HouBen Zhao","email":"","orcid":"","institution":"Chinese Academy of Forestry","correspondingAuthor":false,"prefix":"","firstName":"HouBen","middleName":"","lastName":"Zhao","suffix":""},{"id":281765524,"identity":"4ba7c98d-b38f-4633-8c3e-c2f748a4cfd3","order_by":8,"name":"ZhaoJia Li","email":"","orcid":"","institution":"Chinese Academy of Forestry","correspondingAuthor":false,"prefix":"","firstName":"ZhaoJia","middleName":"","lastName":"Li","suffix":""},{"id":281765525,"identity":"55577235-b1bd-46d2-b520-2b22709ca275","order_by":9,"name":"ZhiJun Qiu","email":"","orcid":"","institution":"Chinese Academy of Forestry","correspondingAuthor":false,"prefix":"","firstName":"ZhiJun","middleName":"","lastName":"Qiu","suffix":""},{"id":281765526,"identity":"e79ad0c0-bbe2-4a07-8352-ebf5e8ee956e","order_by":10,"name":"ZhongMin Wu","email":"","orcid":"","institution":"Chinese Academy of Forestry","correspondingAuthor":false,"prefix":"","firstName":"ZhongMin","middleName":"","lastName":"Wu","suffix":""}],"badges":[],"createdAt":"2024-03-15 07:46:54","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4105908/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4105908/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":53227536,"identity":"9f7dbb85-5199-4b83-b787-54477daa6604","added_by":"auto","created_at":"2024-03-22 06:43:53","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1511417,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial Distribution of Land Use on Hainan Island from 1990 to 2020\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-4105908/v1/793dcfd510c5ba54016e1aed.png"},{"id":53228262,"identity":"4dadae89-8685-4162-b05e-2a8421bff85f","added_by":"auto","created_at":"2024-03-22 06:51:54","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":711443,"visible":true,"origin":"","legend":"\u003cp\u003eLand Use Transfer Matrix of Hainan Island from 1990 to 2020\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-4105908/v1/02948d4d7aeeda5ea5df2f32.png"},{"id":53227538,"identity":"3f36ead0-7261-4aa3-8267-9fc2fa1146ca","added_by":"auto","created_at":"2024-03-22 06:43:54","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1870241,"visible":true,"origin":"","legend":"\u003cp\u003eChanges in forest carbon storage on Hainan Island 1990 to 2020\u003c/p\u003e","description":"","filename":"Fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-4105908/v1/9f50f7d6bced1119686be3d0.png"},{"id":53227539,"identity":"0c6c2b86-20c9-41de-8f31-b4dc1151a6a8","added_by":"auto","created_at":"2024-03-22 06:43:54","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":201677,"visible":true,"origin":"","legend":"\u003cp\u003eCarbon storage of the forest ecosystem in Hainan Island from 1990 to 2020\u003c/p\u003e","description":"","filename":"Fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-4105908/v1/d728ec78888d0a8ce2b26491.png"},{"id":53227537,"identity":"48124d42-2f97-4ba2-ac3f-0decbb0c089d","added_by":"auto","created_at":"2024-03-22 06:43:53","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1439679,"visible":true,"origin":"","legend":"\u003cp\u003eAnselin Local Moran’s I of carbon storage in forest ecosystems on Hainan Island from 1990 to 2020\u003c/p\u003e","description":"","filename":"Fig5.png","url":"https://assets-eu.researchsquare.com/files/rs-4105908/v1/5bdb3ab94b13a21cb8a3a036.png"},{"id":53271004,"identity":"42ed5f0c-bf00-49f9-822a-0797e7c2fe53","added_by":"auto","created_at":"2024-03-22 16:36:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3457472,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4105908/v1/fe79eb3f-e095-4408-a7ff-72e6f856ae35.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Spatiotemporal variation in carbon sequestration in the forest ecosystem of Hainan Island over a 30-year period and its driving factors","fulltext":[{"header":"Introduction","content":"\u003cp\u003eTropical rainforests are crucial forest ecosystems worldwide, covering just 12% of Earth\u0026rsquo;s land area and hosting over 50% of the world\u0026rsquo;s species, accounting for 40% of the carbon reservoirs in terrestrial ecosystems (Phillips et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e1998\u003c/span\u003e). They also represent 55% of global forest carbon stocks (Pan et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Minor alterations in carbon sequestration in tropical rainforest ecosystems can substantially affect ecosystems and global climate change. Estimating forest ecosystem carbon storage, analysing spatiotemporal variation patterns, and exploring the mechanisms influencing carbon sequestration capacity in tropical forest ecosystems are important for carbon management in tropical forest ecosystems.\u003c/p\u003e \u003cp\u003eHainan Island hosts China\u0026rsquo;s largest expanse of tropical forests, forming a concentrated, well-preserved, and contiguous collection of \u0026lsquo;island-like\u0026rsquo; tropical forests within the nation (Chai et al., 2022). It is an essential component of global tropical forest ecosystems (Kothandaraman et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Hai et al., 2014). Since the establishment of Hainan Province in 1988, rapid socioeconomic development and urban expansion have occurred, resulting in notable changes in land use that have profoundly affected the structure and function of the island\u0026rsquo;s ecosystem.\u003c/p\u003e \u003cp\u003eSeveral domestic and international scholars have conducted extensive research on forest carbon storage in Hainan. Cao et al. (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2002\u003c/span\u003e) examined changes in carbon storage within Hainan Island forests from 1979 to 1999. Zhang et al. (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2004\u003c/span\u003e) analysed the variations in carbon storage attributed to different forest vegetation classification systems using forest inventory data and diverse land use/cover classification systems. Ren et al. (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2014\u003c/span\u003e) estimated the distribution pattern of carbon storage across forest ecosystems on Hainan Island over four survey periods from 1993 to 2008. The study incorporated additional estimates from the understory vegetation, litter layer, and soil layer. Liu et al. (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) investigated the relationship between land use/cover and carbon storage on Hainan Island from 1992 to 2019. Gao et al. (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) calculated the carbon storage and spatial distribution characteristics of forest ecosystems on Hainan Island in 2023, using inventory data and remote sensing images. They enhanced the carbon density data for various parts of the forest ecosystem in the Hainan Province through extensive field measurements, thus improving the accuracy of forest carbon storage estimates.\u003c/p\u003e \u003cp\u003eAlthough resource inventory data, remote sensing images, or a combination of both have been used in previous studies to evaluate forest carbon storage on Hainan Island, there is still a gap in understanding the primary factors that influence the spatial differentiation of carbon storage. Geographical detectors offer distinct advantages for the study of geographical phenomena. They can consider the complex relationships among multiple factors and reveal their mechanisms of impact on spatial distribution patterns. However, the use of geographical detectors requires the discretisation of continuous variables, as the driving factors must be continuous (Li, 2020). Continuous factor data are spatial data that describe attributes and spatial features. Attribute features contain either interval or ratio data (Peng, 2001). Discretisation is the process of converting interval or ratio data into nominal or ordinal discrete factor data (Hanning et al., 2003). Discretising continuous variables can be challenging owing to the subjective and random nature involved. Geographical detectors commonly use methods such as equal intervals, equal frequencies, geometric breakpoints, natural breakpoints, K-means, quantiles, and standard deviation distance for discretisation (Li, 2020). In the present study, we used the one-standard deviation method to classify and calculate the decision value q for each driving factor. This method is based on the numerical characteristics and distribution patterns of the original data, considering the mean value of the data as the classification centre, thereby yielding in more objective results (Meng et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn forest carbon storage research, ground survey data are preferred due to their high accuracy. However, adequately capturing the spatial heterogeneity of ground vegetation remains challenging because of the limited number of available survey points. Conversely, simulation methods that require less data and yield precise evaluation outcomes are adept at exploring alterations in regional carbon storage across diverse spatial and temporal scales (Wang et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Fu et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These approaches have garnered considerable attention (Zhang et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Yang et al., 2020; Fan et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The Integrated Valuation of Ecosystem Services and Trade-offs (InVEST) model is widely used in carbon storage assessments, primarily using land-use data (He et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Piyathilake et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Li et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003ePrevious studies have mainly focused on examining the impact of land use changes on carbon storage dynamics. For instance, Zhu et al. (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) used this model to examine the effects of land-use alterations on ecosystem carbon storage in the Qihai River Basin of the Taihang Mountains. Zou et al. (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) illustrated the capability of the InVEST model to simplify the modelling of yearly carbon fixation by forest trees and enhance carbon density approximations for forests. Zhu et al. (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) used this model to investigate the effects of land use changes on carbon storage during land expansion in Guangzhou.\u003c/p\u003e \u003cp\u003eIn this study, we used the InVEST model to examine the total volume and spatial dynamics of forest carbon storage on Hainan Island from 1990 to 2020. We conducted a geographical detector analysis to examine the driving factors and interaction patterns that contribute to the spatial differentiation of carbon storage in the forest ecosystems on Hainan Island. This facilitated identification of the primary factors influencing changes in forest carbon storage within the region. This study provides a theoretical basis for carbon management strategies in Hainan Forest ecosystems and aids in evaluating the role of forests in achieving carbon neutrality.\u003c/p\u003e\n\u003ch3\u003eStudy Area Overview\u003c/h3\u003e\n\u003cp\u003eHainan Island (18.80\u0026deg;\u0026ndash;20.10\u0026deg; N, 108.37\u0026deg;\u0026ndash;111.03\u0026deg; E) is located at the northern extremity of the South China Sea continental shelf, facing the Leizhou Peninsula across the water. It spans approximately 34,000 km\u003csup\u003e2\u003c/sup\u003e, encompasses 18 cities and counties, and has a coastline extending 1,944.4 km. The region exhibits diverse landforms, with towering mountains at its core and gradually declining elevations towards the periphery. The prominent mountainous features in the area are Wuzhi Mountain, with an elevation of 1867 m, and Yingge Ridge, which stands at 1811 m. The terrain gradually transitions from mountainous to hills, plateaus, and plains. The climate is characterised by high temperatures, abundant rainfall, and lengthy summers without winters. Annual precipitation ranges from 1000 to 2500 mm, whereas the average annual temperature fluctuates between 23.8 and 26.2\u0026deg;C, indicating a tropical monsoon maritime climate. The land use on Hainan Island is dominated by forests and farmland, which together account for approximately 90% of the total land area. According to Lei et al. (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), forest coverage has increased substantially from 38.2% in the early stages of provincial establishment to 62.1%.\u003c/p\u003e"},{"header":"Research Methodology","content":"\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eData collection\u003c/h2\u003e \u003cp\u003eThe land-use data for Hainan Island from 1990 to 2020 were obtained from the \u0026ldquo;30 m Annual Land Cover Data for China from 1990 to 2022\u0026rdquo; dataset by Yang et al. (2021). The dataset has a spatial resolution of 30 m \u0026times; 30 m. Normalised difference vegetation index (NDVI) data were obtained from the spatial distribution dataset of the annual vegetation index in China, curated by Xu Xinliang. The dataset covers the period 1998\u0026ndash;2020 and has a spatial resolution of 1 km.\u003c/p\u003e \u003cp\u003eAnnual average temperature and precipitation data for Hainan Island were obtained from the National Earth System Science Data Center. The data have a spatial resolution of 1 km and can be accessed through \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.geodata.cn/\u003c/span\u003e\u003cspan address=\"http://www.geodata.cn/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eThe elevation data used in this study were obtained from Professor Tang Guo\u0026rsquo;an\u0026rsquo;s China Digital Elevation Model (DEM) on a space-time three-pole environmental big data platform. DEM data can capture local terrain features at a certain resolution, facilitating the extraction of abundant surface morphology information, including the hillshade, slope, and aspect, at a spatial resolution of 1 km.\u003c/p\u003e \u003cp\u003eThe population density grids were obtained from WorldPop (\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) at a spatial resolution of 1 km. Nighttime light data were sourced from the \u0026ldquo;DMSP-OLS-like\u0026rdquo; nighttime light remote sensing dataset for China from 1992 to 2019, published online by Wu Yizhen et al. (2021) in the IEEE Transactions on Geoscience and Remote Sensing journal, with a spatial resolution of 1 km.\u003c/p\u003e \u003cp\u003ePer capita GDP, forestry policy information, and other relevant data were extracted from various editions of the \u0026ldquo;Hainan Statistical Yearbook\u0026rdquo;, \u0026ldquo;China Forestry Statistical Yearbook\u0026rdquo;, and \u0026ldquo;China Forestry and Grassland Statistical Yearbook\u0026rdquo;. These publications were compiled based on annual forestry statistical reports and other relevant materials submitted by the forestry management departments of provinces, autonomous regions, municipalities directly under the central government, and units directly under the National Forestry and Grassland Administration.\u003c/p\u003e \u003cp\u003eThe carbon density data used in the InVEST model for the carbon storage function module, with land-use types as evaluation units, were based on the latest research results on carbon density on Hainan Island by Gao et al. (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Carbon density values were assigned to raster objects based on land-use type, with forestland as the primary carbon density table, resulting in the calculation of carbon storage in the forest ecosystem of Hainan Island. The carbon densities are listed in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\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\u003eCarbon storage and density of different carbon pools in the forest ecosystems of Hainan Island.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" 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 \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIndex\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAboveground biomass\u003c/p\u003e \u003cp\u003ecarbon pool\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBelowground biomass\u003c/p\u003e \u003cp\u003ecarbon pool\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLitterfall carbon\u003c/p\u003e \u003cp\u003epool\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSoil carbon\u003c/p\u003e \u003cp\u003epool\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eForest ecosystem\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCarbon storage/(TgC)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e85.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e18.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e231.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e338.15\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCarbon density (MgC/hm\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e37.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e101.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e147.66\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePercentage/%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e25.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5.54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e68.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e100.00\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=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eCarbon storage calculation\u003c/h2\u003e \u003cp\u003eThe study used InVEST 3.12.1 (Sharp et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2014\u003c/span\u003e) and its carbon-storage module to calculate the carbon storage for each year over a 20-year period within the study area. Land use, aboveground biomass, belowground biomass, soil, and dead organic matter data were used for this purpose. The carbon storage pattern was mapped every five years, encompassing four main components of terrestrial ecosystem carbon storage: aboveground biomass carbon (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{above}\\)\u003c/span\u003e\u003c/span\u003e), belowground biomass carbon (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{below}\\)\u003c/span\u003e\u003c/span\u003e), soil carbon (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{soil}\\)\u003c/span\u003e\u003c/span\u003e), and dead organic carbon (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{dead}\\)\u003c/span\u003e\u003c/span\u003e). The formula used was as follows:\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$${C}_{i}={C}_{i,above}+{C}_{i,below}+{C}_{i,soil}+{C}_{i,dead}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$${C}_{total}=\\sum _{i=1}^{n}{C}_{i}\\times {S}_{i}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003erepresents the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th land-use type, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i}\\)\u003c/span\u003e\u003c/span\u003e denotes the total carbon density for the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th land-use type (t\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\cdot\\)\u003c/span\u003e\u003c/span\u003ehm\u003csup\u003e\u0026minus;2\u003c/sup\u003e), \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i,above}\\)\u003c/span\u003e\u003c/span\u003e represents the aboveground carbon density for the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th land-use type (t\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\cdot\\)\u003c/span\u003e\u003c/span\u003ehm\u003csup\u003e\u0026minus;2\u003c/sup\u003e), \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i,below}\\)\u003c/span\u003e\u003c/span\u003e represents the belowground carbon density for the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th land-use type (t\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\cdot\\)\u003c/span\u003e\u003c/span\u003ehm\u003csup\u003e\u0026minus;2\u003c/sup\u003e), \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i,dead}\\)\u003c/span\u003e\u003c/span\u003e represents the dead organic matter carbon density for the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th land-use type (t\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\cdot\\)\u003c/span\u003e\u003c/span\u003ehm\u003csup\u003e\u0026minus;2\u003c/sup\u003e), \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i,soil}\\)\u003c/span\u003e\u003c/span\u003e represents the soil carbon density for the\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th land-use type (t\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\cdot\\)\u003c/span\u003e\u003c/span\u003ehm\u003csup\u003e\u0026minus;2\u003c/sup\u003e), \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({S}_{i}\\)\u003c/span\u003e\u003c/span\u003edenotes the area for the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th land-use type (m\u003csup\u003e2\u003c/sup\u003e), \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i,total}\\)\u003c/span\u003e\u003c/span\u003e represents the total carbon storage for all land areas (t).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of carbon sequestration driving factors\u003c/h2\u003e \u003cp\u003eIn this study, we used the factor and interaction detection modules of geographic detectors to investigate the driving factors and interactions that shape the spatial distribution of carbon storage in the forest ecosystems of Hainan Island. Based on relevant literature, we identified nine key factors influencing spatiotemporal differentiation in carbon storage (Wang C et al., 2023; Wang Z, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). NDVI, elevation, slope, nighttime light index, GDP per capita, population average density, annual cumulative precipitation, annual average temperature, and cumulative sunshine hours were selected as indicators because they partially reflect the environmental conditions for vegetation growth and the impact of human activities on ecosystems.\u003c/p\u003e \u003cp\u003eFactor Detection: The main objective of this method is to identify spatial variations in Y and evaluate the extent to which factor x accounts for spatial variations in attribute Y. The factor detection results were measured using the q-value, which was calculated using the following formula:\u003cdiv id=\"Equ3\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ3\" name=\"EquationSource\"\u003e\n$$q=1-\\frac{\\sum _{i=1}^{1}{n}_{i}{\\delta }_{i}^{2}}{n{\\delta }^{2}}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e3\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(q\\)\u003c/span\u003e\u003c/span\u003e is the strength of the explanation of factor \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(x\\)\u003c/span\u003e\u003c/span\u003e for \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(Y\\)\u003c/span\u003e\u003c/span\u003e, and ranges from 0 to 1, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e Number of different carbon storage classes in the study area, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({n}_{i}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(n\\)\u003c/span\u003e\u003c/span\u003e are the sample sizes of the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th class in the study area and the total sample size of the study area, respectively. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\delta {\\text{i}}^{2}{\\text{i}}^{2}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({\\delta }^{2}\\)\u003c/span\u003e\u003c/span\u003e. The variance of carbon storage in the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th class of the study area and the total variance of carbon storage in the study area.\u003c/p\u003e \u003cp\u003eThe interaction detection module examines whether interactions exist among different driving factors and evaluates the strength of these interactions. It assesses whether the combined influence of factors X1 and X2 enhances or diminishes the explanatory power of the dependent variable Y, or whether the effects of these factors on Y are independent. It determines whether the combined correlation of different driving factors amplifies or diminishes their impact on the dependent variable Y. The interactions between the driving factors are listed in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\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\u003eTypes of interactions\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe criteria for judging\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eType\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eq(x1\u0026cap;x2)\u0026thinsp;\u0026lt;\u0026thinsp;Min[q(x1), q(x2)]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNonlinear attenuation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMin[q(x1),q(x2)]\u0026thinsp;\u0026lt;\u0026thinsp;q(x1\u0026cap;x2)\u0026thinsp;\u0026lt;\u0026thinsp;Max[q(x1),q(x2)]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNonlinear attenuation of a single factor\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eq(x1\u0026cap;x2)\u0026thinsp;\u0026gt;\u0026thinsp;Max[q(x1), q(x2)]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEnhancement by two factors\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eq(x1\u0026cap;x2)\u0026thinsp;=\u0026thinsp;q(x1)\u0026thinsp;+\u0026thinsp;q(x2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIndependent\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eq(x1\u0026cap;x2)\u0026thinsp;\u0026gt;\u0026thinsp;q(x1)\u0026thinsp;+\u0026thinsp;q(x2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNonlinear enhancement\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=\"Section2\"\u003e \u003ch2\u003eSpatial pattern analysis of carbon storage\u003c/h2\u003e \u003cp\u003eThis study used the Anselin Local Moran\u0026rsquo;s I index to describe the global spatial autocorrelation characteristics of carbon storage, which can be calculated based on the neighbouring areas of each region to obtain the local clustering situation of each region. Using this index, we further explored the spatial distribution patterns of carbon storage, identified local hotspots and cold spots, and identified clustering patterns and possible outliers in each region, gaining a deeper understanding of the spatial characteristics of carbon storage in the Hainan Island forest ecosystem. The formula used to calculate this index is as follows:\u003cdiv id=\"Equ4\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ4\" name=\"EquationSource\"\u003e\n$$\\text{Moran\u0026#039;s I }=\\frac{n\\sum _{i=1}^{n} \\sum _{j=1}^{n} {w}_{ij}\\left({x}_{i}-\\stackrel{-}{x}\\right)\\left({x}_{j}-\\stackrel{-}{x}\\right)}{\\sum _{i=1}^{n} {\\left({x}_{i}-\\stackrel{-}{x}\\right)}^{2}\\left(\\sum _{i} \\sum _{j} {w}_{ij}\\right)}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e4\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({x}_{i}\\)\u003c/span\u003e\u003c/span\u003e represents the carbon storage value of the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th region, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\stackrel{-}{x}\\)\u003c/span\u003e\u003c/span\u003e the global mean value, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({w}_{ij}\\)\u003c/span\u003e\u003c/span\u003e denotes the spatial weight between the regions \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e and\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(j\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\delta {\\text{i}}^{2}, {\\text{i}}^{2}\\)\u003c/span\u003e\u003c/span\u003e, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({\\delta }^{2}\\)\u003c/span\u003e\u003c/span\u003e represent the variance of carbon storage in the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e-th region and the overall variance of carbon storage in the study area, respectively.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results and Analysis","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eLand use change in Hainan Island\u003c/h2\u003e \u003cp\u003eUsing land cover data of 1990\u0026ndash;2020, we created spatial distribution maps and land-use transition matrices at 5-year intervals for Hainan Island (Fig.\u0026nbsp;1\u0026ndash;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The island displays substantial spatial heterogeneity in land-use types, with forestland being the most dominant, ranging from 63.9\u0026ndash;70.9% of all land-use types, followed by arable land, which accounts for 24.7\u0026ndash;31.9%. This study found that the built-up land area ranged from 0.6\u0026ndash;1.8%, whereas unused land accounted for only 0.002\u0026ndash;0.14%. The other land-use types remained relatively stable. Notable transitional changes were observed in each land type during the study period.\u003c/p\u003e \u003cp\u003eSubstantial transitions occurred in all land-use types, except for minor changes in water body areas, from 1990 to 2020. Between 1990 and 1995, the most remarkable change was the conversion of 1819.2 km\u003csup\u003e2\u003c/sup\u003e of arable land to forest land, which accounted for 62.15% of the total conversion area. Conversely, 831.1 km\u003csup\u003e2\u003c/sup\u003e of forestland was converted into arable land, accounting for 29.39% of the total conversion area. The built-up and shrub areas experienced relatively minimal changes during this period.\u003c/p\u003e \u003cp\u003eBetween 1995 and 2000, the largest area of forestland was converted to arable land, reaching 1634.0 km\u003csup\u003e2\u003c/sup\u003e, accounting for 62.17% of the total conversion area. Conversely, 753.2 km\u003csup\u003e2\u003c/sup\u003e of arable land was converted into forestland, accounting for 28.66% of the total conversion area. During this period, shrubs, grasslands, and unused land experienced substantial reductions, whereas conversion between forestland and arable land remained relatively high.\u003c/p\u003e \u003cp\u003eBetween 2000 and 2005, arable land expanded, and land types other than built-up land were converted into arable land. The largest conversion during this period was the transformation of forest land into arable land, which totalled 1845.1 km\u003csup\u003e2\u003c/sup\u003e, accounting for 66.05% of the total conversion area. In contrast, 692.9 km\u003csup\u003e2\u003c/sup\u003e of arable land was converted to forestland, comprising 24.80% of the total conversion area.\u003c/p\u003e \u003cp\u003eBetween 2005 and 2010, the forestland area increased substantially, with a growth of 2408.9 km\u003csup\u003e2\u003c/sup\u003e in 2010 compared to 2005, primarily attributed to conversion from arable land. Built-up land received inputs from all land types except shrublands.\u003c/p\u003e \u003cp\u003e From 2010 to 2015, arable land received inputs from forest land, with 2242.2 km\u003csup\u003e2\u003c/sup\u003e of forest land being converted to arable land, whereas the area of built-up land continued to grow. From 2015 to 2020, the forestland and arable land areas remained almost stable, indicating a relatively stable trend. The area of built-up land continued to increase.\u003c/p\u003e \u003cp\u003e \u003cb\u003eFigure\u0026nbsp;1\u003c/b\u003e Spatial Distribution of Land Use on Hainan Island from 1990 to 2020\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eSpatiotemporal characteristics of carbon storage in the forest ecosystem of Hainan Island\u003c/h2\u003e \u003cp\u003eUsing land use classification remote sensing images of Hainan Island from 1990 to 2020 and the InVEST model, we computed the total carbon storage of the forest ecosystem over a 30-year period. The analysis revealed a 'double-peak' curve in the total carbon storage of the forest ecosystem of Hainan Island from 1990 to 2020, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e. The peaks occurred in 1998 and 2011, with carbon storage reaching 369.94 and 372.80 Tg, respectively. In contrast, carbon storage dropped to 335.09 Tg in 2005, following a valley. Substantial fluctuations are observed in 1997, 2005, and 2013.\u003c/p\u003e \u003cp\u003eFurther analysis indicated that carbon storage in the forest ecosystem of Hainan Island showed an increasing trend (Fig.\u0026nbsp;4) from 1991 to 1997. A decreasing trend was observed from 1997 to 2004, followed by an increasing trend from 2004 to 2010. Following that, a decrease occurred from 2010 to 2015, then a stable period from 2015 to 2020 with a slight decline.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eFigure\u0026nbsp;4\u003c/b\u003e Carbon storage of the forest ecosystem in Hainan Island from 1990 to 2020\u003c/p\u003e \u003cp\u003eAt the global spatial scale, there were local variations in forest carbon storage from 1990 to 2020; however, the overall distribution did not change significantly. Carbon storage is mainly concentrated in the central and southern tropical rainforest areas of Hainan Island, including Qiongzhong, Baisha, and Wuzhishan. In contrast, low-value areas were dispersed in coastal regions, such as Lingao, Lingshui, and Haikou.\u003c/p\u003e \u003cp\u003eThe analysis revealed that central tropical rainforest areas, which are characterised by high forest quality, are the main hotspots of carbon storage in the forest ecosystem of Hainan Island. The integration of the local Moran\u0026rsquo;s I index and land-use spatial distribution provided valuable insights into the spatial distribution of carbon storage in the forest ecosystem of Hainan Island. In contrast, coastal areas primarily constitute cold spots, forming a circular distribution pattern across the island.\u003c/p\u003e \u003cp\u003eWhen observing the distribution of carbon storage hot and cold spots from 1990 to 2020, the degree of aggregation of high- and low-carbon storage was more pronounced, indicating a more concentrated aggregation of forests. High\u0026ndash;high clustering primarily occurred in forest and grassland areas, whereas high\u0026ndash;low clustering primarily occurred in urban areas. High\u0026ndash;low clustering can identify the boundary between urban construction and forests on Hainan Island. Low\u0026ndash;low clustering indicates areas with less forest carbon storage and more artificial construction.\u003c/p\u003e \u003cp\u003eFrom 1990 to 2020, patches of increase and decrease were dispersed in the southwestern part of Hainan Island (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e5\u003c/span\u003e). Areas of decreased carbon storage were more numerous than those of increased storage around the built-up coastal areas of Hainan Island. By combining different periods and the local Moran\u0026rsquo;s I index, we found that the spatial distribution of carbon storage in the forest ecosystem of Hainan Island changed. Between 1990 and 2000, carbon storage in the northwestern region of Hainan Island, particularly in Dongfang City, decreased significantly, transitioning from high\u0026ndash;high clustering to non-significance. Areas of low\u0026ndash;low clustering increased in the northern and eastern coastal areas of Hainan Island. Between 2000 and 2010, carbon storage in Ledong Li Autonomous County on Hainan Island increased significantly, transitioning from non-significant to high\u0026ndash;high clustering. From 1990 to 2020, the overall high\u0026ndash;high clustering of carbon storage showed a transition from scattered centres to an overall central aggregation. During the study period, the carbon storage in the northern region of Hainan Island decreased gradually. High\u0026ndash;low clustering of forest carbon storage was commonly found near low\u0026ndash;low clustering areas along the coast of Hainan Island. The region exhibited significant spatial heterogeneity in land-use types.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eDriving factors of carbon sequestration differentiation\u003c/h2\u003e \u003cp\u003eAn analysis of the spatial differentiation of the nine factors driving forest ecosystem spatial differentiation on Hainan Island was conducted at intervals of 10 years from 2000 onwards. The factors analysed were NDVI, elevation, slope, nighttime light index, per capita GDP at the city level, average population density at the city level, annual cumulative precipitation, annual average temperature, and cumulative sunshine. Before 2000, there was insufficient data on the relevant factors. Tables\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e shows the explanatory power index of the geographical detector for the detecting factors. The normalised vegetation index (NDVI) ranged from 0.521 to 0.530, elevation from 0.511 to 0.529, slope from 0.314 to 0.339, nighttime light index from 0.086 to 0.130, and per capita GDP from 0.089 to 0.164. The spatial differentiation of carbon storage was influenced by several factors, including average population density (0.080\u0026ndash;0.100), annual cumulative precipitation (0.066\u0026ndash;0.144), annual average temperature (0.040\u0026ndash;0.124), and cumulative sunshine (0.022\u0026ndash;0.130).\u003c/p\u003e \u003cp\u003eIn different years, NDVI, elevation, and slope, which are the three influencing factors, had explanatory powers exceeding 30%, indicating that they were the dominant factors affecting carbon storage. This analysis is objective and based solely on the data presented. These three factors exhibited fluctuations of less than 2.5% in their explanatory power for carbon storage in forest ecosystems over 20 years, indicating a relatively stable trend of their long-term influence on forest carbon storage. The correlation between NDVI and carbon storage suggests that areas with good vegetation cover usually have higher carbon storage in forests. Similarly, the slope significantly affects forest carbon storage. The latter terrain promotes better vegetation growth and carbon sequestration, whereas steep terrain may lead to soil erosion and organic matter loss, hindering carbon sequestration by plants.\u003c/p\u003e \u003cp\u003eThe nighttime light index, per capita GDP, population density, annual cumulative precipitation, annual average temperature, and cumulative sunshine showed medium-to-low explanatory power for carbon storage in forest ecosystems. The explanatory power of all factors was consistently below 15% across all periods. Per capita GDP and population density had an impact of approximately 10%, indicating that regional development quality and population concentration affected forest carbon storage. The explanatory power remained stable. The impact of climatic factors on the spatial and temporal differentiation of forest carbon storage on Hainan Island was evident, particularly in years of extreme climatic events. The explanatory power of the annual cumulative precipitation, annual average temperature, and cumulative sunshine varied significantly between 0.4% and 10.8% across all periods.\u003c/p\u003e \u003cp\u003eThere are dual-factor and nonlinear enhancement interactions between the driving factors, indicating that any combination of factors enhances the spatial differentiation of carbon storage in the forest ecosystem of Hainan Island. In different years, the combined action of elevation and NDVI had the greatest explanatory power for spatial differentiation of forest carbon storage in the region, ranging from 0.645 to 0.655. This text suggests that areas where elevation and NDVI interact may have rich forest carbon storage resources.\u003c/p\u003e \u003cp\u003eThe interaction between annual average temperature and population density mostly showed a nonlinear enhancement with other factors in different years. This indicates that the synergistic effects of annual average temperature and population density with other factors are notably superior to their individual effects. Therefore, to understand the factors that influence carbon storage in the forest ecosystem of Hainan Island, it is crucial to consider the combined effects of various driving factors.\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\u003eInteraction detection results of driving factors in 2000\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" 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 \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVegetation\u003c/p\u003e \u003cp\u003eCoverage Index\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eElevation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSlope\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNighttime\u003c/p\u003e \u003cp\u003eLight Index\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGDP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePopulation\u003c/p\u003e \u003cp\u003eDensity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eAnnual\u003c/p\u003e \u003cp\u003eCumulative\u003c/p\u003e \u003cp\u003ePrecipitation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAnnual\u003c/p\u003e \u003cp\u003eAverage\u003c/p\u003e \u003cp\u003eTemperature\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eCumulative\u003c/p\u003e \u003cp\u003eSunshine\u003c/p\u003e \u003cp\u003eDuration\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVegetation Coverage Index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.529\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eElevation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.648\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.511\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlope\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.577\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.534\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.339\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNighttime Light Index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.575\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.546\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.404\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.130\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGDP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.551\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.539\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.373\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.155\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.089\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePopulation Density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.548\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.528\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.371\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.195\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.157\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.080\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnnual Cumulative Precipitation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.557\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.566\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.415\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.178\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.139\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.151\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.070\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnnual Average Temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.539\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.547\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.385\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.187\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.144\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.123\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.082\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.040\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCumulative Sunshine Duration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.547\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.565\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.409\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.145\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.115\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.110\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.091\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.099\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.022\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\u003eThe blue area represents nonlinear enhancement, and the green area represents factor enhancement.\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\u003eInteraction detection results of driving factors in 2010\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" 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 \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVegetation\u003c/p\u003e \u003cp\u003eCoverage Index\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eElevation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSlope\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNighttime\u003c/p\u003e \u003cp\u003eLight Index\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGDP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePopulation\u003c/p\u003e \u003cp\u003eDensity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eAnnual\u003c/p\u003e \u003cp\u003eCumulative\u003c/p\u003e \u003cp\u003ePrecipitation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAnnual\u003c/p\u003e \u003cp\u003eAverage\u003c/p\u003e \u003cp\u003eTemperature\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eCumulative\u003c/p\u003e \u003cp\u003eSunshine\u003c/p\u003e \u003cp\u003eDuration\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVegetation Coverage Index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.530\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eElevation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.645\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.518\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlope\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.582\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.537\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.314\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNighttime Light Index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.556\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.549\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.384\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.086\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGDP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.578\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.537\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.367\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.184\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.114\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePopulation Density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.547\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.544\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.361\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.178\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.215\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnnual Cumulative Precipitation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.558\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.586\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.399\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.120\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.193\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.167\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.066\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnnual Average Temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.557\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.550\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.376\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.163\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.174\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.214\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.153\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.124\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCumulative Sunshine Duration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.555\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.582\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.387\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.202\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.168\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.168\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.116\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.156\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.065\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=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eInteraction detection results of driving factors in 2020\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" 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 \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVegetation\u003c/p\u003e \u003cp\u003eCoverage Index\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eElevation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSlope\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNighttime\u003c/p\u003e \u003cp\u003eLight Index\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGDP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePopulation\u003c/p\u003e \u003cp\u003eDensity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eAnnual\u003c/p\u003e \u003cp\u003eCumulative\u003c/p\u003e \u003cp\u003ePrecipitation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAnnual\u003c/p\u003e \u003cp\u003eAverage\u003c/p\u003e \u003cp\u003eTemperature\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eCumulative\u003c/p\u003e \u003cp\u003eSunshine\u003c/p\u003e \u003cp\u003eDuration\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVegetation Coverage Index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.521\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eElevation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.655\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.529\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlope\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.591\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.552\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.334\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNighttime Light Index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.554\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.564\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.418\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.101\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGDP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.580\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.547\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.402\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.254\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.164\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePopulation Density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.531\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.549\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.372\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.164\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.229\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.080\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnnual Cumulative Precipitation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.581\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.614\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.440\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.198\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.247\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.212\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.144\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnnual Average Temperature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.542\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.599\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.452\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.185\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.243\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.149\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.190\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.073\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCumulative Sunshine Duration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.585\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.614\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.436\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.183\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.211\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.203\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.161\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.167\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.130\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"},{"header":"Discussion","content":"\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eFeasibility of using the inVEST model for evaluating carbon storage in forests of Hainan Island\u003c/h2\u003e \u003cp\u003eThe InVEST model was used in this study to calculate the forest carbon storage on Hainan Island. The model relies on a carbon density table derived from field surveys of different vegetation types on the Hainan Island. The data used in the table include information from the 2021 forest resources survey of Hainan Province and observations from national forest ecosystem observation stations, such as Wenchang and Jianfengling. Compared to previous studies, such as Duan Xuan et al. (2022), who derived carbon density tables by correcting carbon density in the Pearl River Delta region based on average rainfall, temperature, and elevation data from Guangdong and Hainan, the results obtained in this study are more accurate. This is true for forest areas, where both aboveground and belowground biomass carbon densities are higher. In addition, the table includes carbon pools from litter, with a soil carbon density that is four times higher. The study by Liu et al. adheres to conventional academic structures and maintains regular author and institution formatting. (2022) used a comprehensive dataset to determine the ecological carbon density of terrestrial areas in China in the 2010s. They selected data points from the profiles of Hainan Island and calculated the corrected carbon density of the region. This study found that the total forest carbon density was like the data measured and compiled by Xuan et al., although there were discrepancies in the aboveground and soil carbon pools. Thus, the carbon density results obtained by Xuan et al. can be used to examine carbon storage in the forest ecosystem of Hainan Island.\u003c/p\u003e \u003cp\u003eObtaining clear satellite images within the same period is challenging because of the tropical and subtropical climate of Hainan Island. Furthermore, discrepancies in satellite image stitching, land cover type delineation, and calibration have been reported by different researchers, particularly at different resolutions. Large-scale remote sensing studies have been conducted on carbon sinks on Hainan Island using satellite data, such as Sentinel-2A from the European Space Agency (ESA), with a spatial resolution of 300 m and vegetation cover data from MODIS from 2000 to 2019. Han et al. (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) used this data. However, comprehensively validating the classifications obtained is challenging in previous studies analysing land cover data from satellite remote sensing, due to the limited number of training samples and computational capabilities. This study uses the \u0026lsquo;China 30 m Annual Land Cover Data from 1990 to 2022\u0026rsquo; by Yang Jie et al. (2021) from Wuhan University, which is based on 5,463 visual interpretation samples. Furthermore, it is worth noting that Liu et al. (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) found that the rate of change of forest area on Hainan Island from 1992 to 2019 was 1.3%. Incorporating this rate of change into the results obtained in this study can effectively account for the variation in forest carbon storage with changes in the forest area. According to the results of the third forest resource survey of Hainan Province conducted in 2021, the total forest area in the province was 2,283,242.74 ha, which closely aligns with the forest area of 2,212,257.33 ha in 2020 used in this study, indicating the reliability of the remote sensing classification data used in this study.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eImpact of policies on forest carbon storage in Hainan Island\u003c/h2\u003e \u003cp\u003eForest carbon storage policies have a crucial impact on the development of industries in Hainan Island. Over the past 30 years, the forest carbon storage on the island has undergone five stages of change.\u003c/p\u003e \u003cp\u003eDuring Stage 1 (1991\u0026ndash;1997), the forest ecosystem carbon storage increased from 350.92 to 368.95 Tg. Two significant forestry policies have played pivotal roles in China. In 1987, the \u0026lsquo;Directive on Strengthening Forest Resource Management in Southern Collective Forest Areas to Strictly Prohibit Illegal Logging\u0026rsquo; was implemented. This directive reduced deforestation and prevented a decrease in forestland area. The second most significant event was the nationwide cessation of natural forest logging in 1994. This has led to the implementation of various systems primarily based on joint production and contract responsibility, resulting in the fastest afforestation period in the history of Hainan. The land development and utilisation were strictly controlled.\u003c/p\u003e \u003cp\u003eDuring Stage 2 (1997\u0026ndash;2004), the carbon storage decreased from 369.95 to 338.73 Tg. The arable land area increased from 884,496.51 to 983,509.74 hm\u003csup\u003e2\u003c/sup\u003e, possibly due to increased demand and higher prices in the tropical fruit and vegetable markets, driving an increase in the planting area for these crops. However, this increase in arable land came at the cost of forest areas, which were encroached upon or converted to farmland, leading to a decrease in the forestland area.\u003c/p\u003e \u003cp\u003eDuring Stage 3 (2004\u0026ndash;2010), carbon storage increased from 338.73 to 372.80 Tg. The increase in forest area could be credited to policy implementation, like the \u0026ldquo;Opinions on Further Improving the Measures for Returning Farmland to Forests\u0026rdquo;. This policy restored arable land with forest attributes to forest land, increasing forest area. In addition, a reform pilot of the collective-forestland tenure system was initiated in Hainan Province in 2007, which played a role in protecting and managing forest resources, increasing forest carbon storage.\u003c/p\u003e \u003cp\u003eDuring Stage 4 (2010\u0026ndash;2015), these policies and reforms continued to have a positive impact on forest conservation and management. Carbon storage decreased from 372.80 to 350.16 Tg, possibly due to policies such as the \u0026lsquo;Opinions on Promoting the Construction and Development of Hainan International Tourism Island\u0026rsquo;, which resulted in urban expansion and increased construction land area.\u003c/p\u003e \u003cp\u003eDuring Stage 5 (2015\u0026ndash;2020), the carbon storage remained relatively stable, with a slight downward trend. In 2017, the Hainan Provincial Government implemented the \u0026lsquo;Regulations on Balanced Management of Land Occupation and Compensation for Forest Land\u0026rsquo; to ensure that the forest land area did not decrease. The Seventh Provincial Party Congress of the same year emphasised continued adherence to the policy of 'putting ecology first'. In 2018, the National Reserve Forest Construction Plan and Guiding Principles for National Ecological Civilisation Construction were implemented as policies.\u003c/p\u003e \u003cp\u003eAn analysis of almost 30 years of changes in forest ecosystem carbon storage on Hainan Island revealed that government forestry policies have played a crucial role in protecting and restoring the forestry ecological pattern in Hainan Province.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eKey factors influencing forest carbon storage on Hainan Island\u003c/h2\u003e \u003cp\u003eThis study found that forest carbon storage on Hainan Island is primarily influenced by the NDVI, elevation, and slope. Other factors, including nighttime light index, per capita GDP, population density, annual cumulative precipitation, mean annual temperature, and cumulative sunshine, also contribute to the spatial differentiation of forest carbon storage within the region. These findings support the results of Gao et al. (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), who used field experiments and remote sensing data to demonstrate that carbon storage in forest ecosystems on Hainan Island decreased with increasing elevation.\u003c/p\u003e \u003cp\u003eThe correlation between slope and forest carbon storage is consistent with the research conducted by Han et al. (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) using geographic detectors, which highlights the influence of the mean annual NDVI on the spatial differentiation of carbon storage. Furthermore, although Wang et al. (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and Pereira et al. (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) suggested that the soil type is a significant factor in the spatial distribution of forest carbon storage, future studies should explore and refine this relationship. In conclusion, it is crucial to understand the factors that influence forest ecosystem carbon storage on Hainan Island. Future studies should examine the interactive relationships between various factors, particularly the significant impacts of elevation and the vegetation coverage index on forest carbon storage. To gain a more comprehensive understanding of the spatial differentiation of forest ecosystem carbon storage on Hainan Island, it is necessary to further explore the synergistic effects of the different factors.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eProspects\u003c/h2\u003e \u003cp\u003eThe InVEST model was used to calculate forest ecosystem carbon storage data for Hainan Island over the past 30 years. The geographical detector method was subsequently used to quantitatively analyse the driving factors that influence the spatial distribution of carbon storage in the region. In contrast to previous large-scale studies on forest carbon storage on Hainan Island (Gao et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Liu Q et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Cao et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2002\u003c/span\u003e), which relied primarily on qualitative analysis, this study adopted a novel approach by adopting the geographical detector method. This method offers a new perspective for investigating ecosystem carbon storage on Hainan Island and provides valuable insights for future research on carbon sequestration issues in the region.\u003c/p\u003e \u003cp\u003eIt is important to note that the geographical detector method does not inherently address the directional effects of the factor importance. Future research could enhance the analysis by incorporating spatial statistical methods such as principal component analysis, analytic hierarchy process, system clustering analysis, and discriminant analysis. In further research on forest ecosystem carbon storage on Hainan Island, the use of geographic weighted regression (GWR) can provide a more comprehensive understanding of both direction and magnitude of the effects of driving factors on forest carbon storage. Guo et al. (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) used this method to estimate regional forest carbon storage. To date, no study has explored the relationship between survey factors and carbon storage using models of forest ecosystem carbon storage on Hainan Island. However, it is crucial to understand the relationship and spatial correlation characteristics between forest carbon storage and survey factors to establish models for estimating regional forest carbon storage and its distribution.\u003c/p\u003e \u003cp\u003eA factor estimation model for forest ecosystem carbon storage on Hainan Island can be developed using satellite bands, topographic, climate, and human activity data. This approach leads to more accurate results in the model estimation and prediction of forest ecosystem carbon storage based on forest inventories (Gao et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Cao et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Lin et al., 2000) or remote sensing image data (Liu et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This will help achieve high-quality monitoring records year after year, advancing the early realisation of the 'double carbon' goal.\u003c/p\u003e \u003cp\u003eBased on the information above, we now comprehensively discuss how to help Hainan Province achieve carbon neutrality through forest carbon sink in the future.\u003c/p\u003e \u003cp\u003eAs we move forward with carbon-neutral initiatives, we should focus on forest carbon sequestration and turn its potential and cost advantages into real benefits for carbon reduction. Focusing on Wuzhishan City, Qiongzhong City, Baisha City, Baoting City, and other places with high forest carbon storage, a forest carbon sequestration pilot base will be established. Regarding policy formulation, afforestation, quality improvement, quota cutting, mechanism innovation, etc., resources should be channelled towards increasing forest carbon sink capacity and improving carbon sequestration. In terms of specifically improving carbon sink capacity, there should a focus on supporting state-owned forest farms, collective economic cooperation organizations, enterprises and large forest farmers and other subjects. The application and demonstration of forestry carbon sequestration methodology and new technologies will be the main direction of the pilot base, and the first exploration of forestry carbon sequestration trading and carbon sequestration capacity potential improvement will be carried out at the same time.\u003c/p\u003e \u003cp\u003eIn addition, there must be a focus on coordinated promotion and pay attention to spillover effects. The terrain of Hainan Province is complex (with many high and low points), and there are great differences in forest resources, economic development level, energy consumption structure and carbon emission level among different regions. Therefore, the pace at which peak carbon emissions and carbon neutrality is achieved across different regions should be adapted to local conditions, rather than generalized. A global market for carbon emissions trading should be introduced, with plans to include forestry carbon sinks in the carbon market. Carbon sink trading between regions with poor economy but rich forest carbon reserves and regions with developed economy but relatively poor carbon reserves will become a reality. An accurate assessment of the ecological and economic value of forests in the region should contributes towards their sustainability.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eResearch limitations\u003c/h2\u003e \u003cp\u003eThis study utilized 30 years of consecutive land use remote sensing image data to classify land use types within urban areas were uniformly classified as forest land, without further subdivision of the classification units, which may have introduced certain errors into the experimental results. Given that Hainan Island has abundant original or contiguous forest resources, according to the Third National Land Survey of Hainan Province, urban land and town land encompassed 35,555.06 and 60,425.69 hectares in 2019, respectively. However, according to the forest resources inventory data of Hainan Province in 2018, the forest area was 1,944,900 hectares in 2018; the total land area only accounted for 4.93% of the forest area. Meanwhile, the remote sensing data used in this study showed that the area of construction land in Hainan Province in 2020 accounted for 2.05%, while forest land accounted for 64.55%. Therefore, treating forest land as a homogenous monolith, may impact experimental results, even though the impact would most likely be negligible. In future forest division studies, it will be necessary to focus on using a large amount of measured data or sufficient on-site monitoring data as the classification learning objects to improve the accuracy of regional ecosystem carbon storage assessment.\u003c/p\u003e \u003cp\u003eIn this study, the geographic detector was used to discretize the data, but only one discretization method was adopted. Specifically, results obtained via multiple discretization methods were not considered. However, the comprehensive analysis of multiple discretization methods may help to understand the relationship between data features and factors more comprehensively, to obtain more accurate results. In addition, although the geographic detector can show the important explanatory force of the factor, it cannot explain the direction of the explanatory force of the factor. Therefore, to comprehensively elucidate the role of various factors, the use of correlation factor analysis will be necessary. However, the value of correlation analysis results will lose some spatial explanatory power, so correlation analysis is mainly used to judge the direction of the influence factors. Overall, the shortcomings of the research are mainly in the selection of data processing and analysis methods, and the comprehensive utilization of various methods is not fully considered, which may lead to the limitation of the results. In the future, when using geo-detectors, it could be beneficial to combine more comprehensive auxiliary data analysis methods. For example, the use of geographic detectors and GWR or principal component analysis can provide a more comprehensive understanding of the relationship and degree that affect the spatial differentiation of forest carbon stocks, to ensure that the obtained data results contain both spatial interpretation and more accurate data interpretation. In this study, the geographic detector was used to discretize the data, but only one discretization method was adopted, and the comprehensive analysis of the results obtained by multiple discretization methods was not considered. The comprehensive analysis of multiple discretization methods may help to understand the relationship between data features and factors more comprehensively, to obtain more accurate results. In addition, although the geographic detector can show the important explanatory force of the factor, it cannot explain the direction of the explanatory force of the factor. Therefore, yield a more accurate understanding of factors, correlation factor analysis is also needed. However, the value of correlation analysis results will lose some spatial explanatory power, so correlation analysis would be mainly used to judge the direction of the influence factors. Overall, the shortcomings of the research are mainly in the selection of data processing and analysis methods, and the comprehensive utilization of various methods is not fully considered, which limit the rigour of the results. In the future, when using geo-detectors, it will be worth combining more comprehensive auxiliary data analysis methods such as: The use of geographic detectors and GWR or principal component analysis can provide a more comprehensive understanding of the relationship and degree that affect the spatial differentiation of forest carbon stocks. This approach would yield results that are not only accurate but that also have a spatial component.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eInnovation points\u003c/h2\u003e \u003cp\u003eIn the past, ess attention was paid to the year-by-year changes of forest carbon stocks in Hainan Island. Here, we highlight the importance of studying the pattern of forest carbon stocks in Hainan Island in a long time series. The use of InVEST model to analyse forest carbon stocks provide a systematic and quantitative method for research. By analysing data over a long period of time, the dynamic change trend of forest carbon stocks can be more accurately understood, and era-specific changes in the area of forests can be easily explored. From the perspective of forestry policy, the causes and impact time of forest area change are viewed to provide scientific basis for the formulation of long-term forest management strategies.\u003c/p\u003e \u003cp\u003eFor the first time in Hainan Island, the correlation analysis of forest carbon storage factors was conducted using a geographic detector. The advantage of this method lies in the fact that it facilitates an exploration of nine spatial factors, including a measure of their spatial differentiation. In addition, this approach allows for an exploration of multiple independent variables and the degree of impact on forest carbon storage. Through the lens of space, we were able to explain whether the spatial distribution of important factors affecting forest carbon stocks is similar to the amount of forest carbon stocks. Simultaneously, by using the interpretation data of the third phase, we determined whether the interpretation strength of factors for forest carbon stocks has been improved, revealing the mechanism underlying the change of forest carbon stocks. Overall, this approach will allow for a better understanding of the carbon cycle process of forest ecosystem, providing a theoretical basis for forest management and protection in other areas in the future. In addition, this study has brought forth new research avenues in the context of analysing factors that disturb forest carbon storage in the context of large space and long-time series.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eBetween 1990 and 2020, there were substantial changes in land-use types on Hainan Island. The proportions, from largest to smallest, were forest land, cropland, water bodies, construction land, grassland, unused land, and shrubbery. The conversion between cropland and forestland had the highest proportion, and changes in forestland area were the primary direct cause of forest carbon storage variation.\u003c/p\u003e \u003cp\u003eBetween 1990 and 2020, the total carbon storage of the forest ecosystem on Hainan Island (excluding Sansha City) fluctuated between 335.09 and 372.80 teragrams of carbon (TgC). The lowest carbon storage was recorded in 2004 at 335.09 TgC, whereas the highest was in 2010 at 372.80 TgC. Between 1991 and 1997, forest ecosystem carbon storage showed an upward trend followed by a downward trend from 1997 to 2004. From 2004 to 2010, there was an upward trend followed by a downward trend from 2010 to 2015. Forest ecosystem carbon storage was stable from 2015 to 2020.\u003c/p\u003e \u003cp\u003eForest ecosystem carbon storage on Hainan Island exhibited spatial clustering (\u003cem\u003ep\u0026thinsp;\u0026lt;\u003c/em\u003e\u0026thinsp;0.01), with higher values concentrated in central mountainous regions and lower values in coastal areas. The carbon storage levels on Hainan Island varied significantly across different regions. The eastern part of the island has higher carbon storage levels than the western part, whereas the southern regions exhibit higher values than the northern regions. The forest ecosystem carbon storage also showed significant variability with increasing elevation.\u003c/p\u003e \u003cp\u003eGeographical detector analysis revealed that topography, climate, and anthropogenic factors were the main factors influencing the spatial differentiation of forest ecosystem carbon storage on Hainan Island. The distribution of forest carbon storage was significantly shaped by the NDVI (0.521\u0026ndash;0.530), elevation (0.511\u0026ndash;0.529), and slope (0.314\u0026ndash;0.339), which emerged as the primary drivers. These factors had a greater impact than others. In addition, the interaction detection results highlighted the synergistic effects of various driving factors, surpassing their individual impacts on forest ecosystem carbon storage. The synergy between the NDVI and elevation exhibited the strongest nonlinear enhancement (0.645\u0026ndash;0.655), which is noteworthy.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAcknowledgments:\u0026nbsp;We would like to thank Editage (www.editage.cn) for English language editing.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Funding: This research is supported by the Special Fund for Basic Research Expenses of Central public welfare research institutes; the grant number is (CAFYBB2021ZE001). Xu Wang received financial support from the project\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Conflicts of interest/Competing interests: There are no conflicts of interest.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Availability of data and material: Not Applicable\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Code availability: Not Applicable\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Authors\u0026apos; contributions: Pi. and Wang and Zang. wrote the main manuscript text; Song. and Wang. and Zhou. drafted the work; Pi. and Guo. and BaoYin. and Zhao. acquisition, analysis of data; Li. and Qiu. and Wu. interpretation of data; All authors reviewed the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eCao J, Zhang Y, Liu Y (2002) Changes in forest biomass carbon storage in Hainan Island over the last 20 years. Geogr Res 21(5):551\u0026ndash;560.\u003c/li\u003e\n\u003cli\u003eChai Y, Yu Y (2022) Research on innovative paths of Hainan tropical rainforest national park system. West For Sci 51(1):155\u0026ndash;160. (In Chinese)\u003c/li\u003e\n\u003cli\u003eDuan X, Gong W, Sun Y, Liu T, Qiu X, Zhang Y (2022) Land use change in the coastal zone of Hainan Island and its spatiotemporal evolutionary impact on carbon storage. Bull Soil Water Conserv 42(5):301\u0026ndash;311. (In Chinese)\u003c/li\u003e\n\u003cli\u003eFan L, Cai T, Wen Q, Han J, Wang S, Wang J, Yin C (2023) Scenario simulation of land use change and carbon storage response in Henan Province. Ecol Indic 154(7):110660.\u003c/li\u003e\n\u003cli\u003eFu Y, Huang M, Gong D, Lin H, Fan Y, Du W (2023) Dynamic simulation and prediction of carbon storage based on land use/land cover change from 2000 to 2040: A case study of the Nanchang urban agglomeration. Remote Sens 15(19):4645. https://doi.org/10.3390/rs15194645\u003c/li\u003e\n\u003cli\u003eGao S, Chen Y, Chen Z, Lei J, Wu T (2023) Carbon storage and spatial distribution characteristics of forest ecosystems on Hainan Island. Acta Ecol Sin 43(9):3558\u0026ndash;3570. (In Chinese)\u003c/li\u003e\n\u003cli\u003eGuan M, Xiong C (2022) The net spatio-temporal impact of the international tourism is-land strategy on the ecosystem service value of Hainan Island: A counterfactual analysis. Land 11(10):1694. https://doi.org/10.3390/land11101694\u003c/li\u003e\n\u003cli\u003eGuo H, Zhang M, Xu L, et al (2015) Estimation of regional forest carbon storage based on geographically weighted regression. J Zhejiang A\u0026amp;F University 32(4):497\u0026ndash;508.\u003c/li\u003e\n\u003cli\u003eHan J, Zhang G, Li W, et al (2022) Analysis of vegetation ecological quality change characteristics in Hainan Island over the past 20 years. Sci Ecol Sin 41(01):20\u0026ndash;30. https://doi.org/10.14108/j.cnki.1008-8873.2022.01.003\u003c/li\u003e\n\u003cli\u003eHan Y, Ding S, Yang T (2023) Spatial and temporal distribution of carbon storage in the ecosystem of Southern Taihang Mountains in Southern Shanxi and its driving factors. Chin J Environ Sci 43(3):1298\u0026ndash;1306.\u003c/li\u003e\n\u003cli\u003eHanning R (2003) Spatial data analysis: Theory and practice. Cambridge University Press\u003c/li\u003e\n\u003cli\u003eHe C, Zhang D, Huang Q, Zhao Y (2016) Assessing the potential impacts of urban expansion on regional carbon storage by linking the LUSD-urban and InVEST models. Environ Model Softw 75:44\u0026ndash;58. https://doi.org/10.1016/j.envsoft.2015.09.015\u003c/li\u003e\n\u003cli\u003eHou K, Li X, Wang JJ, Zhang J (2016). An analysis of the impact on land use and ecological vulnerability of the policy of returning farmland to forest in Yan\u0026rsquo;an, China. Environ Sci Pollut Res Int 23(5):4670\u0026ndash;4680.\u003c/li\u003e\n\u003cli\u003eKang Hou et al (2016) An analysis of the impact on land use and ecological vulnerability of the policy of returning farmland to forest in Yan\u0026rsquo;an, China. Environmental Science and Pollution Research 23: 4670\u0026ndash;4680.\u003c/li\u003e\n\u003cli\u003eKothandaraman S, Dar JA, Sundarapandian S, Dayanandan S, Khan ML (2020) Ecosystem-level carbon storage and its links to diversity, structural and environmental drivers in tropical forests of Western Ghats, India. Sci Rep 10(1):13444. https://doi.org/10.1038/s41598-020-70313-6\u003c/li\u003e\n\u003cli\u003eLei JR, Chen ZZ, Chen XH, et al. (2020) Spatiotemporal changes in land use and ecosystem service value on Hainan Island from 1980 to 2018. Acta Ecol Sin 40(14):4760\u0026ndash;4773.\u003c/li\u003e\n\u003cli\u003eLi P, Chen J, Li Y, Wu W (2023) Using the InVEST-PLUS model to predict and analyze the pattern of ecosystem carbon storage in Liaoning Province, China. Remote Sens 15(16):4050.\u003c/li\u003e\n\u003cli\u003eLi W, Chen J, Zhang ZJFE (2020). Forest quality-based assessment of the Returning Farmland to Forest Program at the community level in SW China Management 461:117938.\u003c/li\u003e\n\u003cli\u003eWenqing L, Chen J, Zhang Z (2020) Forest quality-based assessment of the Returning Farmland to Forest Program at the community level in SW China. Forest Ecology and Management 461: 117938.\u003c/li\u003e\n\u003cli\u003eLin MZ, Zhang YL Dynamic changes and sustainable development of forest resources in Hainan Island. Ecol Sci 2000(4):84\u0026ndash;89.\u003c/li\u003e\n\u003cli\u003eLiu Q, Yang D, Cao L, Anderson B (2022) Assessment and prediction of carbon storage based on land use/land cover dynamics in the tropics: A case study of Hainan Island, China. Land 11(2):244. https://doi.org/10.3390/land11020244\u003c/li\u003e\n\u003cli\u003eMeng Q, Wu ZT, Du ZQ, Zhang H (2021) Quantitative impact of regional vegetation cover based on geographic detector: A case study of the Beijing-Tianjin Sand source area. Chin J Environ Sci 41(2):826\u0026ndash;836.\u003c/li\u003e\n\u003cli\u003ePan YD, Birdsey RA, Fang JY, et al. (2011) A large and persistent carbon sink in the world\u0026rsquo;s forests. Science 333(6045):988\u0026ndash;993. https://doi.org/10.1126/science.1201609\u003c/li\u003e\n\u003cli\u003ePereira OJR, Montes CR, Lucas Y, Santin RC, Melfi AJ (2015) A multi-sensor approach for mapping plant-derived carbon storage in Amazonian podzols. Int J Remote Sens 36(8):2076\u0026ndash;2092. https://doi.org/10.1080/01431161.2015.1034896\u003c/li\u003e\n\u003cli\u003ePhillips OL, Malhi Y, Higuchi N, et al. (1998) Changes in the carbon balance of tropical forests: Evidence from long-term plots. Science 282(5388):439\u0026ndash;442. https://doi.org/10.1126/science.282.5388.439\u003c/li\u003e\n\u003cli\u003ePiyathilake I, Udayakumara E, Ranaweera L, Gunatilake SJMES (2022) Modeling predictive assessment of carbon storage using InVEST model in Uva province, Sri Lanka. Environment 8:2213\u0026ndash;2223.\u003c/li\u003e\n\u003cli\u003eRen H, Li L, Liu Q et al. (2014) Spatial and temporal patterns of carbon storage in forest ecosystems on Hainan Island, Southern China. PLoS One 9(9):e108163. https://doi.org/10.1371/journal.pone.0108163\u003c/li\u003e\n\u003cli\u003eTallis H, Icketts T, Guerry A (2013) In: VEST User\u0026rsquo;s Guide: Integrated valuation of environmental services and tradeoffs. The Natural Capital Project, Stanford.\u003c/li\u003e\n\u003cli\u003eSharp, R., et al (2014) InVEST user\u0026rsquo;s guide: integrated valuation of environmental services and tradeoffs. The Natural Capital Project. In Stanford Woods Institute for the Environment. University of Minnesota\u0026apos;s Institute on the Environment, the Nature Conservancy \u0026amp; WW Foundation Stanford.\u003c/li\u003e\n\u003cli\u003eTang G (2019) Digital elevation model of China (1KM). A big earth data platform for three poles.\u003c/li\u003e\n\u003cli\u003eWang CW, Luo JJ, Tang HH (2023) Analysis of spatiotemporal differentiation driving forces of ecosystem carbon storage in the Taihang Mountains area based on the InVEST model. Ecol Environ Sci 32(2):215\u0026ndash;225.\u003c/li\u003e\n\u003cli\u003eWang JF, Xu CD (2017) Geographic detector: Principles and prospects. Acta Geogr Sin 72(1):116\u0026ndash;134.\u003c/li\u003e\n\u003cli\u003eWang JJ Study on the dynamic characteristics and influencing factors of carbon storage of main tree species in forest arboreal layer in Anze County, Shanxi Province. Master\u0026rsquo;s thesis. Shanxi University of Finance \u0026amp; Economics.\u003c/li\u003e\n\u003cli\u003eWang Junjie (2022) Study on dynamic change characteristics and influencing factors of carbon storage of main tree species in forest tree layer in Anze County, Shanxi Province. Dissertation, Shanxi University of Finance and Economics\u003c/li\u003e\n\u003cli\u003eWang P (2010) Analysis of the development of tropical fruit industry in Hainan Province. Trop Agric Eng 34(2):67\u0026ndash;70.\u003c/li\u003e\n\u003cli\u003eWang Z Study on the spatiotemporal changes and influencing factors of forest carbon storage in Hangzhou city based on the CASA model. Master\u0026rsquo;s thesis. Zhejiang A\u0026amp;F University.\u003c/li\u003e\n\u003cli\u003eWang Z (2021) Study on spatial-temporal changes and influencing factors of forest carbon storage in Hangzhou based on CASA model. Dissertation, Zhejiang A \u0026amp; F University\u003c/li\u003e\n\u003cli\u003eWang ZH, Liu HM, Guan QW, et al (2011) Carbon storage and density of urban forest ecosystems in Nanjing. J Nanjing For Univ 54(4):18. \u003c/li\u003e\n\u003cli\u003eWang ZH, Liu HM, Guan QW, et al (2011) Carbon storage and density of urban forest ecosystems in Nanjing. Journal of Nanjing Forestry University 54(4):18.\u003c/li\u003e\n\u003cli\u003eWei S, Yang Y, Lin Z, Zhang D (2014). A model of the new style of urbanization for Hainan province in the context of international tourism-island construction. Resources 35:14\u0026ndash;18.\u003c/li\u003e\n\u003cli\u003eWei, S, Yang, Y, Lin, Z, Zhang, D. (2014) A model of the new style of urbanization for Hainan province in the context of international tourism-island construction. Shanghai Land and Resources, 35(1), 14\u0026ndash;18.\u003c/li\u003e\n\u003cli\u003eWu Y, Shi K, Chen Z, Liu S, Chang Z (2021) Developing improved time-series of improved DMSP-OLS-like data (1992\u0026ndash;2019) in China by integrating DMSP-OLS and SNPP-VIIRS. IEEE Trans Geosci Remote Sensing 60:1\u0026ndash;14. https://doi.org/10.1109/TGRS.2021.3135333\u003c/li\u003e\n\u003cli\u003eXu XL China annual vegetation index (NDVI) spatial distribution dataset. Data registration and publishing system of resource and environmental science data center. Chinese Academy of Sciences, 2018 https://doi.org/10.12078/2018060601\u003c/li\u003e\n\u003cli\u003eYan LX Research on strategies to enhance the competitiveness of tropical fruit industry in Hainan Province. Master\u0026rsquo;s thesis. South China University of Tropical Agriculture.\u003c/li\u003e\n\u003cli\u003eYan Linxia (2005) Research on countermeasures to enhance the competitiveness of tropical fruits in Hainan Province. Dissertation, South China Tropical Agricultural University\u003c/li\u003e\n\u003cli\u003eYang J, Huang X (2021) The 30 m annual land cover dataset and its dynamics in China from 1990 to 2019. Earth Syst Sci Data 13(8):3907\u0026ndash;3925. https://doi.org/10.5194/essd-13-3907-2021\u003c/li\u003e\n\u003cli\u003eZhang L, Zhou G, Ji Y, Bai Y (2016) Spatiotemporal dynamic simulation of grassland carbon storage in China. Sci China Earth Sci 59(10):1946\u0026ndash;1958. https://doi.org/10.1007/s11430-015-5599-4\u003c/li\u003e\n\u003cli\u003eZhang Y, Zhang W, Ding M (2004) Differences in estimating carbon stocks based on land use/cover classification system: A case study of forests in Hainan Island. Prog Geogr 06:63\u0026ndash;70.\u003c/li\u003e\n\u003cli\u003eZhu W, Zhang J, Cui Y, et al (2019) Assessment of ecosystem carbon stock based on scenarios of land use change: A case study of Qihe River Basin in Taihang Mountains. Acta Geogr Sin 74(3):446\u0026ndash;459.\u003c/li\u003e\n\u003cli\u003eZhu Z, Ma X, Hu H (2021) Spatiotemporal evolution and prediction of ecosystem carbon stock in Guangzhou based on coupled FLUS-InVEST model. Bull Soil Water Conserv 41(2):222\u0026ndash;229 + 239\u003c/li\u003e\n\u003cli\u003eZou W, He Y, Ye B, et al (2021) Study on ecosystem carbon stock of Qianjiangyuan National Park based on InVEST model. J Cent S Univ For Technol 41(3):120\u0026ndash;128.\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":"Carbon storage, forest ecosystem, InVEST model, geographic detector, Hainan Island, Influencing factors","lastPublishedDoi":"10.21203/rs.3.rs-4105908/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4105908/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study examines long-term carbon sequestration in the forest ecosystems of Hainan Island from 1990 to 2020 using the InVEST model and a geographic detector technique. We analysed changes in land use and forest cover, observing an 85.78%, 87.55%, and 256.96% decrease in undeveloped, shrub-covered, and burned urbanised land, respectively. Urbanised land increased by 4.01% annually. Forested land decreased by 3.62%, agricultural land expanded by 5.27%, and aquatic bodies decreased by 2.05%. The forest ecosystems sequestered 335.09\u0026ndash;372.80 TgC of carbon, showing an upward trend from 1991 to 1997, a decline from 1997 to 2004, an increase from 2004 to 2010, a decrease from 2010 to 2015, and overall stability from 2015 to 2020. Spatial clustering analysis revealed substantial clustering of carbon sequestration, with central mountainous regions exhibiting elevated levels, coastal areas having diminished levels, the east experiencing higher levels than the west, and the south showing escalated levels compared to the north. Geographical detector analysis identified NDVI, elevation, and slope as primary drivers of spatial variance in carbon sequestration. Forested area changes and government forestry policies played a pivotal role in enhancing carbon sequestration. The combined effect of NDVI and elevation normalisation on vegetation coverage had the most potent synergistic impact.\u003c/p\u003e","manuscriptTitle":"Spatiotemporal variation in carbon sequestration in the forest ecosystem of Hainan Island over a 30-year period and its driving factors","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-03-22 06:43:49","doi":"10.21203/rs.3.rs-4105908/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":"ad08dab6-14ae-4492-84ae-5f7fc912e1c8","owner":[],"postedDate":"March 22nd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-03-22T16:28:36+00:00","versionOfRecord":[],"versionCreatedAt":"2024-03-22 06:43:49","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4105908","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4105908","identity":"rs-4105908","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}