Most precious studies focused on one single type of ecological functional area during single period, which ignored the differences of evaluation systems and index weights among different ecological areas in different study periods. In addition, the comparability of vulnerability assessment results among different study areas was poor. This paper proposed a novel-quantitative vulnerability evaluation method for multi-type and multi-temporal ecological functional areas using dynamic weighting method: Three-River Source region grassland meadow wetland ecological functional area(TRSR), Guiqiandian karst rocky desertification control ecological functional area(GQD), Hunshandake desertification control ecological functional area(HSDK), and Chuandian forest and biodiversity ecological functional area(CD), and then introduced the net primary productivity (NPP) to realize the determination of multi-type ecological vulnerability thresholds, which is helpful to compare the vulnerability evaluation results of different ecological functional areas in a unified and comparable level. The results showed that: (1) The proposed novel-quantitative vulnerability evaluation method had higher applicability in vulnerability assessment for multi-type ecological functional areas(91.1% for TRSR,91.9%for HSDK, 91.7% for CD, and 94.2% for GQD) based on dynamic weight determination method;(2) The determination of vulnerability thresholds based on NPP could provide a comparable level to investigate the spatial distribution patterns of ecological vulnerability in multi-type ecological functional areas for different periods; (3) The average ecological vulnerability of TRSR, GQD, and CD belonged to mild vulnerability, while that of HSDK belonged to moderate vulnerability; (4) There were significant differences in the dominant factors of ecological vulnerability for multi-type ecological functional areas in different periods. The research results could provide data and method supports for ecological protection for multi-type ecological zones in national scale.
Research Article
A novel-quantitative vulnerability evaluation method for multi-type ecological functional areas based on dynamic weight determination and NPP
https://doi.org/10.21203/rs.3.rs-2192236/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Multi-type functional areas
Ecological vulnerability
NPP
Quantitative evaluation
Driving mechanisms
Evaluation of ecological environment vulnerability has become a hot issue in the research of regional environmental evolution and sustainable development under the influence of global climate change (Che, 2013). With the rapid developments of economy, the intensity of the exploitation and utilization of natural resources are also increasing. As a result, the environment of key ecological functional areas in China has been seriously damaged. Especially since the 20th century, with the intensification of global warming climate warming, environmental problems, including frequent occurrence of extreme weather events, loss of species diversity, frequent occurrence of extreme weather events and intensification of desertification, the overall degradation or even loss of ecological functions, all pose great threats to regional or national ecological security and sustainable socio-economic development (Janpeter et al., 2012;Guo et al.,2022).
Researches on ecological vulnerability evaluation home and abroad mainly include two types of index system: comprehensive index and single index. A series of studies on ecological vulnerability evaluation and its driving mechanism have been conducted by many scholars. Adger (2006) believed that vulnerability was the vulnerability sensitivity of ecosystems due to lack of adaptability under the combined actions of natural and artificial changes. Liu et al. (2001) constructed the ecological vulnerability index in Sanjiang Plain, and then explored the spatial distributions of wetland ecosystem vulnerability. Qiu et al. (2007) proposed the ecological vulnerability model of western Hainan Island based on soil erosion index, landscape index, and desertification index, and then determined the dominant factors that affected the ecological vulnerability.
During the past decades, scholars have proposed many comprehensive evaluation methods of ecosystem vulnerability, including principal component analysis (PCA), analytic hierarchy process (AHP), grey correlation method, and entropy weight method. Aspinall and Pearson (2000) introduced the theories and technical methods of RS, GIS and landscape ecology into the field of ecological hydrology to establish the health index system of catchment area and then systematically evaluated the environmental health and changes at regional scale. Zheng (2016) analyzed the influencing factors of regional ecosystem vulnerability in the upper reaches of Minjiang River, and constructed the index system suitable for the ecological evaluation of the study area based on PSR model, and then analyzed the temporal and spatial change patterns of ecological vulnerability in the study area from multiple angles by using the methods of variation coefficient, gravity center model and transfer matrix. Fully considering the background characteristics of the ecological environment in the alpine region of the Qinghai-Tibet Plateau, Guo et al. (2018) introduced extreme climate factors and human disturbance factors to construct the ecosystem vulnerability evaluation system, and reasonably analyzed the temporal and spatial distribution law and driving factors of ecological vulnerability. Zhao (2019) analyzed the main causes of ecological vulnerability affecting Central Asia according to the principle of index selection, and constructed an index system based on the P-S-R model. Zhang (2019) established the VSD ecological vulnerability evaluation system to analyze the vulnerability status and evolution trend of large-scale (the whole study area) and small-scale (city). Nguyen et al. (2016) combined AHP, GIS and RS to establish a vulnerability assessment framework related to sixteen variables for the largest river system in Vietnam. Zhang et al. (2020) constructed the ecological vulnerability index of Wutai Mountain area based on P-S-R model and then analyzed the spatial and temporal changes of ecological vulnerability of Wutai Mountain area in four periods by using the methods of PCA, gravity center and difference analysis.
However, the above studies were mostly conducted aimed at single type of ecological function zone, and few study of ecological vulnerability evaluations for multi-type of ecological function zones has been reported. In addition, under the interactions of global change and continuous disturbances of human activities, the contribution rates of various factors in regional ecological vulnerability evaluation system in different historical periods changed significantly, but previous studies mostly adopted the same weights of index to carry out the vulnerability assessment. At the same time, due to the differences of the evaluation systems for multi-type ecological functional areas and the index weights in evaluation system for different historical periods, the ecological vulnerability evaluation results in different ecological areas and historical periods were not comparable. However, fewer studies were conducted to focus on the comparability enhancement of vulnerability assessment results among different study areas. How to scientifically determine its vulnerability threshold had become a key problem and important approach to deal with this problem in this research field.
Therefore, this study would establish the evaluation systems of TRSR, HSDK, GQD, and CD according to the regional ecological characteristics, and proposed the dynamic weight determination method to make the weight assignments for different ecological functional areas and different periods and finally introduced the NPP to determine the ecological vulnerability thresholds for multi-type ecological functional areas to confirm the comparability in a unified level. This proposed novel evaluation method could provide a new approach for the vulnerability assessment for multi-type ecological zones worldwide.
2.1 Study area
The Three-River Source region grassland meadow wetland ecological functional area (Fig. 1) is located in the southern part of Qinghai Province, with an average altitude of 3500 ~ 4800 m. There are many rivers, lakes and swamps in this region. The climate in this area belongs to the Qinghai-Tibet Plateau climate system, which is characterized by distinct dry and wet seasons, strong radiation, long sunshine duration, small annual temperature difference, and larger daily temperature difference. The precipitation is concentrated in summer with the average value of 262.2 ~ 772.8 mm.
The Hunshandake desertification control ecological functional area is mainly composed of Xilinguole, Inner Mongolia and Chifeng City, with a total area of 5.3×104km2. The natural environment belongs to the steppe zone of temperate semi-arid region, which is mainly affected by the temperate continental monsoon climate and prevailing west wind. The terrain shows a decreasing trend from southeast to northwest with small fluctuation (Yuan et al., 2021).
Chuandian forest and biodiversity ecological functional area, located in the southeastern margin of Qinghai-Tibet Plateau. The terrain shows a decreasing trend from the northwest to the southeast, with huge topographic relief. The topographic features include plateaus, mountains, plains, hills, sloping valleys, and lakes. The climate belongs to the monsoon climate, which is characterized by high temperature and abundant precipitation in summer, and rain and heat in the same period. However, the monsoon climate is unstable, and the regularity of the advance and retreat of the summer monsoon is abnormal, which will cause natural disasters such as drought and flood (Liu, 2017).
The Guiqiandian karst rocky desertification ecological functional area is a typical karst development area in China, involving Yunnan, Guizhou and Guangxi, with an average altitude of -78 ~ 504 m and a total area of about 3.2×105 km2. This area is dominated by the typical subtropical and tropical monsoon climate. The precipitation and temperature vary significantly with seasons and regions, and the average annual temperature and precipitation show an increasing trend from northwest to southeast. However, due to the vulnerability of the karst environment and the influence of human activities, a series of ecological problems such as soil erosion and rocky desertification have been caused (Ma et al., 2021).
2.2 Data source
Meteorological data are downloaded from China Meteorological Data Sharing Network, mainly including daily precipitation (0.1mm), annual average temperature(0.1℃), extreme high temperature(0.1℃), extreme low temperature(0.1℃), sunshine hours(0.1 hour), accumulated temperature(℃), frost-free period, the grid datasets (250m) are obtained through Kriging interpolation with ArcGIS 10.3. The 1:100,000 land use data are from the Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences. The DEM data derives from Geospatial Data Cloud Platform with a resolution of 90m. Land surface indicators, such as humidity index, bare soil index, improved-adjusted vegetation index, salinity index, surface temperature, and vegetation coverage were obtained from Landsat images with a resolution of 30m. The above datasets can be obtained from the geospatial data cloud platform (http://www.gscloud.cn/). The annual net primary productivity (1000m) is derived from MOD17A3, which can be freely obtained from NASA-EOS (National Aeronautics and Space Administration Earth Observation System). The projection type of all these above datasets is Krasovsky _1940_Albers All the above factors have been tested with higher precisions (Table 1).
|
Evaluation factor |
Data source |
Equations |
Overall accuracy |
RMSE |
|---|---|---|---|---|
|
Annual mean temperature |
823 meteorological stations |
\({\text{AMT}}=\frac{{\sum\limits_{{i=1}}^{N} {{T_d}} }}{N}\) |
91.2% |
0.657 |
|
Annual precipitation |
\({\text{AP}}=\sum\limits_{{i=1}}^{{365}} {{p_d}}\) |
93.4% |
0.964 |
|
|
Variation coefficient of annual precipitation |
\({\operatorname{P} _{cv}}=\frac{{{\sigma _{{\text{pre}}}}}}{{{M_{{\text{pre}}}}}}\) |
89.7% |
1.012 |
|
|
Variation coefficient annual mean temperature |
\({T_{cv}}=\frac{{{\sigma _{{\text{temp}}}}}}{{{M_{{\text{temp}}}}}}\) |
90.8% |
1.354 |
|
|
Extreme high temperature days |
\({\text{EHT}}=\sum\limits_{{i=1}}^{N} {{{\text{n}}_{T>{\text{THR}}}}}\) |
93.6% |
0.458 |
|
|
Extreme low temperature days |
\({\text{ELT}}=\sum\limits_{{i=1}}^{N} {{{\text{n}}_{T<{\text{THR}}}}}\) |
88.6% |
0.653 |
|
|
Proportion of erosive rainfall |
\({\text{PER}}=\frac{{\sum\limits_{{i=1}}^{N} {{p_d}} }}{{\sum\limits_{{i=1}}^{N} {{p_{d>12mm}}} }}\) |
85.4% |
0.854 |
|
|
Accumulated temperature (> 10 ℃) |
|
88.2% |
0.697 |
|
|
Sunshine hours |
\(Sh=\sum\limits_{{i=1}}^{N} {{d_{sunh}}}\) |
93.7% |
0.324 |
|
|
Frost-free period |
|
96.3% |
0.328 |
|
|
Humidity index |
Landsat images |
\(\begin{gathered} {\text{MI}}=0.0315 \times {B_{{\text{blue}}}}+0.2021 \times {B_{{\text{green}}}}+0.3102 \times {B_{{\text{red}}}}+ \hfill \\ 0.1594 \times {B_{{\text{nir}}}} - 0.6806 \times {B_{{\text{swir1}}}} - 0.6109 \times {B_{{\text{swir2}}}} \hfill \\ \end{gathered}\) |
96.1% |
0.425 |
|
Surface albedo |
\(\begin{gathered} {\text{Albedo}}=0.356 \times {B_{{\text{blue}}}}+0.13 \times {B_{{\text{red}}}}+ \hfill \\ 0.085 \times {B_{swir1}}+0.072 \times {B_{swir2}}-0.0018 \hfill \\ \end{gathered}\) |
90.9% |
0.357 |
|
|
Land surface temperature |
\({\text{LST}}=a(1 - C - D)+(b(1 - C - D)+C+D){T_{10}} - D{T_a}\) |
86.9% |
0.247 |
|
|
Bare soil index |
\({\text{BSI}}=\frac{{\left( {\left( {{B_{{\text{swir}}}}+{B_{{\text{red}}}}} \right)-\left( {{B_{{\text{nir}}}}+{B_{{\text{blue}}}}} \right)} \right)}}{{\left( {\left( {{B_{{\text{swir}}}}+{B_{{\text{red}}}}} \right)+\left( {{B_{{\text{nir}}}}+{B_{{\text{blue}}}}} \right)} \right)}}\) |
94.4% |
0.358 |
|
|
Desertification index |
\({\text{DI}}=\sqrt {{\text{LS}}{{\text{T}}^2}+{\text{NDV}}{{\text{I}}^2}}\) |
91.5% |
0.688 |
|
|
Salinization index |
\({\text{SI}}=\sqrt {{B_{{\text{blue}}}} \times B{\text{red}}}\) |
90.8% |
0.665 |
|
|
Rock exposure index |
\({\text{REI}}=\frac{{{B_{{\text{swir1}}}}/{B_{{\text{swir1,mean}}}}}}{{{B_{{\text{nir}}}}/{B_{{\text{nir,mean}}}}}}\) |
86.2% |
0.987 |
|
|
Altitude |
STRM DEM |
-- |
-- |
-- |
|
Slope |
-- |
-- |
-- |
|
|
Topographic relief |
\({\text{TR=}}{H_{\hbox{max} }} - {H_{\hbox{min} }}\) |
96.0% |
0.632 |
|
|
lithology |
Geological and hydrological map |
-- |
-- |
-- |
|
Water resources |
Land use data |
-- |
-- |
-- |
|
Water network density |
-- |
-- |
-- |
|
|
Vegetation coverage |
MODIS NDVI |
\({\text{VC}}=\frac{{{\text{NDVI}}-{\text{NDV}}{{\text{I}}_{\hbox{min} }}}}{{{\text{NDV}}{{\text{I}}_{\hbox{max} }}{\text{-NDV}}{{\text{I}}_{\hbox{min} }}}}\) |
86.7% |
0.359 |
|
Landscape diversity index |
Land use data |
\({\text{SHEI}}=\frac{{ - \sum\limits_{{i=1}}^{m} {({P_i} \times \ln {P_i})} }}{{\ln m}}\) |
-- |
-- |
|
Landscape fragmentation index |
\({C_i}=\frac{{{N_i}}}{{{A_i}}}\) |
-- |
-- |
|
|
Human activity intensity index |
Land use data |
\({\text{HAIS}}=\frac{{{S_i}}}{S} \times 100\) |
93.1% |
0.257 |
| Note: Bblue, Bgreen, Bred, Bnir, Bswir1, Bswir2 are the band reflectance of blue, green, red, swir1, and swir2, respectively; \(\tau\)is the atmospheric transmittance;\(\varepsilon\)is the land surface emissivity; T 10 is the radiance temperature of thermal infrared B10(K); C and D refer to the intermediate parameter; Ta is the average action temperature of the atmosphere(K);Td refers to the average daily temperature (℃); Pd refers to the daily precipitation (mm); \({\sigma _{{\text{temp}}}}\) refers to the standard deviation of temperature; \({M_{{\text{temp}}}}\) refers to the average value of temperature; \({\sigma _{{\text{pre}}}}\)refers to the standard deviation of precipitation;\({M_{{\text{pre}}}}\) refers to the average value of precipitation; dsunh refers to the daily sunshine duration (hour);nTL>2℃refers to the day with daily minimum temperature larger than 2℃;Ni refers to the patch number of landscape type i; Ai refers to the total area of landscape type i ;Pi refers to the area ratio of landscape type i. | ||||
3.1 Comprehensive index model
Comprehensive index evaluation method is the superposition of different evaluation indexes for comprehensive calculation. The comprehensive evaluation index of ecological vulnerability is to choose different indicators to build a comprehensive evaluation system, then determine the weights of different indicators in the system, and finally calculate the comprehensive index.
Considering that different indicators have different contributions to ecosystem vulnerability, it is necessary to introduce the concept of weight and assign different weights to different indicators, so as to highlight the dominant problem of regional ecological environment. The calculation formula is:
$${\text{EVI}}=\sum\limits_{{i=0}}^{n} {{f_i} \times {W_i}}$$
1
where EVI (Ecosystem Vulnerability Index) is the ecosystem vulnerability index; \({f_i}\)is index i; \({W_i}\) is the weight of the ith index; n is the number of indicators.
3.2 Factor Normalization
Considering the dimensional differences between different index factors and their role in the evaluation model, this study uses formula (2) and formula (3) to standardize the indicators.
Positive: \({I_i}=\frac{{{x_i} - {x_{\hbox{min} \cdot i}}}}{{{x_{\hbox{max} \cdot i}} - {x_{\hbox{min} \cdot i}}}}\) (2)
Negative : \({I_i}=\frac{{{x_{\hbox{min} \cdot i}} - {x_i}}}{{{x_{\hbox{max} \cdot i}} - {x_{\hbox{min} \cdot i}}}}\) (3)
where \({I_i}\) represents the normalized value of factor i, \({x_i}\) is the original value of factor i, \({x_{\hbox{max} \cdot i}}\) and\({x_{\hbox{min} \cdot i}}\)represents the maximum and minimum values of factor i, respectively.
3.3 Dynamic weight determination method
With the increasing or decreasing disturbances of climate change and human activity, the roles of different environmental problems in the ecosystem changes greatly. Therefore, it is not scientific and reasonable to adopt one index weight assignment scheme to evaluate the ecological vulnerabilities for different periods. Starting from the idea of dynamic weight determination, this study proposes a dynamic weight determination method based on FAHP, PCA, and the variation coefficient method, and then formulates the dynamic weight assignment schemes for different key ecological functional areas. The coefficient of variation is used to represent the normalized measure of the discrete degree of probability distribution. It can be defined as the ratio of standard deviation to average value, which is used to represent the spatial variation of the indication factor (index layer). The relative importance of each index in the FAHP can be determined by the variation coefficient. FAHP is utilized to determine the index of target and factor layers to ensure the relative consistency of index weight assignment scheme for different periods. This method can not only consider the spatial differentiation characteristics and the changes of ecological significance for each index in different periods, but also avoids the excessive interference of expert knowledge on weight assignment (Thomas and Tinde, 2019; Zhang et al., 2022). PCA is utilized to deal with multidimensional data to obtain the climate vulnerability in this paper.
3.4 Geographic detector
Geographic detector is a model to identify the interactions between driving factors on certain geographic phenomenon (Wang and Xu, 2017,Yao et al., 2022). This method can be utilized to determine the dominant influencing factors of ecological vulnerability with factor detector and interactive detector.
The factor detector can be used to detect the spatial differentiation of dependent variables, the explanatory factor can be used to detect the explanatory degree of dependent variables, and the q value can be used to measure. The formula of this detector is as follows:
$$q=1 - \frac{{\sum\limits_{{h=1}}^{L} {{N_h}\sigma _{h}^{2}} }}{{N{\sigma ^2}}}=1 - \frac{{{W_{{\text{SS}}}}}}{{{T_{{\text{SS}}}}}}$$
4
$${W_{{\text{SS}}}}=\sum\limits_{{h=1}}^{L} {{N_h}\sigma _{h}^{2}}$$
5
$${T_{{\text{ss}}}}=N{\sigma ^2}$$
6
where h refers to the stratification or partition of independent variable Y and dependent variable X; Nh refers to the number of units in layer h ; N represents the number of units throughout the region ; \(\sigma _{h}^{2}\)and \({\sigma ^2}\) are the variance of Y value in layer h and the whole region, respectively. Wss refers to the sum of variances for each layer ; Tss represents the total variance of the entire region. \(q \in [0,1]\), The greater the q value, the stronger the interpretation ability of X to Y ; on the contrary, the weaker.
3.5 Validation method
In this paper, the accuracy of different interpolation methods is evaluated by comparing the differences between the predicted values and the measured values of verification stations. In this article, the root mean square error (RMSE) are selected as the evaluation index (Guo et al.,2022). RMSE is used to measure the deviation between the estimated value and the predicted value, reflecting the extreme value and sensitivity of the data estimation. RMSE can reflect the overall accuracy of the interpolation results. Generally speaking, the smaller the value of RMSE is, the better the interpolation effect is, and the higher the accuracy of the interpolation model is. The equation of RMSE is as followed:
$${E_{RMS}}{\text{=}}\sqrt {\frac{1}{n}\sum\limits_{{i=1}}^{n} {{{({z_i} - {p_i})}^2}} }$$
7
Where \({z_i}\) and \({p_i}\) represent the predicted value and the measured value, respectively; n is the number of test points.
In this paper, the precision index and the basic error matrices are adopted to test the accuracy of the evaluation results of ecological vulnerability.
\({C_{i+}}=\sum\limits_{{{\text{j}}=1}}^{n} {{C_i}_{j}}\) ,\({C_+}_{j}=\sum\limits_{{i=1}}^{n} {{C_i}_{j}}\) (8)
\({P_{u,i}}={C_{ii}}/{C_{i+}}\) ,\({P_{A,j}}={C_{jj}}/{C_{+j}}\) (9)
where Pu,i represents the user accuracy of inversion category i, PA,j represents the cartographic accuracy of field observed category j, n is the number of categories, Ci+ is the sum of the inversion category i, C+ j is the sum of field observed category j, and Cij is the number of the inversion category i, and field observed category j that both occur.
4.1 Ecological vulnerability systems for multi-type key ecological functional areas
Based on the natural characteristics and the dominating ecological environment problems of different key ecological functional areas, the ecological vulnerability evaluation systems of multi-type key functional areas have been determined referring to relevant scientific researches at home and abroad. The vulnerability evaluation systems are as shown in Table 2–5.
| Target layer | Index layer | Factor layer | Direction | |
|---|---|---|---|---|
| Ecological vulnerability | Ecological sensitivity | Climate | Annual mean temperature | Negative |
| Annual precipitation | Negative | |||
| Variation coefficient of annual precipitation | Positive | |||
| Variation coefficient annual mean temperature | Positive | |||
| Extreme high temperature days | Positive | |||
| Extreme low temperature days | Positive | |||
| Proportion of erosive rainfall | Positive | |||
| Accumulated temperature (> 10 ℃) | Negative | |||
| Sunshine hours | Negative | |||
| Frost-free period | Negative | |||
| Soil | Humidity index | Negative | ||
| Surface albedo | Positive | |||
| Land surface temperature | Positive | |||
| Bare soil index | Positive | |||
| Desertification index | Positive | |||
| Salinization index | Positive | |||
| Terrain | Altitude | Positive | ||
| Slope | Positive | |||
| Topographic relief | Positive | |||
| Water | Water resources | Negative | ||
| Water network density | Negative | |||
| Ecological resilience | Vegetation | Vegetation coverage | Negative | |
| Ecological pressure | Land use | Landscape diversity index | Negative | |
| Landscape fragmentation index | Positive | |||
| Social development | Human activity intensity index | Positive | ||
| Target layer | Index layer | Factor layer | Direction | |
|---|---|---|---|---|
| Ecological vulnerability | Ecological sensitivity | Climate | Annual mean temperature | Negative |
| Annual precipitation | Negative | |||
| Variation coefficient of annual precipitation | Positive | |||
| Variation coefficient of annual mean temperature | Positive | |||
| Extreme high temperature days | Positive | |||
| Proportion of erosive rainfall | Positive | |||
| Accumulated temperature(> 10 ℃) | Negative | |||
| Sunshine hours | Negative | |||
| Frost-free period | Negative | |||
| Soil | Humidity index | Negative | ||
| Surface albedo | Positive | |||
| Land surface temperature | Positive | |||
| Bare soil index | Positive | |||
| Desertification index | Positive | |||
| Salinization index | Positive | |||
| Terrain | Altitude | Positive | ||
| Slope | Positive | |||
| Topographic relief | Positive | |||
| Water | Water resources | Negative | ||
| Water network density | Negative | |||
| Ecological resilience | Vegetation | Vegetation coverage | Negative | |
| Ecological pressure | Land use | Landscape diversity index | Negative | |
| Landscape fragmentation index | Positive | |||
| Social development | Human activity intensity index | Positive | ||
| Target layer | Index layer | Factor layer | Direction | |
|---|---|---|---|---|
| Ecological vulnerability | Ecological sensitivity | Climate | Annual mean temperature | Negative |
| Annual precipitation | Negative | |||
| Variation coefficient of annual precipitation | Positive | |||
| Variation coefficient annual mean temperature | Positive | |||
| Extreme high temperature days | Positive | |||
| Extreme low temperature days | Positive | |||
| Proportion of erosive rainfall | Positive | |||
| Accumulated temperature (> 10 ℃) | Negative | |||
| Sunshine hours | Negative | |||
| Frost-free period | Negative | |||
| Soil | Humidity index | Negative | ||
| Surface albedo | Positive | |||
| Bare soil index | Positive | |||
| Terrain | Altitude | Positive | ||
| Slope | Positive | |||
| Topographic relief | Positive | |||
| Water | Water resources | Negative | ||
| Water network density | Negative | |||
| Ecological resilience | Vegetation | Vegetation coverage | Negative | |
| Ecological pressure | Land use | Landscape diversity index | Negative | |
| Landscape fragmentation index | Positive | |||
| Social development | Human activity intensity index | Positive | ||
| Target layer | Index layer | Factor layer | Direction | ||
|---|---|---|---|---|---|
| Ecological sensitivity | Climate | Annual mean temperature | Negative | ||
| Annual precipitation | Negative | ||||
| Variation coefficient of annual precipitation | Positive | ||||
| Variation coefficient of annual mean temperature | Positive | ||||
| Extreme high temperature days | Positive | ||||
| Extreme low temperature days | Positive | ||||
| Proportion of erosive rainfall | Positive | ||||
| Accumulated temperature(> 10 ℃) | Negative | ||||
| Sunshine hours | Negative | ||||
| Soil | Humidity index | Negative | |||
| Surface albedo | Positive | ||||
| Bare soil index | Positive | ||||
| Rock exposure index | Positive | ||||
| Terrain | Altitude | Positive | |||
| Slope | Positive | ||||
| Topographic relief | Positive | ||||
| Geology | lithology | -- | |||
| Water | Water resources | Negative | |||
| Water network density | Negative | ||||
| Ecological resilience | Vegetation | Vegetation coverage | Negative | ||
| Ecological pressure | Land use | Landscape diversity index | Negative | ||
| Landscape fragmentation index | Positive | ||||
| Social development | Human activity intensity index | Positive | |||
4.2 Determinations of the index weight assignment scheme
Utilizing dynamic weight determination method that composed of FAHP, PCA, and CV, the index weights of the target layer and index layer were determined by FAHP, while that of factor layer was made with CV and PCA (for climate vulnerability)(Fig. 2). The index weights of target layer for different types of key ecological functional areas are shown in Table 6–9.
| Target layer | Index layer | Factor layer | Weights of 2000 | Weights of 2018 | |
|---|---|---|---|---|---|
| Ecological vulnerability | Ecological sensitivity (0.4) | Climate(0.19) | Annual mean temperature | -- | -- |
| Annual precipitation | -- | -- | |||
| Variation coefficient of annual precipitation | -- | -- | |||
| Variation coefficient annual mean temperature | -- | -- | |||
| Extreme high temperature days | -- | -- | |||
| Extreme low temperature days | -- | -- | |||
| Proportion of erosive rainfall | -- | -- | |||
| accumulated temperature (> 10 ℃) | -- | -- | |||
| Sunshine hours | -- | -- | |||
| Frost-free period | -- | -- | |||
| Soil(0.31) | Humidity index | 0.13 | 0.08 | ||
| Surface albedo | 0.12 | 0.26 | |||
| Land surface temperature | 0.13 | 0.14 | |||
| Bare soil index | 0.07 | 0.05 | |||
| Desertification index | 0.34 | 0.14 | |||
| Salinization index | 0.21 | 0.33 | |||
| Terrain(0.21) | Altitude | 0.16 | 0.16 | ||
| Slope | 0.47 | 0.47 | |||
| Topographic relief | 0.37 | 0.37 | |||
| Water(0.29) | Water resources | 0.6 | 0.50 | ||
| Water network density | 0.4 | 0.50 | |||
| Ecological resilience (0.25) | Vegetation (1.0) | Vegetation coverage | 1.0 | 1.0 | |
| Ecological pressure (0.35) | Land use (0.6) | Landscape diversity index | 0.6 | 0.5 | |
| Landscape fragmentation index | 0.4 | 0.5 | |||
| Social development (0.4) | Human activity intensity index | 1.0 | 1.0 | ||
| Target layer | Index layer | Factor layer | Weights of 2000 | Weights of 2018 | |
|---|---|---|---|---|---|
| Ecological vulnerability | Ecological sensitivity (0.35) | Climate (0.19) | Annual mean temperature | -- | -- |
| Annual precipitation | -- | -- | |||
| Variation coefficient of annual precipitation | -- | -- | |||
| Variation coefficient of annual mean temperature | -- | -- | |||
| Extreme high temperature days | -- | -- | |||
| Proportion of erosive rainfall | -- | -- | |||
| Accumulated temperature(> 10 ℃) | -- | -- | |||
| Sunshine hours | -- | -- | |||
| Frost-free period | -- | -- | |||
| Soil (0.37) | Humidity index | 0.24 | 0.04 | ||
| Surface albedo | 0.08 | 0.21 | |||
| Land surface temperature | 0.17 | 0.12 | |||
| Bare soil index | 0.17 | 0.10 | |||
| Desertification index | 0.1 | 0.16 | |||
| Salinization index | 0.24 | 0.37 | |||
| Terrain (0.15) | Altitude | 0.09 | 0.09 | ||
| Slope | 0.51 | 0.51 | |||
| Topographic relief | 0.40 | 0.40 | |||
| Water (0.29) | Water resources | 0.7 | 0.6 | ||
| Water network density | 0.3 | 0.4 | |||
| Ecological resilience (0.35) | Vegetation (1.0) | Vegetation coverage | 1.0 | 1.0 | |
| Ecological pressure (0.3) | Land use (0.5) | Landscape diversity index | 0.6 | 0.5 | |
| Landscape fragmentation index | 0.4 | 0.5 | |||
| Social development(0.5) | Human activity intensity index | 1.0 | 1.0 | ||
| Target layer | Index layer | Factor layer | Weights of 2000 | Weights of 2018 | |
|---|---|---|---|---|---|
| Ecological vulnerability | Ecological sensitivity (0.4) | Climate (0.15) | Annual mean temperature | -- | -- |
| Annual precipitation | -- | -- | |||
| Variation coefficient of annual precipitation | -- | -- | |||
| Variation coefficient annual mean temperature | -- | -- | |||
| Extreme high temperature days | -- | -- | |||
| Extreme low temperature days | -- | -- | |||
| Proportion of erosive rainfall | -- | -- | |||
| accumulated temperature (> 10 ℃) | -- | -- | |||
| Sunshine hours | -- | -- | |||
| Frost-free period | -- | -- | |||
| Soil (0.37) | Humidity index | 0.31 | 0.11 | ||
| Surface albedo | 0.53 | 0.56 | |||
| Bare soil index | 0.16 | 0.33 | |||
| Terrain (0.28) | Altitude | 0.26 | 0.26 | ||
| Slope | 0.38 | 0.38 | |||
| Topographic relief | 0.36 | 0.36 | |||
| Water (0.20) | Water resources | 0.7 | 0.6 | ||
| Water network density | 0.3 | 0.4 | |||
| Ecological resilience (0.35) | Vegetation (1.0) | Vegetation coverage | 1.00 | 1.00 | |
| Ecological pressure (0.25) | Land use (0.4) | Landscape diversity index | 0.50 | 0.50 | |
| Landscape fragmentation index | 0.50 | 0.50 | |||
| Social development(0.6) | Human activity intensity index | 1.00 | 1.00 | ||
| Target layer | Index layer | Factor layer | Weights of 2000 | Weights of 2018 | ||
|---|---|---|---|---|---|---|
| Ecological sensitivity (0.4) | Climate (0.11) | Annual mean temperature | -- | -- | ||
| Annual precipitation | -- | -- | ||||
| Variation coefficient of annual precipitation | -- | -- | ||||
| Variation coefficient of annual mean temperature | -- | -- | ||||
| Extreme high temperature days | -- | -- | ||||
| Extreme low temperature days | -- | -- | ||||
| Proportion of erosive rainfall | -- | -- | ||||
| Accumulated temperature(> 10 ℃) | -- | -- | ||||
| Sunshine hours | -- | -- | ||||
| Soil (0.29) | Humidity index | 0.32 | 0.21 | |||
| Surface albedo | 0.26 | 0.17 | ||||
| Bare soil index | 0.30 | 0.32 | ||||
| Rock exposure index | 0.12 | 0.30 | ||||
| Terrain (0.25) | Altitude | 0.36 | 0.36 | |||
| Slope | 0.37 | 0.37 | ||||
| Topographic relief | 0.27 | 0.27 | ||||
| Geology (0.14) | lithology | 1.00 | 1.00 | |||
| Water (0.21) | Water resources | 0.7 | 0.6 | |||
| Water network density | 0.3 | 0.4 | ||||
| Ecological resilience (0.3) | Vegetation (1.0) | Vegetation coverage | 1.00 | 1.00 | ||
| Ecological pressure (0.3) | Land use (0.55) | Landscape diversity index | 0.50 | 0.50 | ||
| Landscape fragmentation index | 0.50 | 0.50 | ||||
| Social development (0.45) | Human activity intensity index | 1.00 | 1.00 | |||
5.1 Determinations of the ecological vulnerability thresholds of multi-type ecological functional areas in different periods based on NPP
In this paper, we have introduced NPP to determine the ecological vulnerability thresholds for multi-type ecological functional areas in different periods, which avoided the randomness of vulnerability threshold definition, and also ensured the comparability of the ecological vulnerability among multi-type ecological functional areas in different periods. The detail processes were as follows (Fig. 3):
(1)The NPP of four key ecological functional areas in different periods (2000 and 2018) were divided into four levels (< 0.25; 0.25 ~ 0.5; 0.5 ~ 0.75;>0.75).
(2)The classifications datasets of NPP of four key ecological functional areas in different periods were utilized to obtain the average value of ecological vulnerability index corresponding to different grades.
(3)The average values of vulnerability index and the standard deviations were utilized to determine the vulnerability thresholds of multi-type ecological functional areas (Table 10).
| Levels of Vulnerability | TRSR | GQD | CD | HSDK | ||||
|---|---|---|---|---|---|---|---|---|
| (2000) | (2018) | (2000) | (2018) | (2000) | (2018) | (2000) | (2018) | |
| Slight vulnerability | < 0.49 | < 0.36 | < 0.27 | < 0.25 | < 0.31 | < 0.36 | < 0.26 | < 0.36 |
| Mild vulnerability | 0.49–0.51 | 0.36–0.49 | 0.27–0.33 | 0.25–0.31 | 0.31–0.47 | 0.36–0.44 | 0.26–0.35 | 0.36–0.48 |
| Moderate vulnerability | 0.51–0.56 | 0.49–0.54 | 0.33–0.39 | 0.31–0.35 | 0.47–0.51 | 0.44–0.49 | 0.35–0.40 | 0.48–0.51 |
| Intensive vulnerability | 0.56–0.59 | 0.54–0.58 | 0.39–0.47 | 0.35–0.38 | 0.51–0.55 | 0.49–0.51 | 0.40–0.44 | 0.51–0.53 |
| Severe vulnerability | > 0.59 | > 0.58 | > 0.47 | > 0.38 | > 0.55 | > 0.51 | > 0.44 | > 0.53 |
5.2 Validation of ecological vulnerability evaluation index
In order to validate the accuracy of the novel-quantitative vulnerability evaluation method for multi-type ecological functional areas, 270 validation samples (Fig. 4, taking the Three-River Source region grassland meadow wetland ecological functional area as example) were selected from different types of landscape areas by using field measured data, Google Earth and GF-2 satellite images. And then the error matrixes of ecological vulnerability index for different types of ecological functional areas (2018) were constructed, shown in Table 11–14. The results showed that the overall accuracies of the vulnerability evaluation method for multi-type ecological functional areas were 91.1%(TRSR), 91.9%(HSDK), 91.7%(CD), and 94.2%(GQD), respectively, indicating that proposed novel-quantitative vulnerability evaluation method had higher applicability for multi-type ecological zones.
| Vulnerability levels | Slight | Mild | Moderate | Intensive | Severe | Sum |
|---|---|---|---|---|---|---|
| Slight | 63 | 1 | 2 | 1 | 1 | 68 |
| Mild | 1 | 54 | 1 | 2 | 2 | 60 |
| Moderate | 1 | 0 | 49 | 1 | 1 | 52 |
| Intensive | 2 | 1 | 2 | 42 | 0 | 47 |
| Severe | 2 | 1 | 1 | 1 | 38 | 43 |
| Sum | 69 | 57 | 55 | 47 | 42 | 270 |
| Vulnerability levels | Slight | Mild | Moderate | Intensive | Severe | Sum |
|---|---|---|---|---|---|---|
| Slight | 70 | 2 | 0 | 1 | 1 | 74 |
| Mild | 2 | 60 | 1 | 1 | 2 | 66 |
| Moderate | 1 | 1 | 53 | 0 | 1 | 56 |
| Intensive | 1 | 2 | 2 | 46 | 2 | 53 |
| Severe | 2 | 0 | 1 | 2 | 58 | 63 |
| Sum | 76 | 65 | 57 | 50 | 64 | 312 |
| Vulnerability levels | Slight | Mild | Moderate | Intensive | Severe | Sum |
|---|---|---|---|---|---|---|
| Slight | 75 | 1 | 2 | 2 | 2 | 82 |
| Mild | 1 | 66 | 1 | 0 | 1 | 69 |
| Moderate | 0 | 2 | 54 | 2 | 1 | 59 |
| Intensive | 2 | 1 | 2 | 51 | 0 | 56 |
| Severe | 2 | 1 | 1 | 2 | 42 | 48 |
| Sum | 80 | 71 | 60 | 57 | 46 | 314 |
| Vulnerability levels | Slight | Mild | Moderate | Intensive | Severe | Sum |
|---|---|---|---|---|---|---|
| Slight | 78 | 0 | 1 | 3 | 1 | 83 |
| Mild | 2 | 74 | 0 | 1 | 1 | 78 |
| Moderate | 1 | 1 | 64 | 0 | 2 | 68 |
| Intensive | 2 | 2 | 1 | 85 | 0 | 90 |
| Severe | 0 | 1 | 2 | 1 | 56 | 60 |
| Sum | 83 | 78 | 68 | 90 | 60 | 379 |
5.3 Ecological vulnerability in multi-type ecological functional areas
Figure 5(a) showed that, the average ecological vulnerability index of Three-River Source region grassland meadow wetland ecological functional area was 0.48, which belonged to mild vulnerability. Zones of slight and mild vulnerability covered the area of 2.85×105 km2, accounting for 67.02%. It was mainly distributed in the south parts of TRSR, including southern Geermu City, western and eastern Zhiduo County, southern Maqin County, southern Xiahe County, the middle of Tongren County and Gonghe County. Zone of moderate vulnerability, covered an area of 7.46×104 km2, accounting for 17.55%, which was mainly distributed in northern Geermu City, northern Zhiduo County, northwestern Gonghe County and northern Maqin county. Zones of intensive and severe vulnerability was mostly located in the northwest part, with an area of 6.56×104 km2, accounting for 15.43%, including the middle of Ulan county and northern Zhiduo county.
Figure 5(b) showed that, the average ecological vulnerability index of Hunshandake desertification control ecological functional area was 0.35, which belonged to the moderate vulnerability. The area of slight and mild vulnerability zones was 1.21×105 km2, accounting for 72.84%, which was mainly concentrated in southeastern Xilinhot, southern Jining, northern Chifeng, eastern Chengde and southern Zhangjiakou. The second is the moderately vulnerable area, with an area of 3.18×104km2, accounting for 19.14% of the total area of the study area, scattered throughout the study area, including the north of Xilinhaote, the northeast of Jining, the south of Chifeng and the west of Chengde. The area of intensive and severe vulnerability zones was 1.33×104 km2, accounting for 8.02%, which were mostly concentrated in the west parts, including southern Xilinhaote City and northern Jining City.
Figure 5(c) showed that, the average ecological vulnerability index of Chuandian forest and biodiversity ecological functional area was 0.38, which belonged to mild vulnerability. Zones of slight and mild vulnerability covered an area of 2.15×105 km2, accounting for 68.60%, which were widely distributed in the whole study area, including northern Jinghong City, Simao City, Gejiu City, southern Mianyang City, western Maerkang County, southern Kangding County, southern Ya'an City, the middle of Xichang City, and northern Zhongdian County. Zone of moderate vulnerability had the second largest area of 6.23×104 km2, accounting for 19.86%, which was mostly located in southern Maqin County, northern Kangding County, southern Zhongdian County, northern Jinghong City and the middle of Nujiang Lisu Autonomous County. The area of intensive and severe vulnerability zones was 3.62×104 km2, accounting for 11.54%, which was scattered throughout the study area, including northern Mianyang City, the middle of Maerkang County, northwestern Yaan City, northwestern Kangding County and southern Xichang City.
Figure 5(d) showed that, the average ecological vulnerability index of Guiqiandian karst rocky desertification prevention and control ecological functional area was 0.30, which belonged to mild vulnerability. Zones of slight and mild vulnerability had the largest area of 5.83×104 km2, accounting for 66.39%, which were mostly concentrated in the south and middle parts, including eastern Wenshan County, the middle of Baise City, southern Xingyi City, southern Duyun City, southern Hechi City, southern Mashan County and the middle of Liupanshui City. Zone of moderate vulnerability was the second most widely distributed, with an area of 2.21 ×104km2, accounting for 25.19%, which was mainly concentrated in southern Wenshan County, the middle of Mashan County, southern Heshan City, northern Duyun City, northern Anshun City, and eatern Bijie City. The area of intensive and severe vulnerability zones was 0.74×104 km2, accounting for 8.42% in total, which were scattered throughout the study area, including southwestern Wenshan County, northern Liupanshui City, northwestern Anshun City and western Bijie City.
5.4 Change intensity of ecological vulnerability intensity changes in multi-type key functional areas
This study utilized the Grid calculator of ArcGIS10.3 to obtain the change intensity of ecological vulnerability, and then graded the vulnerability change intensity (CI) into five categories to further explore the spatial and temporal changes of ecological vulnerability in different types of key functional areas, as shown in Table 15.
| Levels | Severe decrease | Slight decrease | Stable | Slight increase | Severe increase |
|---|---|---|---|---|---|
| CI | CI≤-0.02 | -0.02 < CI≤-0.01 | 0.01 < CI ≤ 0.01 | 0.01 < CI ≤ 0.02 | CI > 0.02 |
During 2000–2018 in TRSR (Fig. 6(a)), zones of slight increase and severe increase were mostly located in the northwest parts, accounting for 33.64%, including western and northern Geermud City and Wulan County. Slight and severe decrease zones account for 44.19%, which were located in southern Zhiduo County, Maqin County, Xiahe County, Tongren County and southeastern Gonghe County. The stable zone accounted for 22.17% of the total area, which was concentrated in eastern Geermu City and western Gonghe County.
During 2000–2018 in HSDK (Fig. 6(b)), zones of slight and severe increase zones accounted for 25.87% a, which were concentrated in the northwest of the study area, including northern Jining City and Xilinhaote City. Slight and severe decrease zones accounted for 40.86%, which were located in the middle of Jining City, the middle and northern Chifeng City, the middle of Chengde City and Zhangjiakou City. The stable zone accounted for 33.26%, which was scattered in the whole study area, including southwestern Jining City, southeastern Xilinhaote City and southwestern Chifeng City.
During 2000–2018 in CD (Fig. 6(c)), the slight and severe increase zones accounted for 21.71%, which were scattered throughout the study area, including southern Mianyang City, western Kangding County, western and southern Ya an City, southwestern Maqin County, southwestern Zhongdian County, southwestern Simao City and southern Gejiu City. Sight and severe decrease zones accounted for 46.89%, which were mainly distributed in the middle of Maqin County, northernYaan City, Maerkang County, southern Jinghong City, Yunlong County, and Nujiang Lisi Autonomous Prefecture. The stable zone accounted for 31.40% and was scattered throughout the study area, including northern Mianyang City, northern of Kangding County, the middle of Xichang City, and southeastern Yaan City.
During 2000–2018 in GQD (Fig. 6(d)), the slight and severe increase zones accounted for 14.62% of the total area, while the slight and severe decrease zones accounted for 52.60%, which were mainly distributed in southwestern Wenshan County, Baise City, Hechi City, southern Xingyi City, southern Duyun City, Mashan County, Heshan City and Pingxiang City. The stable zone accounted for 32.78% of the total area, which was scattered in the whole study area, including southern Baise City, southern Duyun City and northeastern Hechi City.
In order to further investigate the dominant single factors and interactive factors of the change process of ecological vulnerability for multi-type ecological zones in different periods, this study selected eight typical factors, namely vegetation coverage, land use, solar radiation, precipitation, temperature, slope, altitude, and NPP. Utilizing the Geodetector model, the dominant single factors and interactive factors of multi-type key ecological functional areas were determined.
As shown in Table 16–17, the dominant single factor and dominant interactive factor of TRSR in 2000 were vegetation coverage and vegetation coverage ∩ precipitation, with q values of 0.60 and 0.84, respectively. In 2018, the dominant single factor and dominant interactive factor of the study region were vegetation coverage and vegetation coverage ∩ NPP, with q values of 0.95 and 0.97, respectively. The vegetation still played a dominant role in the regional ecosystem evolution and its contribution rates had been greatly enlarged. For HSDK in 2000, the dominant single factor and dominant interaction factor were vegetation coverage and vegetation coverage ∩ solar radiation, with the q values of 0.71 and 0.86, respectively. The vegetation and solar radiation contributed greatly to the health of regional ecosystem. The dominant single factor and dominant interaction factor of the study region in 2018 were NPP and vegetation coverage ∩ NPP, with q values of 0.91 and 0.98, respectively. With the effective implementations of afforestation project, the dominant role of vegetation in regional ecosystem had been enhanced. For CD in 2000, the dominant single factor was vegetation coverage with q value of 0.79, while the dominant interactive factor was altitude ∩ vegetation coverage with q value of 0.87. The vegetation and terrain had contributed greatly to the ecosystem health. In 2018, the dominant single factor was vegetation coverage with q value of 0.83, while the dominant interactive factor was NPP ∩ vegetation coverage with q value of 0.89. For GQD in 2000, the dominant single factor was solar radiation with q value of 0.37, while the dominant interactive factor was altitude ∩ vegetation coverage with q value of 0.66. In 2018, the dominant single factor was precipitation, with q value of 0.48, while the dominant interactive factor was altitude ∩ vegetation coverage, with q value of 0.72.
| Ecological subregion | Dominant single factor | q-value | Dominant interactive factor | q-value |
|---|---|---|---|---|
| TRSR | precipitation | 0.60 | Vegetation coverage∩precipitation | 0.84 |
| HSDK | Vegetation coverage | 0.71 | Vegetation coverage∩Solar radiation | 0.86 |
| CD | Vegetation coverage | 0.79 | Vegetation coverage∩altitude | 0.87 |
| GQD | solar radiation | 0.37 | Vegetation coverage∩altitude | 0.66 |
| Ecological subregion | Dominant factor | q-value | Dominant interactive factor | q-value |
|---|---|---|---|---|
| TRSR | Vegetation coverage | 0.95 | Vegetation coverage∩Precipitation | 0.97 |
| HSDK | NPP | 0.91 | Vegetation coverage∩NPP | 0.98 |
| CD | Vegetation coverage | 0.83 | Vegetation coverage∩NPP | 0.89 |
| GQD | Precipitation | 0.48 | Vegetation coverage∩Altitude | 0.72 |
7.1 Advantages of the novel-quantitative vulnerability evaluation method
Previous studies have conducted a series of studies on the ecological vulnerability evaluation of single-type ecological areas in a specific historical period, and have achieved certain results (Zhang et al.,2022). However, few studies on multi-type ecological zones had been reported. In this paper, a complete set of ecological vulnerability evaluation system oriented to multi-type ecological zones was proposed, which fully considered the differences of regional eco-environmental characteristics and dominant eco-environmental problems. Traditional subjective methods, sun as AHP and FAHP were greatly disturbed by expert experience and knowledge, which led to significant differences in the evaluation results of different studies in the same region (Zhang et al.,2014). Moreover, objective methods, such as PCA, entropy weight method, and variation coefficient method, could considered the image information features and spatial differentiation characteristics of all the evaluation index, but ignored the ecological significance of the index itself in the evaluation system (Ma, 2021). In addition, with the increasing disturbances of natural and artificial factors, the dominant factors of different ecological zones would change (Boori et al.,2022). Therefore, the proposed dynamic weight determination method by combining the variation coefficient method, FAHP and PCA could not only consider the spatial differentiation characteristics and ecological roles of the weight index itself, but also avoid the excessive interference of expert knowledge (Cai et al.,2022). Therefore, it could more accurately reflect the importance of each factor in different evaluation systems and different periods. However, there are some great differences in the evaluation system of multi-type key ecological functional areas, and the index weights in the evaluation system of different periods were also different. Therefore, the ecological vulnerability evaluation results were not comparable. NPP referred to the total amount of organic dry matter accumulated by green plants in unit time and area, which is part of the total organic matter generated by photosynthesis after deducting autotrophic respiration (Zhang et al., 2014). It was an important indicator of the health status of ecosystem and sustainable development (Chen et al.,2021). In this study, we introduce the NPP to determine the ecological vulnerability thresholds in multi-type ecological zones and different periods. It could indicate the sensitivity and vulnerability of regional ecosystem to a certain extent, which was helpful for the continuous spatial expression of ecological vulnerability among different ecological zones, avoiding the arbitrariness of the definition of vulnerability thresholds (Xiang et al.,2022). This method could ensure the comparability of ecological vulnerability among multi-type ecological zones in different periods (Yang et al.,2022). The results showed that the ecological vulnerability index evaluation method based on dynamic weighting methods and NPP had high applicability in multi-type key ecological functional areas, which could provide important reference for the ecological vulnerability evaluation of other regions worldwide.
7.2 Causes of ecological vulnerability in multi-type key ecological functional areas
The average ecological vulnerability index of TRSR was 0.48, which belonged to mild vulnerability. The TRSR was located in the Qinghai-Tibet Plateau with high altitude, scare precipitation, and low temperature. However, the abundant water resources that derived from glaciers and melting snow and widely distributed grassland had improved the ecological condition (Guo et al., 2020a). Zones with serious (intensive and severe) vulnerability were mostly distributed in the northwest of TRSR, where the precipitation was scare and the wind erosion and desertification were severe. The slight vulnerability zone was located in the source regions of Yellow River and Lancang River, and in these regions, the surface water resources were abundant and the vegetation coverage was higher. The average ecological vulnerability index of HSDK was 0.35, which belonged to moderate vulnerability. Zones of serious vulnerability were mostly concentrated in the west parts, where the evapotranspiration was very large and the Gobi was widely distributed. Meanwhile, the frequent sandstorm and low vegetation coverage had exacerbated the ecological vulnerability (Chen et al., 2021; Lalyer et al., 2021). The average ecological vulnerability index of CD was 0.38, which belonged to mild vulnerability as a whole. Zones of slight and mild vulnerability were mainly distributed in the middle of south of the study area. The reason was that in these regions, the vegetation coverage was higher, and the water and heat were sufficient (Liu et al., 2020; Felix et al., 2019). The intensive and severe vulnerability zones were mostly concentrated in the west and east parts of the study region. In the west parts, the freeze-thaw erosion was severe and the climate was characterized by low temperature and scare precipitation. In the east parts, the frequent earthquakes caused serious geological disasters, such as debris flow and landslide, which led to severe soil erosion (Chen et al.,2021). The average ecological vulnerability index of GQD was 0.30, which belonged to mild vulnerability as a whole. Zones of serious vulnerability were mainly distributed in the north and west parts, where the rocky desertification was severe (Guo et al.,2022).
7.3 Causes of evolution of ecological vulnerability in multi-type key ecological functional areas
In 2000, precipitation was the dominant single factor in the ecological functional area of the Three-River Source region grassland meadow wetland, and in 2018, vegetation cover was the dominant factor. This was because that during the past decades, a series of ecological protection measures, such as afforestation, prohibition of deforestation, and returning farmland to forest and grassland, had been effectively implemented, which achieved remarkable results with large area restorations of forest and grassland and strengthened the ecosystem diversity and species diversity (Guo et al., 2020b). The dominant factor of HSDK in 2000 and 2018 was vegetation. In this functional area, the climate was characterized by sandstorm, drought, scare precipitation and extreme high temperature, which directly affected and limited the grown and recovery of regional plant ecosystem and further accelerated the degradation process of grassland and forest land (Qi et al., 2021). The dominant factor of CD was vegetation and altitude. In this study region, the plateau and mountain were widely distributed, so that the climate and the natural environment were both significantly influences by the altitude and topographic features (Wu et al.,2021). The vertical climate types were characterized by abundant orographic precipitation and large temperature difference. In addition, there was rich ecological diversity and various types of vegetation (Mukesh et al., 2022). The dominant factors for GQD in 2000 were solar radiation and precipitation, while that of 2018 were precipitation, altitude, and vegetation coverage (Guo et al., 2019).Since 2003, the implementations of rocky desertification control projects and closing mountains for afforestation had greatly improved the conditions of vegetation. The increasing precipitation with the climate change had positive influences on the local ecological environment and microclimate, which became the dominant factor to affect the regional ecosystem (Sun et al., 2016).
Fully considering the geographical and climatic characteristics of multi-type key ecological functional areas, a set of ecological vulnerability evaluation systems oriented to different ecological functional areas has been constructed and a novel dynamic weight determination method has been proposed to determine the index weights. Finally, the NPP was introduced to define the vulnerability thresholds for multi-type ecological zones in different periods. The temporal and spatial changes of ecological vulnerability and its driving mechanisms were explored based on Geodetector. The main results were as follows:
(1) The novel-quantitative vulnerability evaluation method for multi-type ecological functional areas had higher applicability for multi-type key ecological functional areas and the evaluation precisions were 91.3%(TRSR), 92.6%(HSDK), 89.9%(GQD),and 90.8%(CD), respectively.
(2) The proposed dynamic weight determination method could better consider the difference of contribution rate of factors in the vulnerability evaluation system for different periods. In addition, the introduction of NPP to determine the vulnerability thresholds could confirm the comparability of the assessment results among different study regions in a unified level than previous studies.
(3) The average ecological vulnerability of TRSR, GQD, and CD belonged to mild vulnerability, while that of HSDK belonged to moderate vulnerability.
(4) There were significant differences in the dominant factors of ecological vulnerability for multi-type ecological functional areas in different periods, namely, TRSR: precipitation and vegetation (2000→2018, unchanged); HSDK: vegetation and solar radiation (2000) →vegetation (2018); GQD: solar radiation, vegetation, and altitude (2000) → precipitation, vegetation, and altitude(2018); CD: vegetation and altitude (2000) →vegetation (2018).
Authorship contributions
Xingming Yuan: Conceptualization, Methodology, Validation, Writing - original draft, Writing - review & editing; Bing Guo: Methodology, Validation, Visualization, Writing review & editing, Funding acquisition.
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Conflict of interest
The authors declare no competing interests.
Data availability
All data can be requested from the authors on reasonable request.
Funding
This work was supported by National Natural Science Foundation of China (grant no.42101306 and 42171413); Natural Science Foundation of Shandong Province (grant no. ZR2021MD047); Open Fund of Key Laboratory of Urban Land Resources Monitoring and Simulation, Ministry of Natural Resources (grant no.KF-2020-05-001), and the Major Project of High Resolution Earth Observation System of China (grant no.GFZX0404130304).
Acknowledgements
Not applicable.
- Adger WN (2006) Vulnerability Global Environ Chang 16(3):268–281
- Aspinall R, Pearson D (2000) Integrated geographical assessment of environmental condition in water catchments: Linking landscape ecology, environmental modelling and GIS. J Environ Manage 59(4):299–319
- Boori MS, Choudhary K, Paringer R, Kupriyanov A (2022) Using RS/GIS for spatiotemporal ecological vulnerability analysis based on DPSIR framework in the Republic of Tatarstan, Russia. Ecological Informatics,67,101490.
- Cai XR, Li ZQ, Liang YQ (2022) Tempo-spatial changes of ecological vulnerability in the arid area based on ordered weighted average model. Ecol Ind 133:108398
- Che LG (2013) Assessment of eco-environmental vulnerability over Guangxi Beibu Gulf economic zone. Guangxi Teachers Education University
- Chen ST, Guo B, Zhang R, Zang WQ, Wei CX, Wu HW, Yang X, Zhen XY, Li X, Zhang DF, Han BM, Zhang HL 2021.Quantitatively determine the dominant driving factors of the spatial–temporal changes of vegetation NPP in the Hengduan Mountain area during 2000–2015.Journal of Mountain Sciences, 18(2),427–445
- Chen ZQ, Zhang J, Zhang YL, Liu R (2021) Spatio-temporal patterns variation of ecological vulnerability in Otindag Sandy Land based on a vulnerability scoping diagram. Arid Zone Research 38(05):1464–1473
- Felix J, Russell T, Karin F (2019) Ecological vulnerability through Insurance? Potential unintended consequences of Livestock drought insurance. Ecol Econ 157:357–368
- Guo B, Chen ST, Han BM, Yang X, Fan YW, He TL, Yang WN, Jiang L (2019) Quantitative assessment of ecological vulnerability of the Silk Road Economic Belt, China based on the Partition-integration concept. Resour Environ Yangtze Basin 28(11):2601–2611
- Guo B, Jiang L, Fan YW (2018) Analysis of spatial and temporal changes and its driving mechanism of ecological vulnerability of alpine ecosystem in Qinghai Tibet Plateau. Ecol Sci 37(3):96–106
- Guo B, Wei CX, Yu Y, Liu YF, Li JL, Meng C, Cai YM ,2022.The Dominant Influencing Factors of Desertification Changes in the Source Region of Yellow River: Climate Change or Human Activity?Sci. Total. Environ., 813,152512
- Guo B, Yang F, Fan JF, Lu YF (2022) The changes of spatiotemporal pattern of rocky desertification and its dominant driving factors in typical karst mountainous areas under the background of global change[J]. Remote Sens 14:2351
- Guo B, Zang WQ, Luo W (2020a) Spatial-temporal shifts of ecological vulnerability of Karst Mountain ecosystem-impacts of global change and anthropogenic interference. Sci Total Environ 741:0048–9697
- Guo B, Zang WQ, Yang F, Han BM, Chen ST, Liu Y, Yang X, He TL, Chen X, Liu CT, Gong R 2020b. Spatial and temporal change patterns of net primary productivity and its response to climate change in the Qinghai–Tibet Plateau of China from 2000 to 2015.Journal of Arid Land, 12(1):1–17
- Janpeter S, Elke H, Jürgen S (2012) Climate change, vulnerability and adaptation in North Africa with focus on Morocco. Agr Ecosyst Environ 156:12–26
- Lalyer CR, Sigsgaard L, Giese B (2021) Ecological vulnerability analysis for suppression of Drosophila suzukii by gene drives. Glob Ecol Conserv 32:2351–9894
- Liu J (2017) Sichuan Yunnan image in the travel writings by foreigners in the 19th and 20th century. Yunnan Unversity
- Liu YM, Yang SN, Han CL (2020) Variability in regional ecological vulnerability: a case study of Sichuan Province, China. Int J Disast Risk SC 11(6):696–708
- Liu ZQ, Liu YH, Lv XG (2001) Ecological fragility of wetlands in Sanjiang Plain. Chin J Appl Ecol 12(2):241–244
- Ma BX, Jing JL, Xu Y, He HC, Liu B (2021) Spatial-temporal changes of NPP and its relationship with climate change in karst areas of Yunnan, Guizhou and Guangxi from 2000 to 2019. Ecol Environ Sci 30(12):2285–2293
- Ma CX (2021) Construction of multi-factor city remote sensing ecological environmental quality index-A case study of Tongchuan City. Stand Surveying Mapp 37(02):17–23
- Mukesh SB, Komal C, Rustam P, Alexander K (2022) Using RS/GIS for spatiotemporal ecological vulnerability analysis based on DPSIR framework in the Republic of Tatarstan. Russia Ecol Inform 67:1574–9541
- Nguyen AK, Liou YA, Li MH (2016) Zoning eco–environmental vulnerability for environmental management and protection. Ecol Indic 69:100–117
- Qi DH, Yang HX, Lu Q, Chu JM, Gan HH, Zhao S, Chen JW, Xu XQ (2021) Biodiversity of plant communities and its environmental interpretation in the Otindag Sandy Land, China. J Desert Res 41(06):65–77
- Qiu PH, Xu SJ, Xie GZ (2007) Analysis on the ecological vulnerability of the western Hainan Island based on its landscape pattern and ecosystem sensitivity. Acta Ecol Sinica 27(4):1258–1267
- Sun TT, Lin WP, Chen GS, Guo PP, Zeng Y (2016) Wetland ecosystem health assessment through integrating remote sensing and inventory data with an assessment model for the Hangzhou Bay. China Sci Total Environ 566–567:627–640
- Thomas HW, Tinde VA (2019) A social-ecological perspective on ecosystem vulnerability for the invasive creeper coralita (Antigonon leptopus) in the Caribbean: A review.Glob. Ecol. Conserv.18, e00605
- Wang JF, Xu CD (2017) Geodetector: Principle and prospective. Acta Ecol Sinica 72(01):116–134
- Wu HW, Guo B, Fan JF, Yang F, Han BM, Wei CX, Lu YF, Zang WQ, Zhen XY, Meng C (2021) A novel remote sensing ecological vulnerability index on large scale: A case study of the China-Pakistan Economic Corridor region, Ecological Indicators, 107955
- Xiang JW, Li XM, Xiao RB, Wang Y (2022) Effects of land use transition on ecological vulnerability in poverty-stricken mountainous areas of China: A complex network approach.Journal of Environmental Management,297,113206.
- Yang X, Liu S, Jiao C, Liu Y, Yu CC (2021) Vulnerability assessment and management planning for the ecological environment in urban wetlands. J Environ Manage 298:113540
- Yao K, Zhou B, He L, Liu B, Luo H, Liu DL, Li YX (2022) Changes in soil erosion in the upper reaches of the Minjiang River based on geo-detector. Res Soil Water Conserv 29(02):85–91
- Yuan ZH, Sa CL, Yin S (2021) Research on vegetation phenological changes in the Otindag sandy land Based on MODIS NDVI and EVI. China Environ Scince 41(11):5254–5263
- Zhang HL, Wu PS, Hou YJ (2020) Ecological vulnerability evaluation and Its Temporal and Spatial Changes in Wutai Mountain Area. J Ecol Rural Environ 36(08):1026–1035
- Zhang Q (2019) The impact assessment of ecological environment vulnerability based on land use/ cover change in Changbai Mountain area. Yanbian University
- Zhang Q, Yuan RY, Keke F, Shen ZX, Wang G, Zhao JQ (2022) Dynamic vulnerability of ecological systems to climate changes across the Qinghai-Tibet Plateau, China. Ecol Indic 134:108483
- Zhang YL, Qi W, Zhou CP (2014) Spatial and temporal variability in the net primary production of alpine grassland on the Tibetan Plateau since 1982. J Geogr Sci 24(2):269–287
- Zhang XY, Liu K, Wang SD, Wu TX, Li XK, Wang JN, Wang DC, Zhu HT, Tan C, Ji YH (2022) Spatiotemporal evolution of ecological vulnerability in the Yellow River Basin under ecological restoration initiatives. Ecol Ind 135:108586
- Zhao Y (2019) Evaluation and Zoning of ecological vulnerability in Central Asia. Beijing Forestry University
- Zheng J (2016) Study on the ecological fragility in the upper reaches of Minjiang River based on PSR model. Chengdu University of Information Engineering