MODIS-NDVI (2001-2020) response to natural factors and engineering activities along railway in Sichuan-Tibet Region

doi:10.21203/rs.3.rs-3624264/v1

Download PDF

Article

MODIS-NDVI (2001-2020) response to natural factors and engineering activities along railway in Sichuan-Tibet Region

https://doi.org/10.21203/rs.3.rs-3624264/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The ecological environment of Sichuan-Tibet region is sensitive and fragile, the impact of railway construction on its ecological environment cannot be ignored. This research used the NDVI, a theoretical model of spatial-temporal pattern evolution of vegetation cover along the railway and its driving factors was innovatively proposed by integrating Geodetector and other methods. And the model was applied to an empirical study of the vegetation coverage along the major railway projects in Sichuan-Tibet region from 2001 to 2020, based on the MODIS-NDVI dataset. The results showed that: (1) There is significant spatial heterogeneity in the response of vegetation dynamics to natural and anthropogenic factors along the railway. (2) There are two-factor and nonlinear enhancement interaction of the driving factors on vegetation cover, and the interaction effect is greater than the independent effect. (3) Human activities can not only negatively affect vegetation growth and lead to vegetation degradation, but also positively promote vegetation restoration. The general theoretical model proposed in this study not only provides data support for ecological environment monitoring and restoration in Sichuan-Tibet area and areas with similar geographical characteristics, but also provides theoretical support for the study of vegetation dynamic change along major international railway projects.

Biological sciences/Ecology

Biological sciences/Plant sciences

Earth and environmental sciences/Environmental sciences

Earth and environmental sciences/Environmental social sciences/Environmental impact

Vegetation dynamic

Normalized Difference Vegetation Index

Geodetector

Driving factors

Railway project

Spatial and temporal evolution

The construction of railway increased the degree of human disturbance to the ecosystem ^{[1, 2]}. Under the background of the increasing expansion of railway construction, the impact of engineering activities along the railway on the ecological environment cannot be ignored. Major railway projects in Sichuan-Tibet region are located in the southeast of Qinghai-Tibet Plateau (QTP) ^[3]. As a "sensor" and "sensitive region" of ecological change in Asia and even the Northern Hemisphere, the region where the railway is located plays an important role in biodiversity maintenance, water carbon supplement and ecosystem service regulation ^{[4, 5]}. However, the ecosystem along major railway projects is sensitive and fragile ^[6], affected by the unique climatic conditions and geographical location of QTP, and if damaged, it is very difficult to repair and may even cause serious environmental problems ^[7]. Therefore, timely, dynamic, objective and accurate monitoring of the changes of ecological environment along the major railway projects in Sichuan-Tibet region is very important for the local ecological restoration, even the global ecological balance.

Vegetation is the most important part of the ecosystem along the railway, which plays an important role in maintaining biodiversity and global carbon balance ^{[8, 9]}. Vegetation dynamics affect environmental evolution ^[10]. Therefore, the quality of local ecological environment is mainly determined by vegetation^[11]. In recent years, many scholars have devoted themselves to the study of eco-environmental impact along the railway in Sichuan-Tibet region. For example, Ding et al. found that engineering activities in the process of railway construction (subgrade filling, soil removal and disposal site and construction walkway) were the main factors affecting vegetation cover^[12]. Liu et al., Zhu et al. and Monson et al. believed that among the factors affecting each district and vegetation ecosystem along the railway, ecological factors such as climate were the main controlling factors ^[13–15]. However, previous studies on driving factors of vegetation change along major railway projects in Sichuan-Tibet region mainly focus on one aspect (natural factors or human factors), and there are problems such as incomplete selection of driving factors of vegetation dynamics and unclear details of vegetation growth change due to short research time series. In addition, it is still unclear whether and to what extent vegetation dynamics can be attributed to changes in temperature, precipitation, slope and other ecological factors ^[16–18], and how much it is affected by the intensity of engineering activities ^[18], which still cannot meet the practical guidance requirements for sustainable development of ecological environment in Sichuan-Tibet region. Therefore, the problem of this study is how to comprehensively consider ecological factors and engineering activities to more accurately determine the driving factors of vegetation change along the major railway projects in Sichuan-Tibet region.

Traditional methods based on linearity (e.g., linear regression and residual trend analysis) have been widely used to reveal the relationships between the monotonic trend of vegetation change and its driving forces in previous studies^[19]. However, the coupling relationship between natural and anthropogenic factors and vegetation change, or even between these factors, is not simply linear due to the complex responses of vegetation growth to drivers ^{[20, 21]}. Therefore, rigid models based on linear assumptions cannot accurately describe this intrinsic connection. To better understand this complex relationship, the researchers proposed a nonlinear model called the Geodetector. Geodetectors based on the spatial hierarchical heterogeneity theory have no linearity assumption and have a clear physical meaning to measure spatial differentiation, detect drivers, and analyze the interactions between variables ^[22]. It has been widely used to quantify the nonlinear effects of factors and their interactions on vegetation change. In addition, remote sensing data with high spatial coverage and long time series provide an effective way to monitor large-area vegetation dynamics due to the rapid development of GIS technology^{[23, 24]}. Normalized difference vegetation index (NDVI) is the most effective indicator for monitoring regional or global vegetation and ecological environment changes in remote sensing methods ^{[25, 26]}. It can reflect the vegetation growth status and distribution pattern on a global or regional scale, reflect the change trend of long-term vegetation sequence ^[27], have been widely used to study vegetation change in different regions, such as the Sahel ^[28], South Africa^[29] and on a global scale ^[30], with good results.

Therefore, the purpose of this study is to explore the driving factors of vegetation dynamics and the evolution of temporal and spatial patterns along the railway in Sichuan-Tibet region by integrating ecological factors and engineering activities. Specifically, this study innovatively proposes a general theoretical model for the spatial-temporal pattern evolution of vegetation coverage along the railway and its driving factors, select section a-b of a major railway project in Sichuan-Tibet region with complex geographical elements as the case study area, combine with Slope trend analysis and other methods, and analyze the long-term spatial-temporal pattern evolution of the MODIS-NDVI dataset along the a-b railway from 2001 to 2020. Using night light as the index factor to characterize engineering activities, the influence of engineering activities and ecological factors (single factor and interaction) on the local vegetation ecological environment before and after railway construction is discussed by using geographical detector. The significance of this empirical study is that, on the one hand, the driving factors of vegetation dynamics in section a-b of major railway projects in Sichuan-Tibet region are analyzed by integrating natural and human factors, which is helpful to understand the impact of railway projects on ecological environment of complex geographical regions such as plateau. On the other hand, the general model for the evolution of spatial and temporal patterns of vegetation cover and its drivers along the railway proposed in this study is applicable to the study under the conditions of extreme terrain and climate, which not only tests the applicability of the theoretical model, but also provides theoretical support for the ecological restoration and ecological environment improvement in other similar regions of the world.

2.1 Study area

The Major railway project in Sichuan-Tibet area (Fig. 1) is one of the main railway lines in southwest China and a fast channel connecting Sichuan and Chengdu. The railway stretches across five geomorphic units (Sichuan basin, western Sichuan alpine canyon area, western Sichuan plateau basin area, southeast Tibet Mt. Hengduan area and southern Tibet valley area) and six large rivers (Dadu River, Yalong River, Jinsha River, Lancang River, Nujiang River and Yarlung Zangbo River) from east to west. The terrain is generally high in the north and low in the south, and high in the west and low in the east, with an elevation difference of more than 3,000 meters that characterize the unique ecological and geological conditions. At the end of 2014, the construction of section a and section b began, and have been in operation in 2018 and 2021 respectively.

Section a-b crosses various climatic zones including the mid-subtropical humid climate zone, plateau temperate monsoon semi-humid climate zone and semi-arid monsoon climate zone, characterized by the strong wind, heavy rain, low average temperatures, a large temperature difference and strong ultraviolet radiation^[31]. The forests in the study area are mostly distributed along the eastern section of Nyingchi City and Ya 'an City. There are obvious artificial surface in Lhasa city and Chengdu city. The overall altitude of section b is low, the climate is humid, and the cultivated land and artificial surface are mostly.

2.2 Data collection and processing

Vegetation cover data for the study need to be calculated using NDVI remote sensing data, the NDVI dataset was obtained from MOD13Q1, a vegetation index product with a spatial resolution of 250 m, provided by the National Aeronautics and Space Administration, NASA (https://search.earthdata.nasa.gov). A total of 920 raw images (h26v05 and h26v06 images covering the study area, 46 images per year) from 2001–2020 were obtained from the website. The raw images were formatted and projected using the Modis Reprojection Tool (MRT), and the data were smoothed and denoised using the Savitzky-Golay filtering method to eliminate the interference of clouds, atmosphere, and solar altitude angle on the remotely sensed images ^{[32, 33]}. August is the period of maximum vegetation growth in the study area ^[34], therefore, the August NDVI dataset from 2001 to 2020 was selected for the calculation of vegetation cover. Pixels with NDVI less than 0.1 during the growing season were masked to exclude sparse and unvegetated areas ^[35].

The administrative division data were obtained from the latest national basic geographic information data published by the National Center for Basic Geographic Information (http://www.ngcc.cn) in 2021. The temperature (TEM) and precipitation (PRE) data from 2001 to 2020 were obtained from the National Meteorological Science Data Sharing Service Platform (http://data.cma.cn/site/index.html). And the generalized kriging interpolation method in ArcGIS was used to generate the annual mean temperature and precipitation data with 250-meter resolution. Elevation (DEM) data were obtained from Geospatial Data Cloud (https://www.gscloud.cn/) 90-meter resolution data. Slope (SLOPE) and slope direction (ASPECT) data were extracted using the ArcToolbox-3D Analyst tool. Nighttime light intensity (NLI) data were obtained from the DMSP-OLS-like dataset (1992–2022) calculated by combining calibrated DMSP-OLS data (1992–2013) and DMSP-OLS-like data (2013–2022) transformed from SNPP-VIIRS data.

2.3 Methods

2.3.1 Theoretical model

This research study synthesized the previous division of vegetation cover drivers and their measurement methods, and proposed a theoretical model for the evolution of spatial and temporal patterns of vegetation cover and its drivers along the railway analysis (Fig. 2). Specifically, this study divided the driving factors into five major ecological factors (TEM, PRE, DEM, SLOPE, ASPECT) and one engineering activity factor (NLI), and used buffer zone analysis, coefficient of variation and Slope trend analysis, kriging interpolation and natural breakpoint method, and Geodetector to analyze the evolution of spatial-temporal pattern of vegetation cover, the single-factor and interaction effects of the driving factors, and the effect distance of the railway on the vegetation cover, etc. It provides theoretical support for the analysis of the evolution of spatial and temporal patterns of vegetation cover and its driving factors along the railway.

2.3.2 Calculation of vegetation cover

Vegetation coverage refers to the percentage of the vertical projection of vegetation on the ground in the statistical area, which is an important indicator for ecological environment monitoring ^[36]. In this paper, the pixel dichotomy method was used to obtain the vegetation cover in the study area. The vegetation cover calculation formula is shown in Eq. (1).

$$FVC=\frac{NDVI-{NDVI}_{soil}}{{NDVI}_{veg}-{NDVI}_{soil}}$$

Where, FVC is the vegetation coverage, NDVI_soil is the minimum NDVI value in the study area, that is, the NDVI value of the bare land and non-vegetation covered pixels, and NDVI_veg is the maximum NDVI value in the study area, that is, the NDVI value of the whole vegetation covered pixels. Combined with the specific situation of the image, the confidence of 2% and 98% were selected as the values of NDVI_soil and NDVI_veg, and the NDVI value less than NDVI_soil was set as 0, and the NDVI value greater than NDVI_veg was set as 1, so as to estimate the vegetation coverage of the study area ^[37].

2.3.3 Buffer zone analysis

Buffer zone analysis is a common spatial geometric relationship analysis tool used in ArcGIS to analyze spatial neighborhoods. A buffer zone is a banded buffer zone of a specific distance around spatial objects of different geometric types to represent a specific sphere of influence or service area of a spatial target ^[38]. Due to the different geometries of spatial objects, buffers are constructed in different ways ^[39]. For road construction, the difference in the range of buffer distances reflects the impact of road construction on the ecology along the road ^[40], and can be used to check the relationship between the distance along the road and the vegetation cover index. The mathematical expression (2) is:

${B}_{i}=（x：d（{x}_{i}，{O}_{i}）\le R）$

(2)

where O_i is the spatial object, R is the neighborhood radius, d is the minimum Euclidean distance, and B_i is the set of all points whose distance to O_i is less than or equal to R.

As an independent data layer for overlay analysis, buffer zones can be applied to spatial analysis of roads, rivers and residential areas ^[41]. In order to accurately assess the impact of railway projects on vegetation coverage, this study used the buffer function of ArcGIS to generate a 5-km buffer zone on both sides of major railway projects in Sichuan-Tibet region.

2.3.4 Trend analysis

Slope trend analysis method was used to fit the change trend of vegetation coverage. This method can simulate the change trend of each grid and reflect the spatial-temporal evolution characteristics of regional vegetation pixel level ^[42]. This paper takes the start time of the railway in 2014 as the node to calculate the Slope values of 2001–2014 (before the railway construction) and 2015–2020 (after the railway construction), and the calculation formula (3) is as follows:

$$Slope=\frac{n{\sum }_{i=1}^{n}\left(i{FVC}_{i}\right)-{\sum }_{i=1}^{n}i{\sum }_{i=1}^{n}{FVC}_{i}}{n{\sum }_{i=1}^{n}{i}^{2}-{\left({\sum }_{i=1}^{n}i\right)}^{2}}$$

Where Slope is the slope of the regression equation, Slope > 0 indicates that the vegetation coverage in the studied time range presents an upward trend, and vice versa. n is the number of monitoring years, and FVC_i is the vegetation coverage in the year i. In addition, the change trend of FVC is divided into 5 types, and the classification standards are set in Table 1. Where σ represents the standard deviation and S the vegetation cover average, respectively.

Table 1

Classification standard of tendency value
Growing	Obvious deterioration	Slight degradation	No change	Slight improvement	Significant improvement
Slope	<S-σ	S-σ ~ S-0.5σ	S-0.5σ ~ S + 0.5σ	S + 0.5σ ~ S + σ	>S + σ
Note: S refers to the average of a 20-year pixel-by-pixel Slope for a city; σ refers to the standard deviation of the 20-year pixel per pixel Slope.

2.3.5 Coefficient of variation analysis

Coefficient of variation is a statistical index describing the degree of change of time series values, indicating the relative change (fluctuation) degree of geographic data ^[43], reflecting the dispersion degree of NDVI and the inter-annual fluctuation and stability of vegetation ^[44]. Specifically, a lower CV indicates less fluctuation of vegetation change and more stability, while a higher CV indicates more fluctuation of vegetation. The calculation formula (4) is as follows:

$CV=\frac{1}{\stackrel{-}{NDVI}}\sqrt{\frac{1}{n-1}\sum _{i=1}^{n}{\left({\varDelta NDVI}_{i}-\stackrel{-}{NDVI}\right)}^{2}}$

(4)

Where, n is the total number of years, equal to 20; NDVI_i is the NDVI value for year i (2001 is the first year); $\stackrel{-}{NDVI}$ is the NDVI average for the years 2001–2020. In this study, CV of NDVI was calculated on a pixel-by-pixel basis to represent vegetation fluctuation in the study area from 2001 to 2020. The natural intermittent point method is used to divide the fluctuation results into five levels: Low (CV < 0.05); Slightly low (0.05 ≤ CV < 0.1); Medium (0.1 ≤ CV < 0.15); Slightly high (0.15 ≤ CV < 0.2); High (CV ≥ 0.2).

2.3.6 Geodetector

In quantitative analysis, Geodetector is used to determine the spatial variability of factors affecting vegetation cover and the interaction between factors, to find the drivers of vegetation change ^[45]. The Geodetector can detect the interpretation ability of single factor and interactive factor to vegetation coverage. For factor detection, the interpretability of NDVI can be detected by potential impact factor X, and the calculation formula (5) is as follows:

$$q=1-\frac{1}{N{\sigma }^{2}}\sum _{i=1}^{k}{N}_{i}{\sigma }_{i}^{2}$$

Where q value is the magnitude of the explanatory power of vegetation cover influencing factors with values ranging between [0, 1]; N_i is the number of samples of class i, and N is the total number of samples in the study area; ${\sigma }_{i}^{2}$ is the variance of samples of class i, and σ² is the variance of the whole study area. The larger q is, the better it can explain the spatial heterogeneity of X to NDVI, and vice versa^[46].

Interaction detection was used to assess interaction between two factors. The q values of individual factors were first calculated separately, and the q value of two-factor interaction was calculated. The results are defined by comparing the q value of individual factor and two-factor interaction as shown in Table 2.

Table 2

Categories of factor interactions.
Foundation	Interaction
C = A + B	Independent
C > A + B	Non-linear enhancement
C < D	Non-linear weakening
D < C < E	Single-factor non-linear weakening
C > E	Dual-factor enhancement
Note: A = q(X1), B = q(X2), C = q(X1∩X2), D = Min(q(X1), q(X2)), E = Max(q(X1), q(X2)).

3.1 Spatial and temporal distribution characteristics of vegetation cover

Section a is located in southern Tibet, the vegetation is mainly alpine shrubs and alpine meadows, and the overall FVC value is low. The spatial differentiation of FVC along the railway is obvious (Fig. 3a), showing a distribution feature of "high in the east and low in the west". The low-value areas with FVC of 0-0.25 are mainly distributed in Lhasa section and west Shannan section as well as along the Yarlung Zangbo River basin. The area with FVC less than 0.5 accounts for 64.1%, and the area with FVC greater than 0.75 accounts for only 1.1%.

Section b is located in the eastern low altitude area, mostly cultivated land and artificial surface along the route. The remote sensing data in August was selected for the study, which is the season of vigorous crop growth. The overall vegetation coverage along the railway is higher than that in Section a (Fig. 3b), and the spatial distribution is generally "high in the west and low in the east". The area with FVC between 0.5–0.75 accounts for 71.1% of the total area.

In this study, the average FVC value along major railway projects in Sichuan-Tibet region from 2001 to 2020 was selected to describe the changing trend of FVC in the study area (Fig. 4), and the average FVC value in 2001 was used as the reference value to analyze the changes of it in each year from 2002 to 2020. On the whole, FVC along section a and b showed a fluctuating upward trend, and the vegetation coverage of section b was always higher than that of section a.

For section b, the altitude is low, the climate is humid, most of them are forest and cultivated land, except in 2003, the FVC growth value along the line is positive, and the growth curve fluctuates steadily; Section a is located in the high altitude area, and the change curve of FVC growth value along the line fluctuates greatly, with obvious negative areas appearing in 2010, 2015 and 2018. However, the annual cumulative growth value is 0.4, which is greater than 0, indicating that the vegetation growth trend in section a shows a better trend in the long-term series observation. Moreover, the growth rate of the two railway sections showed a peak in 2003. Combined with the local policies at that time, Tibet implemented the work of returning farmland to forest in 2003, which led to a significant growth trend in vegetation coverage from 2003 to 2004.

3.2 Change trend of vegetation cover

Taking 2014 as the time node, the change trend of vegetation coverage before and after railway construction was analyzed. According to the classification criteria in Table 1, the vegetation change types were classified into five groups: Obvious deterioration, Slight degradation, No change, Slight improvement and Significant improvement.

Comparing the spatial distribution of vegetation cover change types before and after the construction of the railway, as shown in Fig. 5 and Fig. 6, most of the types in the two sections showed a “No change” type. After the construction of the railway, the vegetation growth along section a showed an improved trend, while the vegetation growth along section b showed a worsening trend. According to statistics, from 2001 to 2014, the vegetation cover change type of section a and section b was “No change”, which accounted for 46.46% and 48.39%, respectively, and the vegetation proportion of “Obvious deterioration” or “Slight degradation” was 35.39% and 21.07%, respectively. During the period of railway construction from 2015 to 2020, the area of “Obvious deterioration” or “Slight degradation” vegetation cover in section a decreased from 35.39–28.91%, while that in section b increased to 41.75%.

We calculated the transfer matrix of each vegetation cover change type into other types before and after the construction of the railway based on the vegetation cover change types from 2001–2014 and 2015–2020. The trend of vegetation change from deterioration to improvement is called upward transfer, and vice versa is called downward transfer.

From the transfer matrix of vegetation cover change types (Table 3), it can be seen that the vegetation cover change types of “No change” and “Significant improvement” in section a are the most stable, with the probability of maintaining their own stability reaching 47.07% and 39.16%. Section b has the highest probability of maintaining its own stability with the vegetation change type of "No change", reaching 52%. However, in general, the vegetation cover of the two sections is not stable, and its diagonal probability is lower than the sum of the non-diagonal probabilities, i.e., the probability of the vegetation cover type changing to other types after the construction of the railroad is higher than the probability of the type not changing. In addition, the probability of downward shift of vegetation change type in section a (28.69%) is lower than the probability of upward shift (39.77%), which means that the vegetation growth along this section is not disturbed too much by the construction of the railroad, but rather improves; In section b, the probability of downward shift (49.23%) is higher than the probability of upward shift (18.55%), which means that the vegetation growth along this section deteriorates significantly after construction of the railroad. The construction of the railway may have affected the growth of vegetation, but whether it is the main factor needs further research.

Table 3

Transfer probability matrix of vegetation coverage change types from 2001 to 2020.
	Before building	After building					Area share (%)
	Before building	Obvious deterioration	Slight degradation	No change	Slight improvement	Significant improvement	Area share (%)
Section a	Obvious deterioration	7.17	15.59	45.21	11.85	20.18	14.52
	Slight degradation	7.26	18.19	56.21	7.85	10.50	20.87
	No change	10.11	24.11	47.07	7.43	11.28	46.46
	Slight improvement	11.56	16.65	35.47	11.73	24.58	8.28
	Significant improvement	9.66	11.24	27.84	12.11	39.16	9.87
	Area share (%)	9.16	19.75	45.85	8.98	16.26
Section b	Obvious deterioration	18.49	29.75	44.57	4.82	2.36	8.33
	Slight degradation	14.53	29.36	45.69	6.21	4.21	12.74
	No change	11.04	28.83	52.00	4.32	3.81	48.39
	Slight improvement	10.65	31.87	48.75	4.72	4.01	17.03
	Significant improvement	9.21	32.30	46.16	5.15	7.19	13.51
	Area share (%)	11.79	29.96	49.24	4.78	4.23

3.3 Stability of vegetation cover

According to the spatial distribution map (Fig. 7), the areas with higher stability of vegetation cover in section a are mainly distributed in the Lhasa and Shannan sections, and along the Yarlung Tsangpo River basin. According to the statistics, the area proportion of the stability grade of this section was High (32.95%) > Slightly high (27.8%) > Medium (16.56%) > Slightly low (13.85%) > Low (8.85%) from large to small.

Whereas the areas with higher stability of vegetation cover in section b are mainly distributed in the Ya'an and the southern part of Chengdu near the Meishan section. The area proportion of the stability grade of this section was Medium (32.41%) > Slightly high (30.63%) > High (16.95%) > Slightly low (16.11%) > Low (3.89%) from large to small. The proportion of areas with a vegetation coverage stability grade of medium or above in the two sections is 77.31% and 79.99%, respectively, and the stability of vegetation coverage was high.

Also, compared with Fig. 3, it can be seen that the vegetation coverage of the region with higher stability in section a is mostly lower than 0.5, while the vegetation coverage of the region with higher than 0.5 is relatively low. The vegetation coverage stability of section b does not show similar characteristics to that of segment a.

In order to further analyze the difficulty of improving the ecological environment of vegetation along major railway project in Sichuan-Tibet Region, the differences of vegetation growth change types under different vegetation coverage stability were analyzed.

As can be seen from Table 4, in section a, the areas with “Slightly high” vegetation stability accounted for 27.8%, and the sum of the corresponding vegetation change types of "Obvious deterioration" and "Slight degradation" accounted for 47.92%; The areas with “High” vegetation stability accounted for 32.95%, and the sum of the corresponding vegetation change types of "Obvious deterioration" and "Slight degradation" accounted for 73.55%; 90.12% of the regions with "Slightly low" and "Low" stability levels showed an improvement trend (“Slight improvement” and “Significant improvement”). Although the proportion of vegetation improvement in the low-stability area is very large, the area of the low-stability area only accounts for 8.85% of the total study area, while the area of the high-stability area accounts for 60.75%. It can be seen that there is a correlation between the stability of vegetation coverage in section a and the trend of vegetation change: the higher the stability, the worse the vegetation growth trend. This indicates that it is difficult to improve the ecological environment of vegetation along section a.

In section b, the “Low” stability area has 69.35% of the area where the type of vegetation change is well transitioned. But again, the “Low” stability area accounts for only 3.89% of the total area. Except for "low" stability, the area of "No change" vegetation change type accounted for the largest proportion of the other four stability types, followed by the area of "Slight improvement". Unlike section a, section b does not show a higher stability in the area of vegetation deterioration, which makes ecological restoration easier.

Table 4

Corresponding relationship between vegetation coverage change types and stability.
	Stability	Change types					Area share (%)
	Stability	Obvious deterioration	Slight degradation	No change	Slight improvement	Significant improvement	Area share (%)
Section a	Low	1.67	0.72	7.29	47.57	42.75	8.85
	Slightly low	3.30	2.31	23.36	53.86	17.17	13.85
	Medium	16.01	8.27	36.57	34.17	4.99	16.56
	Slightly high	31.85	16.07	38.37	12.88	0.83	27.80
	High	54.13	19.42	20.15	6.00	0.29	32.95
Section b	Low	9.25	2.58	18.82	35.59	33.76	3.89
	Slightly low	22.64	11.20	31.57	28.77	5.82	16.11
	Medium	13.46	13.10	48.03	24.06	1.34	32.41
	Slightly high	8.97	12.77	53.08	24.20	0.98	30.63
	High	4.05	2.99	58.19	33.24	1.53	16.95

3.4 Effect distance of railway on vegetation coverage

Five buffer zones of 0 ~ 250, 250 ~ 500, 500 ~ 1000, 1000 ~ 2000 and 2000 ~ 5000 meters were selected to calculate the vegetation coverage change type area proportion of each section of major railway projects in Sichuan-Tibet region with different buffer distances. This paper analyzes the influence of railway construction on vegetation ecological environment in different spatial scales.

As can be seen from Fig. 8a, vegetation change in section a shows a significant improvement trend when the distance is 2000m from the railway line, and the effect distance of the railway on the vegetation coverage in this section is 2000m. Within the effect distance, with the increase of the distance from the railway, the area of "Obvious deterioration" decreased from 11–10%, and the area of "Significant improvement" increased from 14–16%.

As can be seen from Fig. 8b, The change type of vegetation cover in section b was always relatively stable, and there was no obvious change trend along the two sections with the increase of buffering distance. The proportion of “Obvious deterioration” area in the two sections is relatively small, only 5%-16% in different spatial scales of the respective sections. It can be seen that the railway construction has caused damage to the vegetation along the line to a certain extent, but the effect on the vegetation growth in the whole study area is not obvious. Therefore, the vegetation growth situation in the study area may also be affected by other related factors such as terrain and climate.

3.5 Driving factors of vegetation cover

Precipitation (PRE), temperature (TEM), elevation (DEM), slope (SLOPE), slope direction (ASPECT) and night light intensity (NLI), which characterizes the intensity of engineering activities, were selected as driving factors to analyze the influence of these factors on the vegetation cover from 2001 to 2020. The geographic detector should discretize the influence factors before conducting the detection. The natural intermittent point have the advantage of classifying the influence factors according to the distribution law of the data itself to maximize the differences among various types. Therefore, the natural intermittent point method is chosen in this study, and then the geographic detector is used to detect the explanatory power of the classified factors.

Table 5

Detection results of single factor.
Influence factor	DEM	SLOPE	ASPECT	NTL	PRE	TEM
Section a	0.1461	0.2602	0.0660	0.0289	0.2169	0.2597
Section b	0.2387	0.0699	0.0016	0.4839	0.2541	0.2892

Table 6

Detection results of interaction of influencing factors.
Section a		Section b
Interaction factor	Interaction force	Interaction factor	Interaction force
PRE∩TEM	0.4549	DEM∩NTL	0.5711
SLOPE∩PRE	0.3967	NTL∩TEM	0.5576
DEM∩TEM	0.3863	NTL∩PRE	0.5530
DEM∩PRE	0.3803	SLOPE∩NTL	0.5172
DEM∩SLOPE	0.3707	ASPECT∩NTL	0.4858

As shown in Table 5, The correlation analysis of each factor with vegetation cover shows that the degree of its respective explanatory power in section a decreased by SLOPE > TEM > PRE > DEM >ASPECT >NLI, the explanatory power of SLOPE was just slightly higher than that of TEM). The influence of temperature and precipitation on vegetation cover in section a was also significant. Slope direction and engineering activities have little effect.

The explanatory power of the single factor in section b for FVC is as follows: NLI > TEM > PRE > DEM > SLOPE > ASPECT, the explanatory power of night light is significantly higher than other factors, and the correlation between FVC and night light intensity is the strongest, that is, the FVC of this section is mainly directly affected by the intensity of engineering activities. In addition, temperature and precipitation also have strong explanatory power for FVC. The explanatory power of slope direction in both routes is less than 0.1, indicating that slope direction has no significant effect on vegetation growth.

As shown in Table 6, Five groups of factors with high q value for two-factor interaction detection were selected for analysis. The explanatory power of the interaction between precipitation and temperature in section a is the strongest, reaching 0.4549. The interaction between slope and precipitation is second, with explanatory power of 0.3967. In section b, the interaction between altitude and night light had the strongest explanatory power for vegetation cover, with explanatory power of 0.5711. Night light had the strongest effect on the spatial pattern of vegetation cover (in fact, the interaction with any influence factor had a strong significance).

4.1 Spatial and temporal distribution of vegetation cover

Continuous and accurate monitoring of changes in spatial-temporal distribution of vegetation cover is of great significance to the study of global ecosystem and climate change ^[47]. Normalized difference vegetation index (NDVI) can reflect vegetation growth status, vegetation type and biomass, and is linearly correlated with vegetation coverage, so it has been widely used to study the vegetation dynamic ^[48]. From 2001 to 2020, NDVI during the growing season in the whole study area presents a fluctuating upward trend [65], with significant spatial heterogeneity, showing a pattern of overall improvement and local degradation ^[49], which is consistent with previous studies on population vegetation dynamics in Sicho-Tibet region ^[50–52]. ,

Due to the special geographical characteristics of section a, such as high altitude, dry climate and little precipitation, there are mostly sparse alpine meadows along the line, and animal husbandry is the main local industry, and regional economic development is highly dependent on meadow resources. Therefore, overgrazing is considered to be one of the main causes of regional meadow degradation ^{[53, 54]}. In addition, due to the lack of firewood, the inhabitants collected yak dung as fuel ^[55]. The nutrients that should be provided to the meadow in yak manure are deprived, leading to the degradation of the regional meadow to a certain extent ^[56]. These factors together lead to the overall low vegetation coverage of this section, and with the increase of altitude, vegetation coverage decreases, and the spatial distribution is higher in the east and lower in the west. It can be seen that the temporal and spatial distribution characteristics of vegetation coverage along the railway are the result of the joint action of natural factors and human factors, which is also consistent with the existing research conclusions ^{[57, 58]}.

Section b is located in the eastern low altitude area, and the land type along the line is mainly arable land. August data were selected for the study, which is the peak season for vegetation and crop growth, coupled with the fact that section b is located in a humid area with low altitude, high temperature, and abundant inter-annual rainfall, which is conducive to the propagation and growth of soil microorganisms. Therefore, the vegetation cover of this section is relatively high on the whole, and the average vegetation coverage reaches 0.58. However, after the construction of the railway, the vegetation change type of this section showed an obvious downward shift trend, indicating that in addition to the impact of ecological factors on vegetation, human activities also had a profound impact on the vegetation change along the railway, and in some areas, the impact exceeded the ecological environment itself ^[59].

It is worth noting that in 2003, the growth rate of vegetation coverage along section a and section b railway showed a peak, which may be related to the implementation of the return of farmland to forest in Tibet in 2003. In fact, human activity is a common result of many factors. Rapid urbanization driven by human development has altered original ecosystems, and urban expansion is the cause of reduced vegetation cover and environmental degradation, which has been observed in many cities around the world ^[60–62]. However, large-scale human activities actually pay more attention to the global environment, including vegetation cover, and accelerate the positive changes of some terrestrial ecological environments ^[63].

4.2 Driving factors of vegetation cover

In this study, 6 factors were used to study the driving factors of vegetation coverage, including 5 ecological factors (PRE, TEM, DEM, SLOPE and ASPECT) and 1 human factor (NLI). Because there is no statistically significant linear relationship between these factors and vegetation cover, or even between these factors ^[64], traditional correlation and regression methods are ineffective. Geographical detectors have no linear assumptions and have clear physical significance, and can measure spatial differentiation, detect drivers, and analyze the interaction between variables ^[22]. The results of Geodetector show that both natural factors and human factors have a certain impact on the vegetation coverage ^[65]. However, factors affecting the growth and distribution of vegetation vary greatly in different regions ^[66].

Natural factors are generally considered to be critical to the growth and distribution of vegetation. Some studies have pointed out that vegetation coverage increases exponentially with the increase of annual precipitation ^[67]. In section a, located at high altitude, the influence of natural factors on vegetation cover is relatively strong. Among them, slope has the strongest explanatory power to vegetation coverage. The altitude drop of section a is large, and the larger the slope, the more easily the soil is lost, and the more difficult the water and nutrients are to be maintained, resulting in the decrease of vegetation coverage ^[68]. In addition, the effects of temperature and precipitation on vegetation cover were also more significant, which is consistent with the findings of existing studies that vegetation in mainland China is most significantly affected by the interaction of temperature and precipitation ^[69]. For this section, due to its location in QTP, temperature is a limiting factor for vegetation cover at high altitudes ^[70]. With the rise of altitude, drought has also become another climatic feature of this region, and precipitation is the limiting factor of vegetation cover in low-precipitation areas, so precipitation also has the most important explanatory influence on vegetation change in this region ^[71]. This is consistent with previous research that vegetation in arid and semi-arid regions is most sensitive to precipitation ^[72].

Although natural factors have an important impact on temporal and spatial changes of vegetation, the impact of human activities cannot be ignored^[73]. In section b, which is located in the eastern part of the study area at a lower altitude, the explanatory power of human activity intensity on vegetation coverage is significantly higher than other factors, and vegetation growth in this area deteriorates with the construction of railway. This may be because rapid urbanization, accompanied by increased population density, changes natural habitats ^[24], leading to the loss of forests and grasslands, and thereby reducing vegetation cover ^[74]. It is worth noting that the impact of human activities on vegetation dynamics can not only lead to vegetation degradation through negative effects, but also promote vegetation restoration through positive effects ^[20]. Appropriate population increase and human disturbance are conducive to vegetation growth ^[73]. This may also be the reason why the construction of section a makes the upward transfer probability along the line greater than the downward transfer probability.

4.3 Suggestions for ecological restoration of vegetation coverage

Section a is located in the higher altitude area of Sichuan-Tibet region. Although there are grassland along the line, the grassland yield is small and the distribution is sparse. The harsh climate environment makes the survival rate of vegetation low. In order to improve the survival rate of vegetation and improve the soil, large areas of shrubs can be planted and drip irrigation can be implemented. The forest of section a is widely distributed, and a large number of fir, spruce and other trees are soaring into the clouds. It is necessary to divide the protected areas and establish ecological protection forests. In addition, government departments should introduce corresponding policies to put an end to indiscriminate logging.

Section b is mainly cultivated land and construction land, so the vegetation protection of this section should be carried out from the two aspects of establishing a reasonable cultivation system and limiting human economic activities. First of all, under the premise of ensuring the economic income of local residents, the farmland should be returned to forest, afforestation and planting grass according to local conditions, and the forest and grass vegetation should be restored. Secondly, the retained cultivated land should be planted green, the use of chemical fertilizers and pesticides should be reduced, the use of organic fertilizers should be increased, and intercropping should be carried out to improve the growing conditions of crops and regulate the soil. The population distribution along the section b is dense, and it is connected with the economically developed Chengyu metropolitan area. Rapid urban development will inevitably lead to urban expansion. Therefore, government departments need to reasonably plan urban land and delimit urban boundaries to reduce inefficient urban expansion.

4.4 Limitations and prospects

Vegetation dynamics are complex and related to many factors. In this study, six drivers were identified with reference to previous research, however, they are still not comprehensive. Some studies have shown that carbon monoxide concentration and nitrogen deposition are key factors in vegetation change, and further studies should be conducted when sufficient data is available ^[75]. However, due to the limitation of data sources, this paper does not take them into account. In addition, spatial resolution is crucial for studying vegetation cover. This study is based on the MODIS-NDVI dataset. Although the time series is long, the spatial resolution is low, which makes it impossible to analyze the different growth conditions of specific vegetation types. Future studies aim to introduce data with higher spatial resolution (such as Landsat NDVI) to improve the research results.

In this paper, a theoretical model of spatial-temporal pattern evolution of vegetation cover and its driving factors along the railway were proposed, and the a-b section of a major railway project in Sichuan-Tibet region was taken as the research area. The spatial and temporal distribution characteristics, change trend and stability of vegetation coverage along the railway were analyzed by buffer analysis and Slope trend analysis, based on the MODIS-NDVI data set and six driving factors from 2001 to 2020. And the spatial driving mechanism of various ecological factors and human factors on vegetation coverage was analyzed by geographical detector. The main conclusions are as follows:

(1) The construction of the railway has caused some damage to the vegetation along the line, but the effect on the vegetation growth in the whole study area is not obvious. NDVI in the growing season of a-b section of major railway projects showed an overall fluctuating upward trend.

(2) Both natural and human factors have a certain impact on vegetation coverage, but the spatial differentiation of vegetation coverage along the railway is obvious. The driving factors affecting vegetation growth and distribution vary greatly in different regions. For the high altitude area, slope, temperature and precipitation are the main factors. For low altitude areas, human activities dominates.

(3) There is a two-factor and nonlinear enhancement interaction between driving factors on FVC, and the interaction effect is greater than the independent effect. In section a, the interaction between precipitation and temperature has the strongest influence on FVC change. In paragraph b, the interaction between engineering activities and any of the impact factors is highly significant.

(4) The influence of human activities on vegetation dynamics is manifested in two aspects: negative effect leads to vegetation degradation, and positive effect promotes vegetation restoration. The excessive expansion of urbanization leads to the degradation of vegetation, and the appropriate increase of population is conducive to ecological restoration.

Author contributions

Jingxiao Zhang: Conceptualization, Formal analysis, Methodology, Software, Validation, Writing – original draft, Writing – review & editing, Conceptualization, Supervision, Methodology, Writing – original draft, Writing – review & editing. Ruixue Zhang: Data curation, Funding acquisition, Validation, Writing – review & editing. Shaoqi Sun: Project administration, Resources. Ballesteros-Pérez Pablo: Supervision, Validation, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This research is supported by Sichuan-Tibet Railway Major Fundamental Science Problems Special Fund (No. 71942006); Scientific Research Project of China Road and Bridge Engineering Co., Ltd (No. 2020-zlkj-04); Open Fund of China Academy of Railway Sciences Group Corporation (No. 2021YJ049); The National Social Science Fund projects (No.19FJYB017; No.20BJY010); List of Key Science and Technology Projects in China's Transportation Industry in 2018-International Science and Technology Cooperation Project (No. 2018-GH-006 and No. 2019-MS5-100); Fundamental Research Funds for the Central Universities (No.300102282601); Going Global Partnership: UK-China-ASEAN, Education Partnership Initiative funded by British Council (“Integrated Built Environment Teaching & Learning in the Joint Curriculum Development amid Digital-Driven Industry 4.0 among China, Vietnam, and UK”); Humanities and Social Sciences Research Project of the Ministry of Education of China (No. 21XJA752003); Shaanxi Social Science Funding(2023R001).

Data availability statement

Data can be accessed by contacting the corresponding author.

SHI C. Study on Ecological Compensation Mechanism of Railway Construction Projects [J]. E3S Web of Conferences, 2020, 145: 02047 (4 pp.)- (4 pp.).
HONG D, YONGHUAN J I N, JIANGUO C U I, et al. Review on ranges of ecological road-effect zones [J]. Journal of Zhejiang Forestry College, 2008, 25(6): 810-6.
PENG C, OUYANG H, GAO Q, et al. Building a" green" railway in China [J]. Science, 2007, 316(5824): 546-7.
ZHONG L, MA Y, SALAMA M S, et al. Assessment of vegetation dynamics and their response to variations in precipitation and temperature in the Tibetan Plateau [J]. Climatic Change, 2010, 103(3): 519-35.
VIVIROLI D, DüRR H H, MESSERLI B, et al. Mountains of the world, water towers for humanity: Typology, mapping, and global significance [J]. Water Resources Research, 2007, 43(7): n/a-n/a.
NIU F, GAO Z, LIN Z, et al. Vegetation influence on the soil hydrological regime in permafrost regions of the Qinghai-Tibet Plateau, China [J]. Geoderma, 2019, 354: 113892.
LI X M, MU P Q, WEN J K, et al. Carrier-Mediated and Energy-Dependent Uptake and Efflux of Deoxynivalenol in Mammalian Cells [J]. Scientific Reports, 2017, 7.
LIU S, HUANG S, XIE Y, et al. Spatial-temporal changes in vegetation cover in a typical semi-humid and semi-arid region in China: Changing patterns, causes and implications [J]. Ecological indicators, 2019, 98: 462-75.
TONG X, WANG K, BRANDT M, et al. Assessing future vegetation trends and restoration prospects in the karst regions of southwest China [J]. Remote Sensing, 2016, 8(5): 357.
DUAN H, XUE X, WANG T, et al. Spatial and Temporal Differences in Alpine Meadow, Alpine Steppe and All Vegetation of the Qinghai-Tibetan Plateau and Their Responses to Climate Change [J]. Remote Sensing, 2021, 13(4): 669.
QIN Y, ZHENG B. The Qinghai–Tibet Railway: A landmark project and its subsequent environmental challenges [J]. Environment, Development and Sustainability, 2010, 12(5): 859-73.
DING M, ZHANG Y, SHEN Z, et al. Land cover change along the Qinghai-Tibet Highway and Railway from 1981 to 2001 [J]. Journal of Geographical Sciences, 2006, 16(4): 387-95.
LIU Z, MENZEL L. Identifying long-term variations in vegetation and climatic variables and their scale-dependent relationships: A case study in Southwest Germany [J]. Global and Planetary Change, 2016, 147: 54-66.
ZHU Z, PIAO S, MYNENI R B, et al. Greening of the Earth and its drivers [J]. Nature Climate Change, 2016, 6(8): 791-5.
MONSON R K, TURNIPSEED A A, SPARKS J P, et al. Carbon sequestration in a high-elevation, subalpine forest [J]. Global Change Biology, 2002, 8(5): 459-78.
DU J, HE Z B, PIATEK K B, et al. Interacting effects of temperature and precipitation on climatic sensitivity of spring vegetation green-up in arid mountains of China [J]. Agricultural and Forest Meteorology, 2019, 269: 71-7.
DUAN J P, MA Z G, LI L, et al. August-September Temperature Variability on the Tibetan Plateau: Past, Present, and Future [J]. Journal of Geophysical Research-Atmospheres, 2019, 124(12): 6057-68.
FU G, SHEN Z X, ZHANG X Z. Increased precipitation has stronger effects on plant production of an alpine meadow than does experimental warming in the Northern Tibetan Plateau [J]. Agricultural and Forest Meteorology, 2018, 249: 11-21.
DU J, SHU J, YIN J, et al. Analysis on spatio-temporal trends and drivers in vegetation growth during recent decades in Xinjiang, China [J]. International Journal of Applied Earth Observation and Geoinformation, 2015, 38: 216-28.
LIU H, JIAO F, YIN J, et al. Nonlinear relationship of vegetation greening with nature and human factors and its forecast – A case study of Southwest China [J]. Ecological Indicators, 2020, 111: 106009.
XU X, LIU H, JIAO F, et al. Nonlinear relationship of greening and shifts from greening to browning in vegetation with nature and human factors along the Silk Road Economic Belt [J]. Science of The Total Environment, 2021, 766: 142553.
POLYKRETIS C, GRILLAKIS M G, ARGYRIOU A V, et al. Integrating Multivariate (GeoDetector) and Bivariate (IV) Statistics for Hybrid Landslide Susceptibility Modeling: A Case of the Vicinity of Pinios Artificial Lake, Ilia, Greece [J]. Land, 2021, 10(9): 973.
GILLESPIE T W, OSTERMANN-KELM S, DONG C, et al. Monitoring changes of NDVI in protected areas of southern California [J]. Ecological Indicators, 2018, 88: 485-94.
ZHANG Z, CHANG J, XU C-Y, et al. The response of lake area and vegetation cover variations to climate change over the Qinghai-Tibetan Plateau during the past 30years [J]. Science of The Total Environment, 2018, 635: 443-51.
TUCKER C J, SLAYBACK D A, PINZON J E, et al. Higher northern latitude normalized difference vegetation index and growing season trends from 1982 to 1999 [J]. International Journal of Biometeorology, 2001, 45(4): 184-90.
DING M, ZHANG Y, LIU L, et al. The relationship between NDVI and precipitation on the Tibetan Plateau [J]. Journal of Geographical Sciences, 2007, 17(3): 259-68.
BISWAS K, CHATTERJEE A, CHAKRABORTY J. Comparison of Air Pollutants Between Kolkata and Siliguri, India, and Its Relationship to Temperature Change [J]. Journal of Geovisualization and Spatial Analysis, 2020, 4(2): 25.
JAMALI S, SEAQUIST J, EKLUNDH L, et al. Automated mapping of vegetation trends with polynomials using NDVI imagery over the Sahel [J]. Remote Sensing of Environment, 2014, 141: 79-89.
MAXWELL S K, SYLVESTER K M. Identification of “ever-cropped” land (1984–2010) using Landsat annual maximum NDVI image composites: Southwestern Kansas case study [J]. Remote Sensing of Environment, 2012, 121: 186-95.
EASTMAN J R, SANGERMANO F, MACHADO E A, et al. Global Trends in Seasonality of Normalized Difference Vegetation Index (NDVI), 1982–2011 [J]. Remote Sensing, 2013, 5(10): 4799-818.
XU B, LI J, LUO Z, et al. Analyzing the Spatiotemporal Vegetation Dynamics and Their Responses to Climate Change along the Ya’an–Linzhi Section of the Sichuan–Tibet Railway [J]. Remote Sensing, 2022, 14(15): 3584.
CHEN J, JöNSSON P, TAMURA M, et al. A simple method for reconstructing a high-quality NDVI time-series data set based on the Savitzky–Golay filter [J]. Remote sensing of Environment, 2004, 91(3-4): 332-44.
CAO R, CHEN Y, SHEN M, et al. A simple method to improve the quality of NDVI time-series data by integrating spatiotemporal information with the Savitzky-Golay filter [J]. Remote Sensing of Environment, 2018, 217: 244-57.
LI Y, ZHOU J, WU X. Effects of the construction of Qinghai-Tibet railway on the vegetation ecosystem and eco-resilience [J]. Geographical Research, 2017, 36(11): 2129-40.
LIU Z, WANG H, LI N, et al. Spatial and temporal characteristics and driving forces of vegetation changes in the Huaihe River Basin from 2003 to 2018 [J]. Sustainability, 2020, 12(6): 2198.
ZHANG K, LYU Y, FU B, et al. The effects of vegetation coverage changes on ecosystem service and their threshold in the Loess Plateau [J]. Acta Geographica Sinica, 2020, 75(5): 949-60.
ZHANG X, XIANGNAN L, ZHIPENG Z, et al. Spatial-temporal pattern analysis of the vegetation coverage and geological hazards in Yanchi County based on dimidiate pixel model [J]. Remote Sensing for Natural Resources, 2018, 30(2): 195-201.
NARUMALANI S, ZHOU Y, JENSEN J R. Application of remote sensing and geographic information systems to the delineation and analysis of riparian buffer zones [J]. Aquatic Botany, 1997, 58(3-4): 393-409.
CUI M Q, ZHENG X J, LI Y, et al. Analysis of NDVI Trends and Driving Factors in the Buffer Zone of the Aral Sea [J]. Water, 2023, 15(13).
NEDBAL V, BROM J. Impact of highway construction on land surface energy balance and local climate derived from LANDSAT satellite data [J]. Science of The Total Environment, 2018, 633: 658-67.
XUE Y, BROWN D E. Spatial analysis with preference specification of latent decision makers for criminal event prediction [J]. Decision Support Systems, 2006, 41(3): 560-73.
D. STOW, DAESCHNER S, A. HOPE E A. Variability of the Seasonally Integrated Normalized Difference Vegetation Index Across the North Slope of Alaska in the 1990's [J]. International Journal of Remote Sensing, 2003, 24(5): 1111-7.
LEI Y, ZHU Y, WANG B, et al. Extreme lake level changes on the Tibetan Plateau associated with the 2015/2016 El Niño [J]. Geophysical Research Letters, 2019, 46(11): 5889-98.
JIANG W, YUAN L, WANG W, et al. Spatio-temporal analysis of vegetation variation in the Yellow River Basin [J]. Ecological Indicators, 2015, 51: 117-26.
ZHU L, MENG J, ZHU L. Applying Geodetector to disentangle the contributions of natural and anthropogenic factors to NDVI variations in the middle reaches of the Heihe River Basin [J]. Ecological Indicators, 2020, 117: 106545.
CHEN T, XIA J, ZOU L, et al. Quantifying the Influences of Natural Factors and Human Activities on NDVI Changes in the Hanjiang River Basin, China [J]. Remote Sensing, 2020, 12(22): 3780.
XIE Y, SHA Z, YU M. Remote sensing imagery in vegetation mapping: a review [J]. Journal of Plant Ecology, 2008, 1(1): 9-23.
ZHANG Y, JIANG X, LEI Y, et al. The contributions of natural and anthropogenic factors to NDVI variations on the Loess Plateau in China during 2000–2020 [J]. Ecological Indicators, 2022, 143: 109342.
CHEN B, ZHANG X, TAO J, et al. The impact of climate change and anthropogenic activities on alpine grassland over the Qinghai-Tibet Plateau [J]. Agricultural and Forest Meteorology, 2014, 189-190: 11-8.
PAN T, ZOU X, LIU Y, et al. Contributions of climatic and non-climatic drivers to grassland variations on the Tibetan Plateau [J]. Ecological Engineering, 2017, 108: 307-17.
LI L, ZHANG Y, LIU L, et al. Current challenges in distinguishing climatic and anthropogenic contributions to alpine grassland variation on the Tibetan Plateau [J]. Ecology and Evolution, 2018, 8(11): 5949-63.
LIU L, WANG Y, YOU N, et al. Changes in aridity and its driving factors in China during 1961–2016 [J]. International Journal of Climatology, 2019, 39(1): 50-60.
LIU L, WANG Y, WANG Z, et al. Elevation-dependent decline in vegetation greening rate driven by increasing dryness based on three satellite NDVI datasets on the Tibetan Plateau [J]. Ecological Indicators, 2019, 107: 105569.
OUYANG Z, ZHENG H, XIAO Y, et al. Improvements in ecosystem services from investments in natural capital [J]. Science, 2016, 352(6292): 1455-9.
WANG Y, WESCHE K. Vegetation and soil responses to livestock grazing in Central Asian grasslands: a review of Chinese literature [J]. Biodiversity and Conservation, 2016, 25(12): 2401-20.
BAI W, ZHANG Y, XIE G, et al. [Analysis of formation causes of grassland degradation in Maduo County in the source region of Yellow River] [J]. Ying Yong Sheng Tai Xue Bao, 2002, 13(7): 823-6.
HUI C, JINSONG L, YU C, et al. Ecological risk assessment of regions along the roadside of the Qinghai-Tibet highway and railway based on an artificial neural network [J]. Human and Ecological Risk Assessment, 2007, 13(4): 900-13.
LIU J. Calculation and Analysis of Ecological Environment Impacts along High-speed Railway [J]. Journal of Engineering Studies, 2017, 9(6): 601-8.
FANG Y, QIN D, DING Y, et al. The impacts of permafrost change on NPP and implications: A case of the source regions of Yangtze and Yellow Rivers [J]. Journal of Mountain Science, 2011, 8: 437-47.
DE CARVALHO R M, SZLAFSZTEIN C F. Urban vegetation loss and ecosystem services: The influence on climate regulation and noise and air pollution [J]. Environmental Pollution, 2019, 245: 844-52.
DE LA BARRERA F, HENRíQUEZ C. Vegetation cover change in growing urban agglomerations in Chile [J]. Ecological Indicators, 2017, 81: 265-73.
WU S, LIANG Z, LI S. Relationships between urban development level and urban vegetation states: A global perspective [J]. Urban Forestry & Urban Greening, 2019, 38: 215-22.
FENG D, FU M, SUN Y, et al. How Large-Scale Anthropogenic Activities Influence Vegetation Cover Change in China? A Review [J]. Forests, 2021, 12(3): 320.
YAMORI W, HIKOSAKA K, WAY D A. Temperature response of photosynthesis in C3, C4, and CAM plants: temperature acclimation and temperature adaptation [J]. Photosynthesis Research, 2014, 119(1): 101-17.
WU H, ZHANG J, BAO Z, et al. The impacts of natural and anthropogenic factors on vegetation change in the Yellow-Huai-Hai River Basin [J]. Frontiers in Earth Science, 2022, 10.
HAO J, XU G, LUO L, et al. Quantifying the relative contribution of natural and human factors to vegetation coverage variation in coastal wetlands in China [J]. CATENA, 2020, 188: 104429.
ZHANG M, WANG K, LIU H, et al. Spatio-Temporal Variation and Impact Factors for Vegetation Carbon Sequestration and Oxygen Production Based on Rocky Desertification Control in the Karst Region of Southwest China [J]. Remote Sensing, 2016, 8(2): 102.
JIA Z, LEI R, LIU Y, et al. Spatial–Temporal Variations and the Driving Factors of Vegetation Coverage in the Loess Hilly and Gully Region of China [J]. Forests, 2023, 14(6): 1238.
CHEN H, REN Z. Response of vegetation coverage to changes of precipitation and temperature in Chinese mainland [J]. Bull Soil Water Conserv, 2013, 2: 18.
GAO S, DONG G, JIANG X, et al. Quantification of Natural and Anthropogenic Driving Forces of Vegetation Changes in the Three-River Headwater Region during 1982–2015 Based on Geographical Detector Model [J]. Remote Sensing, 2021, 13(20): 4175.
CLELAND E E, CHUINE I, MENZEL A, et al. Shifting plant phenology in response to global change [J]. Trends in Ecology & Evolution, 2007, 22(7): 357-65.
OSTAD-ALI-ASKARI K, GHORBANIZADEH KHARAZI H, SHAYANNEJAD M, et al. Effect of climate change on precipitation patterns in an arid region using GCM models: case study of Isfahan-Borkhar Plain [J]. Natural Hazards Review, 2020, 21(2): 04020006.
HUO H, SUN C. Spatiotemporal variation and influencing factors of vegetation dynamics based on Geodetector: A case study of the northwestern Yunnan Plateau, China [J]. Ecological Indicators, 2021, 130: 108005.
WANG J, WANG K, ZHANG M, et al. Impacts of climate change and human activities on vegetation cover in hilly southern China [J]. Ecological Engineering, 2015, 81: 451-61.
PIAO S, YIN G, TAN J, et al. Detection and attribution of vegetation greening trend in China over the last 30 years [J]. Global Change Biology, 2015, 21(4): 1601-9.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

MODIS-NDVI (2001-2020) response to natural factors and engineering activities along railway in Sichuan-Tibet Region

Status:

Version 1

Abstract

Figures

1 Introduction

2 Material and methods

2.1 Study area

2.2 Data collection and processing

2.3 Methods

2.3.1 Theoretical model

2.3.2 Calculation of vegetation cover

2.3.3 Buffer zone analysis

2.3.4 Trend analysis

2.3.5 Coefficient of variation analysis

2.3.6 Geodetector

3 Results and Discussion

3.1 Spatial and temporal distribution characteristics of vegetation cover

3.2 Change trend of vegetation cover

3.3 Stability of vegetation cover

3.4 Effect distance of railway on vegetation coverage

3.5 Driving factors of vegetation cover

4 Discussion

4.1 Spatial and temporal distribution of vegetation cover

4.2 Driving factors of vegetation cover

4.3 Suggestions for ecological restoration of vegetation coverage

4.4 Limitations and prospects

5 Conclusion

Declarations

References

Additional Declarations

Status:

Version 1