Ecological Indicators 157 (2023) 111210

Contents lists available at ScienceDirect

Ecological Indicators
journal homepage: www.elsevier.com/locate/ecolind

Development of an adverse outcome pathways approach for land resource

and environment management at the regional scale
Kai Guo a, Xiaojin Liang b, *, Xinchang Zhang a, Renbo Luo a, Tianqi Qiu b, Zhenhao He a,
Kexin Zhang c
a
School of geography and Remote Sensing, Guangzhou University, Guangzhou 510006, China
b
Guangzhou Urban Planning & Design Survey Research Institute Co., Ltd, Guangzhou 510060, China
c
Map Institute of Guangdong Province, Guangzhou 510620, China

A R T I C L E I N F O A B S T R A C T

Keywords: At the regional scale, the complexity of ecological deterioration limits effective decision-making and preventive
Adverse outcome pathway measures. In this study, an adverse outcome pathway approach was developed and applied to land resource
Regional risk assessment management and environmental decision-making in Daye, a traditional mining city in China. First, adverse
Weight of evidence
outcome pathways for land degradation were hypothesized. Second, a weight of evidence approach was used
TOPSIS
based on the data strength and causal chain strength of each degradation pathway. Then, the technique for order
Land resource and environmental management
Daye preference by similarity to an ideal solution (TOPSIS) was used with entropy information and integrated with the
geographic information system (GIS) to evaluate the credibility of each pathway. The most likely adverse
outcome pathway and the spatial distribution for different sub-regions were obtained. The apparent spatial
heterogeneity suggests that land degradation pathways are closely associated with economic production activ­
ities in each sub-region. This study is expected to facilitate the identification of key mechanisms underlying
ecological degradation and to guide the management of land systems to ensure long-term functioning.

1. Introduction sustainable management of public lands (Brice et al., 2020). These

problems not only inhibit economic production but also directly
Unsustainable human development has caused the increasingly threaten human survival and sustainable development. Therefore,
serious degradation of land resources and the environment worldwide innovative ideas for land use and environmental management at a
(Bai et al., 2014; Bryan et al., 2018). Examples of this degradation regional scale are needed.
include monotonous surface landscapes caused by high-intensity land Various methods and technologies have been utilised for land
development (Brandolini et al., 2018), increases in unusable land due to resource management at a regional scale. For example, studies have
serious environmental pollution (Chen and Ye, 2014; Zhang et al., focused on a series of learning algorithms for optimizing habitat man­
2022), declines in land productivity resulting from soil depletion (Bor­ agement, such as artificial neural network, genetic algorithm and sys­
relli et al., 2017; Alam, 2018), losses of ecological function due to the tems thinking methods (Liu et al., 2013; Dawson, et al., 2017;
destruction of system resilience (Turner et al., 2016; Zhang et al., 2020), Chakraborti et al., 2018), propose and assess a model for evaluating and
and the exhaustion of resources, leading to stagnation and poverty (Ge mapping the land-use suitability used for sustainable utilization (Ristic
et al., 2019). At the same time, global warming and extreme weather et al., 2018; Luan et al., 2021), the construction of soil database and a
events caused by irrational development have direct or indirect impacts global digital map of soil pollution for protecting soils (Hou and Ok,
on global land resources and environmental management (Baron et al., 2019; Hou, 2022), satellite remote sensing combined with the multi-
2009; Joyce et al., 2009; Ellenwood et al., 2012), and in some regions objective algorithms to ease the monitoring of agricultural and grass­
climate change will exacerbate the ecological damage caused by the land loss (Schaefer and Thinh, 2019; Mao et al., 2018), and the use of
development of land resources, thus posing a number of threats to the multi-scenario simulation based on ecological network and causal

* Corresponding author.
E-mail addresses: [email protected] (K. Guo), [email protected] (X. Liang), [email protected] (X. Zhang), [email protected] (R. Luo), [email protected].
cn (T. Qiu), [email protected] (Z. He), [email protected] (K. Zhang).

https://doi.org/10.1016/j.ecolind.2023.111210
Received 19 July 2023; Received in revised form 22 October 2023; Accepted 1 November 2023
Available online 8 November 2023
1470-160X/© 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
K. Guo et al. Ecological Indicators 157 (2023) 111210

inference to guide future resources development (Su et al., 2019; Guo Heckera and LaLone, 2019). It allows for the introduction of other
et al., 2020a; Guo et al., 2020b). intelligent algorithms for expanded applications (Gust et al., 2015;
The aforementioned approaches and tools, however, can be further Collier et al., 2016). This approach has been successfully applied in
improved when applied to research on land resources and the environ­ environmental biology and ecology. However, this method has not been
ment. First, the combined effects of multiple factors were less concerned utilised, to our knowledge, in the management of land resources and
in land systems (Laurance et al., 2014; Maciel et al., 2023). Second, environments with similar ecological mechanisms to those of successful
there is a lack of understanding of the mechanisms underlying ecological cases.
degradation, leading to a failure to propose timely and effective coun­ Therefore, we applied QAOP-WOE to the management of a land-
termeasures. For example, a large area of ore pile that seems to be system, Daye City in China. In particular, we designed an approach to
managed may actually undergo a leaching process after long-term rain quantify hazard paths to guide the management and maintenance of
wash, causing the potential hazard of heavy metal contamination to regional land resources and environments. Risk path assumptions were
surrounding residential areas, farmland, water bodies and groundwater. based on “risk sources, stresses, receptors, and ecological endpoints” and
Third, owing to the difficulty in controlling the spatio-temporal scale, hazard paths were quantified using the weight analysis method. Po­
systematic issues are overlooked. After all, the deterioration effects of tential paths were ranked using an improved technique for order pref­
land ecosystem cannot be reflected unless on a large scale. Further, the erence by similarity to ideal solution (TOPSIS). Decisions corresponding
long action cycle of some hazardous factors (e.g., heavy metal accu­ to different stages of hazard paths were proposed to promote the
mulation entails a long latent period in causing soil contamination) also development of effective management plans for regional land resources.
brings difficulty in assessing the environmental degradation. Fourth,
there is an over-dependence on quantitative computational methods and 2. Data and methodology
model simulations. Finally, large-scale land census and statistics entail
huge costs and decision-making, and preventive plans often emphasise 2.1. Study area
treatment over prevention, without applying ecological principles to
land resource management. To address the above drawbacks of previous Daye is located in south-eastern Hubei Province in Central China
work, in this study, we developed a method for the quantitative analysis (114◦ 31′–115◦ 20′ E, 29◦ 40′–30◦ 15′ N) near the Yangtze River (Fig. 1).
of adverse outcome pathways (QAOPs) combined with weight of evi­ The area is rich in mineral deposits and has well-developed mining and
dence (WOE) to the land resource and environment management, is a metallurgy industries. The ecosystem in and around the city is
reasonable and significant endeavour. comprehensive, including lakes, rivers, forests, mines, arable lands,
QAOP-WOE framework is widely used in risk analysis both to health gardens, and urban and rural residential areas. The entire study area was
and to the environment. The approach interprets lines of evidence for 1566.3 km2. Growing industries have harmed the ecosystem in the
identifying complicated interactions, thereby providing assistance to entire area, resulting in dysfunctions in ecosystem services, including
assessors in verifying multiple hypotheses in complex ecosystems soil contamination, water pollution, and declines in land productivity
(Conolly et al., 2017; Moe et al., 2020; Guo et al., 2021). For example, it (Zhang et al., 2018; Guo et al., 2020a; Guo et al., 2020b). Daye city is
improves our understanding of the systematic degradation process and within the scope of the current comprehensive land improvement
complex ecological mechanisms (Kramer et al., 2011; Fay et al., 2017). It project in China, which involves the basic work of natural resource
provides a basis for the development of causal hypotheses and identi­ management, improvements in the eco-environment, and intensive
fication of risk factors (Ankley et al., 2010; Perkins et al., 2018). It can be land-use (Cai et al., 2019; Zhang et al., 2021). To improve the man­
flexibly applied at different scales to gain consensus among stakeholders agement of land resources and the environment under the regional risk
(Becker et al., 2015; LaLone et al., 2017). Its technological imple­ perspective, the study area was divided into six sub-regions based on
mentation is not limited by a lack of field experiments and data (Lee economic, ecological, and geographical features. The boundaries of the
et al., 2015). It combines qualitative and quantitative approaches to sub-regions were defined based on administrative divisions (Fig. 1).
avoid over-dependence on model computation (Kleinstreuer et al., 2016;

Fig. 1. Map of Daye City and the distribution of mining areas.

2
K. Guo et al. Ecological Indicators 157 (2023) 111210

2.2. AOP analysis-based research framework (land resources officers, representatives of industrial and mining en­
terprises, local residents, and other participants). After the problem
China has incorporated territorial space planning into the 14th Five- statement was formulated, a list of candidate AOPs that influence the
Year Plan, focusing on promoting the comprehensive improvement of stability and health of land ecosystems was established. A series of
land resources (including the reuse of land resources and the restoration possible hypotheses should be effectively proposed to explain the
of ecological functions) (Zhang et al., 2018; Guo et al., 2019). This manner by which these candidate AOPs exert negative ecological effects
highlights the importance of a scientific design for land resource and while diverging from the earlier hypothesis (Platt, 1964; Suter II and
environmental management in a regional sustainable economic model, Cormier, 2011).
requiring intelligent decision-making tools. In this study, we developed
a method for the quantitative analysis of adverse outcome pathways 2.3.2. Development of a conceptual model of adverse outcome pathways
(QAOPs) combined with weight of evidence (WOE) to represent the The initial list of alternative hypotheses should include all AOPs that
reliability of AOPs. Drawing on risk assessment techniques of the United captured the causal relationships between potential sources and end­
States Environmental Protection Agency (USEPA), land degradation in points, describing the complex ecological mechanisms. This process
each sub-region was analysed, and AOPs were developed on “risk requires an extensive literature review to ensure that the hypotheses are
source, stress, receptor, and adverse outcomes”. The weight of evidence rigorous and logical. A conceptual model (Fig. 3) was constructed to
approach was used to evaluate evidence, including the data strength and represent the complex interactions among sources, stressors, habitats,
causal chain strength. TOPSIS was then combined with GIS to calculate and endpoints, and their linkages generated the AOP through the
and sort the reliability of each AOP (Fig. 2). On this basis, land resource abovementioned causal hypotheses. The model eventually characterised
management strategies were designed for each sub-region. the degenerative process of land resources and the environment (i.e. risk
formation) by describing various ecological mechanisms. Not all AOPs
were included in this model; however, they represent a concise profile of
2.3. Preparing the adverse outcome pathway
land degradation in the study area.
2.3.1. Risk analysis of the land system
Following USEPA risk assessment concepts (Suter, 1993; USEPA, 2.4. Calculation of AOP confidence at the regional scale
2010), an AOP of regional land resources and the environment was
developed. A risk analysis of the land system was conducted by the The WOE methodology combines different types of evidence to
organisation of risk sources, stressors, acceptors, and eco-endpoints validate an apparent association between causes and effects or to obtain
(Guo et al., 2020a; Guo et al., 2020b), and determined as described in support for multiple hypotheses (Linkov et al., 2009; Suter and Cormier,
the Table 1. Risk analysis and variable determination were listed in the 2011). This approach has been used successfully to evaluate various
supplementary material. impaired ecosystems to reveal complex mechanisms and identify risk
One risk source can have various environmental impacts and can variables (Giubilato et al., 2014; Bridges et al., 2017; Leonarda et al.,
result in large-scale land degradation via multiple paths. Accordingly, 2019). We applied this method to quantify the AOPs of the land
each of these pathway hypotheses should be considered. A group of ecosystem at a regional scale, including the assignment of evidence
exerts was formed including scholars in relevant fields (ecology, geol­ strength based on “data and causality” and calculation of the confidence
ogy, aquaculture, and urban environment) and various stakeholders coefficient score using TOPSIS.

Fig. 2. Procedure for the AOP analysis of land degradation on a regional scale.

3
K. Guo et al. Ecological Indicators 157 (2023) 111210

Table 1
Criteria for risk sources ranking and data processing in each sub-region.
Source (0.28) Stressors (0.24) Habitats (0.21) Endpoints (0.27)
Chemical action Physical damage Biological Humanistic society
(0.32) (0.27) invasion (0.19) (0.22)

Storms and floods Organic pollutants Altered hydrology Pathogens Behavioral and physical Lakes Soil depletion
Geological disasters Heavy metals Destroyed earth disturbances Rivers Water shortage
accumulation surface
Acid rain Altered earth-surface Wetlands Soil contamination
Urbanization Sulfur deposition runoff Forests Changes in
Mining Soil salinization Desertification Grasslands sedimentation patterns
Agricultural Water Eutrophication Habitat removal Arable lands Water quality
intensification deterioration
Lake-area reclamation Nutrient Runoff Soil erosion Urban and rural Usable land shortage
Aquaculture Soil acidification Space occupation settlements Decline in land
productivity
Sewage emission Leaching Altered underground Biodiversity decreased
Solid-waste pile Solubility inorganic runoff Public safety and health
Application of salt Atmospheric threats
pesticides deposition
Application of Reduce in Eco-resiliency
fertilizers
Irrigation Landscape aesthetics
Biological resource dysfunction
consumption

Fig. 3. Identification of adverse outcome pathways for land degradation based on a conceptual model for the Daye area. In the figure, the variables S1-14, P1-20, H1-
7 and E1-12 represented the fourteen sources of risk, twenty pressures, seven habitat receptors, and twelve effect endpoints, respectively. The graph size of the
variables in this graph indicates the strength of the data, while the thickness of the lines between them characterizes the causal chain strength.

2.4.1. Criteria weighting and scoring (Table 1).

Based on data strength and causal chain strength, the WOE was (2) Based on the strength of causal relationships between risk factors,
analysed and values were assigned. an evidence analysis was conducted and values were assigned. The
(1) Data-level evidence analysis included risk sources, receptors, and causality analysis involved a pathway analysis of “risk source-stress”,
the strength of adverse outcomes. (a) Risk source data included the “stress-receptor”, and “receptor-endpoint”. (a) The “risk source-stress”
actual release and input intensity of human production activities, pathway represented whether a risk source generates/releases stress
emission intensity of industrial wastewater and waste gas, and urban factors, and values of 1.5, 1.0, and 0 were assigned, indicating highly
sprawl intensity. (b) The receptor data were measured based on biotope likely, generally likely, and unlikely release, respectively. (b) The
area ratios to represent vulnerability. (c) The analysis of adverse out­ “stress-receptor” pathway analysis targeted different exposure levels of
comes was based on the strength of field monitoring data, such as water stressors to biotope receptors in different sub-regions to reflect spatial
pollution levels, heavy metal accumulation, and soil organic matter heterogeneity, which was assigned values of 3, 2, 1, or 0. (c) Parameter
levels. Values of 1.0, 1.5, 2.0, and 3 were assigned to indicate high, λ1 was designed in light of the properties of stressors, such as chemical
medium, average, and low data intensities, respectively. Analytic hier­ toxicology, physical damage, biological invasion, and human-social
archy process (AHP) was adopted to identify evaluation weights ac­ stressors. The parameter was assigned values of 4, 3, 2, or 1 to indi­
cording to the evidential importance of adverse outcomes, risk sources, cate different types of stress, which varied in severity. (d) The pathway
and receptor data (Nyimbili et al., 2018; Guo et al., 2020a; Guo et al., analysis of “receptor-endpoint” in different sub-regions represented
2020b). Such as the weights of risk source, stressors, habitats and end­ different effects of biotope damage on the environment, with values of 3,
points were 0.28, 0.24, 0.21 and 0.27, respectively. Among the stressors, 2, 1, or 0. (e) Parameter λ2 was designed for different receptors, such as
the weights of chemical action, physical damage, biological invasion, rivers, lakes, wetlands, forests, grasslands, arable land, and urban and
and humanistic society were 0.32, 0.27, 0.19, 0.22, respectively rural settlements. AHP was employed to determine the weights to reflect

4
K. Guo et al. Ecological Indicators 157 (2023) 111210

the different effects of biotope degradation on the environment. 2.6. Data collection and processing

2.4.2. AOP confidence calculation The research team worked with local governments, businesses,
We calculated the confidence index of AOPs on a regional scale by environmental groups, and other stakeholders to explore the causes of
considering the strength of evidence of “data and causality”. That is, the the serious degradation of land resources and the environment in Daye
confidence coefficient score for each pathway was based on the city. This included those familiar with the regional economic develop­
assumption that the pathway had the “highest level of data for risk ment model and the production and emission status of local polluting
variables” and the “strongest causal chain between variables”. The enterprises. With the help of local governments, multiple field surveys
confidence coefficient ranking of AOPs for all sub-regions was thus were then organised for major land degradation sites. A large amount of
derived with the following formula: data was obtained from the field investigation, social surveys, and from
the Daye Statistical Yearbook, such as data for sewage emission, water
WAOPi = α(risksource)β(λ1 Stressors)γ(λ2 Habitats)(endpoints) (1)
quality, and acid rain (particularly acid rain intensity) released by the
Daye Environmental Protection Bureau, and data on natural disasters
WAOPi is the confidence degree of a certain path i and α, β, γ were pa­
collected from the historical information of Daye (1949–2020). The use
rameters for the causal relationship.
category was based on land use maps of Daye in 2019. The data pro­
TOPSIS is a discrete mufti-criteria decision analysis method that
cessing of variables used in the AOP analysis were listed in Supple­
encompasses a set of techniques to support decision-making processes
mentary materials Table S1, S2, and S3. All of the above processes were
(Guo et al., 2020a; Guo et al., 2020b; Wang et al., 2020). Based on the
performed in ArcGIS 10.2 for follow-up spatial analyses.
TOPSIS combined with the entropy weight method, we calculated the
spatial confidence index of AOPs in each sub-region (Sun et al., 2017;
3. Results
Lin et al., 2022; Wang et al., 2022). The design process of the AOP
confidence calculation was listed in Supplementary materials.
3.1. Determining most likely AOP for sub-regions
Based on this design, the division and ranking of the confidence
degrees of the AOPs can be obtained for all sub-regions, such as areas
Based on the representative risk sources in sub-region 1 and repre­
classified as “very low”, “low”, “moderate”, “high”, and “very high”. In
sentative stress in sub-region 2, we selected hazard paths with high
addition, the spatial visualisation of the results can help decision-makers
confidence levels (Fig. 5). Hazard paths by which urbanization causes a
make appropriate decisions about regional ecological protection and
shortage of land resources in urban and rural settlements through
allow precise management.
“habitat removal and a destroyed earth surface” had confidence levels of
0.639, 0.538, and 0.512 (range of confidence levels is 0–1). The confi­
2.5. AOP-based management decision-making design
dence levels of hazard paths by which industrial sewage emissions act on
the habitats “rivers, urban areas and arable lands” through the stress
The AOP of the land system was identified by quantitative calcula­
“heavy metal accumulation” to cause “water quality deterioration and
tions, uncovering complex ecological mechanisms and illustrating each
public safety and health threats” were 0.591, 0.554, and 0.516. The
stage of system degradation. Therefore, decision-making strategies were
confidence levels of hazard paths by which agricultural intensification
designed for (four) different AOP stages, including a prevention pro­
causes a “decline in land productivity and reduction in eco-resiliency” in
gram, blocking mechanism, adverse outcome reduction/mitigation, and
“arable land, wetlands, and grasslands” through “nutrient runoff,
governance/restoration, described as follows (Fig. 4).
habitat removal, and destroyed earth surface” were 0.747, 0.726, and
At I (risk source), a preparedness plan should aim to contain the risk
0.611.
at the source. At II (stressors), a containment mechanism should be
In sub-region 2, the confidence levels of hazard paths by which the
generalised, and effective solutions for repression should be developed
stressor “destroyed earth surface” causes urban settlements to have
by identifying how stressors affect the “life-non-life feedback mecha­
“public health threats and landscape aesthetics dysfunction” were
nism” of the land system. At III (receptor), a rapid response mechanism
0.857, 0.665, and 0.647. The major risk sources were geological di­
should be adopted to determine the receptor’s response to adverse
sasters and mining. The confidence levels of hazard paths by which
outcomes and the spatiotemporal range to prevent further adverse
heavy metals accumulate due to mining causes “soil contamination,
outcomes. At IV (effect endpoint), a government plan should be
public safety threats and usable land shortage” in “urban settlements
designed to minimise adverse outcomes, and comprehensive area-wide
and arable lands” were 0.853, 0.665, and 0.594. In addition, the confi­
remediation and ecological reconstruction should be emphasised to
dence levels of hazards paths by which geological disasters, mining, and
restore the overall function of the land system. Last, feedback instruction
urbanization disrupt landscape aesthetics and reduce eco-resiliency in
V should be available at each stage to provide flexible and timely
habitats of “urban settlements and grasslands” via desertification were
adjustment plans (Fig. 4).
0.651, 0.634, and 0.582 (Fig. 5).
Representative habitat receptors in sub-region 3 and endpoints in

Fig 4. Design of a management concept based on AOP.

5
K. Guo et al. Ecological Indicators 157 (2023) 111210

Fig. 5. AOP analysis for sub-regions 1 and 2. Each AOP of land degradation all includes risk sources, stressors, habitats, and endpoints. And the AOP’s confidence
scores were presented.

sub-region 4 were showed in Fig. 6, indicating the hazard paths with The ratios of major “risk sources, s tresses, habitats and endpoints” of
high confidence levels. In sub-region 3, the confidence levels of hazard sub-region 5 in respective hazard paths were presented in Fig. 7. Spe­
paths by which stressors “habitat removal, organic pollutants, and water cifically, with respect to risk sources, the ratios of the mining, applica­
eutrophication” caused by the risk sources “lake-area reclamation and tion of pesticides, agriculture intensification, and urbanization in land
aquaculture” contribute to reduction in eco-resiliency and water quality hazard paths were 14.17 %, 13.46 %, 9.67 %, and 8.22 %, respectively.
deterioration in lake habitats were 0.681, 0.654, and 0.663. The confi­ Among multiple stressors, the ratios of habitat removal, behavioral and
dence levels of hazard paths by which urbanization, lake-area recla­ physical disturbances, nutrient runoff, and heavy metals accumulation
mation, and application of pesticides cause “biodiversity decreased and in hazard paths were 14.54 %, 11.63 %, 8.83 %, and 7.85 %, respec­
public safety threats” in wetlands through “nutrient runoff, altered tively. With respect to habitat types, the ratios of forests, grasslands, and
earth-surface runoff, and organic pollutants” were 0.634, 0.597, and arable lands were 28.07 %, 27.09 %, and 25.70 %, respectively. Finally,
0.538. The confidence levels of hazard paths by which grasslands are the ratios of land degradation problems including landscape aesthetics
subject to the stress of “nutrient runoff, destroyed earth surface, and dysfunction, water resource shortage, biodiversity decreased, and soil
organic pollutants” caused by risk sources “agricultural intensification, depletion were 12.17 %, 11.13 %, 10.65 % and 10.15 %, respectively
biological resources consumption and application of pesticides”, leading (Fig. 7).
to “soil depletion, decline in land productivity and public safety threats” The ratios of major “risk sources, stresses, habitats, and endpoints” of
were 0.558, 0.508, and 0.479 (Fig. 6). sub-region 6 in respective hazard paths were presented in Fig. 8. With
In sub-region 4, the confidence levels of hazard paths by which respect to risk sources, the ratios of mining, the application of pesticides,
“sewage emission and solid-waste pile” cause soil contamination in agricultural intensification, and geological disaster in land hazard paths
“urban and rural settlements and arable lands” through the release of the of sub-region 6 were 13.54 %, 13.42 %, 8.90 %, and 7.56 %, respec­
stressor “heavy metal accumulation” were 0.825, 0.727, and 0.695. The tively. Among multiple stressors, the ratios of habitat removal, patho­
confidence levels of hazard paths by which the release of the stress gens, heavy metals accumulation, and behavioral and physical
“leaching and heavy metal accumulation” causes water quality deteri­ disturbances in hazard paths were 11.75 %, 10.75 %, 9.40 %, and 8.24
oration in “lakes, wetlands and rivers” were 0.712, 0.681, and 0.623. %, respectively. For habitats, the ratios of arable lands, grasslands, and
Additionally, the confidence levels of hazard paths by which heavy urban and rural settlements in the hazard paths were 20.61 %, 18.55 %,
metal accumulation resulting from “sewage emission and solid-waste and 17.15 %, respectively. Finally, the ratios of land degradation
pile” causes public health threats in “urban and rural settlements” problems, including landscape aesthetics dysfunction, soil contamina­
were 0.719, 0.668, and 0.575 (Fig. 6). tion, water quality deterioration, and biodiversity decreased were 12.04

Fig. 6. AOP analysis for sub-regions 3 and 4.

6
K. Guo et al. Ecological Indicators 157 (2023) 111210

Fig. 7. The ratios of major “risk sources, stresses, habitats and endpoints” of sub-region 5 in respective hazard paths.

Fig. 8. The ratios of major “risk sources, stresses, habitats and endpoints” of sub-region 6 in respective hazard paths.

%, 1043 %, 10.43 %, and 9.89 %, respectively (Fig. 8). hazard pathways were selected from the sub-regions and their spatial
distributions were presented in Fig. 9.
3.2. Spatial distribution of confidence levels of hazard paths In sub-region 1, the spatial pattern of the confidence levels of the
hazard path “urbanization - habitat removal - arable lands - usable land
The credibility for each pathway was calculated in ArcGIS 10.2. Six shortage” was characterised, as shown in Fig. 9(a). High confidence

7
K. Guo et al. Ecological Indicators 157 (2023) 111210

Fig. 9. Spatial distribution of confidence levels for the representative hazard paths from each sub-region. The AOPs selected as follows: a. “urbanization - habitat
removal - arable lands - usable land shortage”, b. “geological disasters - destroyed earth surface - urban and rural land use - public health threats”, c. “lake area
aquaculture - habitat removal - wetlands - reduction in eco-resiliency”, d. “solid-waste pile - leaching - urban and rural settlements - soil contamination”, e. “bio­
logical resources consumption - habitat removal - forests - biodiversity decreased”, f. “mining - heavy metals accumulation - urban and rural settlements - soil
contamination”.

levels were mainly concentrated around the area of overlap between 4. Discussion
“economic development and industrial aggregation zones” and arable
lands, accounting for 21.8 % of the grids in the sub-region, equivalent to The AOP methodology developed in this study that presents poten­
approximately 58 km2. tial hazard paths towards land degradation on a regional scale. An
In sub-region 2, the spatial pattern of the confidence level of the innovative management perspective can be developed based on targeted
hazard path of “geological disasters - destroyed earth surface - urban and land resource management with the characteristics of AOP.
rural land use - public health threats” was characterised, as shown in
Fig. 9(b). High confidence levels were concentrated in the traditional 4.1. Management decisions based on the AOP analysis
iron mining areas in the central region, where dense mining settlements
were distributed, accounting for 8.4 % of grids in the sub-region, Land resources and environmental degradation in the six sub-regions
equivalent to 20 km2. were analysed according to the identified hazard paths. Management
Sub-region 3 had high confidence levels for the path “lake area countermeasures for all sub-regions were designed in accordance with
aquaculture - habitat removal - wetlands - reduction in eco-resiliency”, the design policy described in 2.5. By the way, we are willing to share
located in traditional lake aquaculture zone and accounting for 11.1 % our findings. For example, we can provide the analysis process and case
of the grids, equivalent to 19 km2, as shown in Fig. 9(c). materials in the form of data sets, attached with URLs. Meanwhile, our
For sub-region 4, the spatial patterns of the confidence levels of the experimental results will be offered to local governments as evidence for
hazard path of “solid-waste pile - leaching - urban and rural settlements - sustainable land resource planning in the future communication.
soil contamination” are shown in Fig. 9(d). High confidence levels were In this paper, a management strategy, involving a prevention plan,
concentrated along the metallurgical processing belt, accounting for 6.4 blocking mechanism, hazard reduction or limitation, and restoration,
% of the grids and equivalent to 12 km2. was designed based on the hazard path analysis results for the sub-
For sub-region 5, the spatial pattern of the confidence levels of the region 6 (Fig. 10).
hazard path “biological resources consumption - habitat removal - for­ For the wastewater discharge from copper mining, the supervision of
ests - biodiversity decreased” is shown in Fig. 9(e). High confidence the implementation of strict discharge standards by regional copper
levels were concentrated in the traditional forest zone of Daye City, enterprises is a prioritised preventive measure. An emergency plan for
accounting for 24.6 % of the grids and equivalent to 140 km2. heavy metal water pollution is critical. In particular, more water quality
In sub-region 6, the spatial pattern of the confidence levels of the monitoring points should be set up in the basin to ensure all-weather
hazard path of “mining - heavy metals accumulation - urban and rural early warning (Fig. 10).
settlements - soil contamination” are shown in Fig. 9(f). High confidence Effects of river pollution caused by copper mining (pollution of
levels were concentrated in the eastern traditional copper mining area downstream water sources) on the safety of downstream habitats should
and the peripheral metallurgical processing belt, accounting for 9.1 % of be avoided. Overall, it is still necessary to treat heavy metal water
the grids and equivalent to 21 km2. pollution in an all-round way from the source to the downstream, and to
establish a complete regulatory system and long-term risk management
to ensure the recovery of water quality (Brankov et al., 2012; Fei et al.,
2017).

8
K. Guo et al. Ecological Indicators 157 (2023) 111210

Fig. 10. Decision-making based on AOP analysis for the sub-region 6.

Regarding protection against the hazard of solid waste stacking, with the loss of substantial water, wetland, and forest habitat (Fig. 10).
effective environmental protection measures should be first taken for The government has attempted to improve the utilisation of urban land
large-scale open-pit copper mine reserves and smelting slag. Soil, lake, by implemented strict control over the proportion of construction land,
and groundwater pollution can be easily caused by surrounding heavy with an emphasis on reasonable urban planning and the optimisation of
metal pollution through long-term rain wash. In particular, preventive industrial structure (Li et al., 2019).
blocking is essential to address pollution caused by the leaching effect.
And in certain circumstances, the reutilization of abandoned mining
waste can help mitigate the environmental impact of waste exposure. 4.2. Implications for land resource and environment management
For example, waste rock and tailings can also be employed as supple­
mentary source materials for the production of construction materials The path analysis concept was used to identify ecosystem hazards,
(Bian et al., 2012). uncover complex underlying mechanisms, and describe the specific
A high copper content in the soil seriously endangers the life and process of land resource and environmental degradation at the regional
health of local residents, the survival of animals and plants, and the scale. Management decisions focus on preventing degradation based on
safety of agricultural products and groundwater quality. However, it is the underlying mechanism, rather than land consolidation and recla­
evident that the complexity and high costs of methods to remediate soil mation projects in the original city area. The proposed method can
polluted with heavy metals make it difficult for managers to design provide a basis for the development and verification of scientific hy­
effective long-term remediation projects. Therefore, the government has potheses. The original intention of this method is to prompt action once
invested in a long-term land remediation plan. the land environment has potential safety hazard. The management
Extensive urban expansion, as a difficult issue in land resource paradigm developed from the perspective of hazard paths can effectively
degradation, is a main cause of the loss of ecological elasticity, together achieve common maintenance by improving the level of understanding
of land stakeholders with different professional backgrounds. In

9
K. Guo et al. Ecological Indicators 157 (2023) 111210

particular, the most effective decisions can be made when decision- interests or personal relationships that could have appeared to influence
making designs are adopted for different management objectives and the work reported in this paper.
stages. Moreover, the proposed method can provide management de­
cisions for the government in a timely and low-cost manner under the Data availability
limited condition of field surveys without large-scale blind census
techniques. The authors do not have permission to share data.
On the whole, most approaches were aimed at seeking revealed the
mechanisms of environmental degradation by integrating prior knowl­ Acknowledgments
edge and data collected in the field, which could help policy makers in
their work (Fay et al., 2017; Moe et al., 2020). Meanwhile, the AOP This research was supported by the National Key R&D Program of
approach the hypothesis regarding the field investigation conducted China (Grant No. 2018YFB2100702), and two National Natural Science
before the experiment was verified, and it reflects the close relationship Foundation of China (Grant No. 42071441; Grant No.42071443). The
between human activities and ecological process (Perkins et al., 2018). authors would like to thank Dr Yiyun Chen (Wuhan University, China)
Moreover, this method process not only ensures the transparency of for offering data related to the land resources of Daye City. We also
policy implementation, but also promotes the establishment of a plat­ thank the Land Resources Bureau of Daye City for their kind help with
form for discussing resources and environment governance in the future the field investigation.
(LaLone et al., 2017). However, the analysis of hazard paths for large-
scale environmental degradation is incomplete, and the results should Appendix A. Supplementary data
be verified. Therefore, the management decisions derived from the AOP
approach are still required to be evaluated and confirmed in practice for Supplementary data to this article can be found online at https://doi.
land-resource management. In particular, with respect to the factors of org/10.1016/j.ecolind.2023.111210.
safety of human life, we should focus our attention on the investigation
of first-line resources and the environment in mining cities according to References
each specific situation in order to implement timely measures, and we
must not be confined to experimental simulations. Alam, M., 2018. Ecological and economic indicators for measuring erosion control
services provided by ecosystems. Ecol. Ind. 95, 695–701.
Ankley, G.T., Bennett, R.S., Erickson, R.J., Hoff, D.J., Hornung, M.W., Johnson, R.D.,
5. Conclusions et al., 2010. Adverse outcome pathways: a conceptual framework to support
ecotoxicology research and risk assessment. Environ. Toxicol. Chem. 29, 730–741.
Bai, X.M., Shi, P.J., Liu, Y.S., 2014. Realizing China’s urban dream. Nature 509,
Using an AOP identification approach, a paradigm for land resource 158–160.
and environmental management at the regional scale was developed. Baron, J.S., Gunderson, L., Allen, C.D., Fleishman, E., McKenzie, D., Meyerson, L.A.,
Based on priori knowledge and field-collected data, the method revealed Oropeza, J., Stephenson, N., 2009. Options for national parks and reserves for
adapting to climate change. Environ Manage 44, 1033–1042. https://doi.org/
the pathways of land system degradation. The hypotheses proposed in 10.1007/s00267-009-9296-6.
the pre-experimental field survey were verified, implying a close tie Becker, R.A., Ankley, G.T., Edwards, S.W., Kennedy, S.W., Linkov, I., Meek, B.,
between human economic activities and ecological processes. The key Sachana, M., Segner, H., Van Der Burg, B., Villeneuve, D.L., Watanabe, H., Barton-
Maclaren, T.S., 2015. Increasing scientific confidence in adverse outcome pathways:
theoretical contribution of this study is the generalisation of adverse
application of tailored bradford-hill considerations for evaluating weight of
outcomes from a complex environment, illuminating the ecological evidence. Regul. Toxicol. Pharm. 72, 514–537.
mechanisms underlying regional land degradation and characterising Bian, Z., Miao, X., Lei, S., Chen, S., Wang, W., Struthers, S., 2012. The challenges of
the evolution of an ecosystem at a large scale. This approach, which reusing mining and mineral-processing wastes. Science 337, 702–703. https://doi.
org/10.1126/science.1224757.
relies on the quantification/simulation of AOPs, is an innovation in land Borrelli, P., Robinson, D.A., Fleischer, L.R., Lugato, E., Ballabio, C., Alewell, C.,
resource management. It can guide management decisions at different Meusburger, K., Modugno, S., Schutt, B., Ferro, V., 2017. An assessment of the global
stages, such as the preparedness plan, containment mechanism, adverse impact of 21st century land use change on soil erosion. Nat. Commun. 8, 2013.
Brandolini, P., Pepe, G., Capolongo, D., Cappadonia, C., Cevasco, A., Conoscenti, C.,
outcome prevention/mitigation, governance/recovery, and reconstruc­ Marsico, A., Vergari, F., Del Monte, M., 2018. Hillslope degradation in representative
tion, and is endorsed by different stakeholders. While ensuring the Italian areas: just soil erosion risk or opportunity for development? Land Degrad.
effective implementation of strategies, this helps shape the future multi- Dev. 29 (9), 3050–3068.
Brankov, J., Milijašević, D., Milanović Pešić, A., 2012. The assessment of the surface
dimensional management of land resources and the environment. water quality using the water pollution index: a case study of the Timok River (The
Furthermore, the approach uses visual simulations to pinpoint locations Danube River Basin). Serbia. Arch Environ Prot 38. https://doi.org/10.2478/
to help policymakers identify land conservation priorities and avoid v10265-012-0004-x.
Brice, E.M., Miller, B.A., Zhang, H., Goldstein, K., Zimmer, S.N., Grosklos, G.J.,
costly blind action. Nevertheless, the study had limitations. It failed to Belmont, P., Flint, C.G., Givens, J.E., Adler, P.B., et al., 2020. Impacts of climate
represent the dynamics of land degradation, and the superposition of change on multiple use management of Bureau of Land Management land in the
multiple pathways may cause uncertainty in simulation results, which is Intermountain West, USA. Ecosphere 11, e3286.
Bridges, J., Sauer, U.G., Buesen, R., Deferme, L., Tollefsen, K.E., Tralau, T.,
expected to be addressed in subsequent studies.
Ravenzwaay, B., Poole, A., Pemberton, M., 2017. Framework for the quantitative
weight-of-evidence analysis of ‘omics data for regulatory purposes. Regul. Toxicol.
CRediT authorship contribution statement Pharm. 91, S46–S60.
Bryan, B.A., Gao, L., Ye, Y.Q., 2018. China’s response to a national land-system
sustainability emergency. Nature 559 (7713), 193–204.
Kai Guo: Conceptualization, Methodology, Resources, Writing – Chakraborti, S., Das, D.N., Mondal, B., Shafizadeh-Moghadam, H., Feng, Y.J., 2018.
original draft. Xiaojin Liang: Methodology, Formal analysis, Data A neural network and landscape metrics to propose a flexible urban growth
curation, Supervision. Xinchang Zhang: Validation, Supervision, boundary: a case study. Ecol. Ind. 93, 952–965.
Chen, R., Ye, C., 2014. Resolving soil pollution in China. Nature 505, 483.
Funding acquisition, Project administration. Renbo Luo: Investigation, Collier, Z.A., Gust, K.A., Gonzalez-Morales, B., Gong, P., Wilbanks, M.S., Linkov, I.,
Resources, Writing – review & editing. Tianqi Qiu: Methodology, Perkins, E.J., 2016. A weight of evidence assessment approach for adverse outcome
Software, Formal analysis, Validation. Zhenhao He: Investigation, pathways. Regul. Toxicol. Pharm. 75, 46–57.
Conolly, R.B., Ankley, G.T., Cheng, W., Mayo, M.L., Miller, D.H., Perkins, E.J.,
Formal analysis, Data curation, Visualization. Kexin Zhang: Formal Villeneuve, D.L., Watanabe, K.H., 2017. Quantitative adverse outcome pathways and
analysis, Visualization, Validation. their application to predictive toxicology. Environ. Sci. Technol. 51 (8), 4661–4672.
Dawson, L., Elbakidze, M., Angelstam, P., Gordon, J., 2017. Governance and
management dynamics of landscape restoration at multiple scales: learning from
Declaration of Competing Interest successful environmental managers in Sweden. J. Environ. Manag. 197, 24–40.

The authors declare that they have no known competing financial

10
K. Guo et al. Ecological Indicators 157 (2023) 111210

Ellenwood, M.S., Dilling, L., Milford, J.B., 2012. Managing United States public lands in Lin, Y.Z., Peng, C., Chen, P., Zhang, M.J., 2022. Conflict or synergy? Analysis of
response to climate change: a view from the ground up. Environ. Manage. 49, economic-social- infrastructure-ecological resilience and their coupling coordination
954–967. https://doi.org/10.1007/s00267-012-9829-2. in the Yangtze River economic Belt. China. 142, 109194.
Fay, K.A., Villeneuve, D.L., Lalone, C.A., Song, Y., Tollefsen, K.E., Ankley, G.T., 2017. Linkov, I., Loney, D., Cormier, S., Satterstrom, F.K., Bridges, T., 2009. Weight-of-
Practical approaches to adverse outcome pathway development and weight of evidence evaluation in environmental assessment: review of qualitative and
evidence evaluation as illustrated by eco-toxicological case studies. Environ. Toxicol. quantitative approaches. Sci. Total Environ. 407, 5199–5205.
Chem. 36 (6), 1429–1449. Liu, Y.L., Jiao, L.M., Liu, Y.F., He, J.H., 2013. A self-adapting fuzzy inference system for
Fei, J., Min, X., Wang, Z., Pang, Z., Liang, Y., Ke, Y., 2017. Health and ecological risk the evaluation of agricultural land. Environ. Model. Softw. 40, 226–234.
assessment of heavy metals pollution in an antimony mining region: a case study Luan, C.X., Liu, R.Z., Peng, S.C., 2021. Land-use suitability assessment for urban
from South China. Environ. Sci. Pollut. R 24, 27573–27586. https://doi.org/ development using a GIS-based soft computing approach: a case study of Ili Valley
10.1007/s11356-017-0310-x. China. Ecol. Indicators. 123, 107333.
Ge, Y., Hu, S., Ren, Z.P., Jia, Y.X., Wang, J.H., Liu, M.X., Zhang, D., Zhao, W.H., Luo, Y. Maciel, E.A., Martins, V.F., Torres, R.R., Martins, F.R., 2023. How do intrinsic and
W., Fu, Y.Y., 2019. Mapping annual land use changes in China’s poverty-stricken extrinsic causes interact in the extinction vulnerability of South American savanna
areas from 2013 to 2018. Remote Sens. Environ. 232, 111285. shrub and tree species? J. Environ. Manage. 343, 118256.
Giubilato, E., Zabeo, A., Critto, A., Giove, S., Bierkens, J., Hond, E.D., Marcomini, A., Mao, D., Wang, Z., Wu, B., Zeng, Y., Luo, L., Zhang, B., 2018. Land degradation and
2014. A risk-based methodology for ranking environmental chemical stressors at the restoration in the arid and semiarid zones of China: Quantified evidence and
regional scale. Environ. Int. 65, 41–53. implications from satellites. Land Degrad. Dev. 29, 3841–3851.
Guo, S.S., Zhang, Y.F., Qian, X.Y., Ming, Z., Nie, R., 2019. Urbanization and CO2 Moe, S.J., Wolf, R., Xie, L., Landis, W.G., Kotamaki, N., Tollefsen, K.E., 2020.
emissions in resource-exhausted cities: Evidence from Xuzhou City, China. Nat. Quantification of an adverse outcome pathway network by bayesian regression and
Hazards 99, 807–826. bayesian network modeling. Integr. Environ. Assess. Manag. 17 (1), 147–164.
Guo, K., Zhang, X.C., Kuai, X., Wu, Z.F., Chen, Y.Y., Liu, Y., 2020a. A spatial bayesian- Nyimbili, P.H., Erden, T., Karaman, H., 2018. Integration of GIS, AHP and TOPSIS for
network approach as a decision-making tool for ecological-risk prevention in land earthquake hazard analysis. Nat. Hazards. 92, 1523–1546.
ecosystems. Ecol. Model. 419, 108929. Perkins, E.J., Gayen, K., Shoemaker, J.E., Antczak, P., Burgoon, L., Falciani, F.,
Guo, K., Zhang, X., Liu, J., Wu, Z., Chen, M., Zhang, K., Chen, Y., 2020b. Establishment of Gutsell, S., Hodges, G., Kienzler, A., Knapen, D., McBride, M., Willett, C., Doyle III, F.
an integrated decision-making method for planning the ecological restoration of J., Garcia-Reyero, N., 2018. Chemical hazard prediction and hypothesis testing using
terrestrial ecosystems. Sci. Total Environ. 741, 139852. quantitative adverse outcome pathways. ALTEX 36 (1), 91–102. https://doi.org/
Guo, K., Chen, Y.Y., Chen, M., Wang, C.J., Chen, Z.Y., Cai, W.N., Li, R.J., Feng, W.M., 10.14573/altex.1808241.
Jiang, M., 2021. Causal analysis of ecological impairment in land ecosystem on a Platt, J.R., 1964. Strong Inference: Certain systematic methods of scientific thinking may
regional scale: applied to a mining city daye China. Land. 10, 530. produce much more rapid progress than others. Science 146, 347–353.
Gust, K.A., Collier, Z.A., Mayo, M.L., Stanley, J.K., Gong, P., Chappell, M.A., 2015. Ristic, V., Maksin, M., Nenkovic-Riznic, M., Basaric, J., 2018. Land-use evaluation for
Limitations of toxicity characterization in life cycle assessment: can adverse outcome sustainable construction in a protected area: a case of Sara mountain national park.
pathways provide a new foundation? Integr. Environ. Assess. Manag. 12 (3), J. Environ. Manag. 206, 430–445.
580–590. Schaefer, M., Thinh, N.X., 2019. Evaluation of land cover change and agricultural
Heckera, M., LaLone, C.A., 2019. Adverse outcome pathways: moving from a scientific protection sites: a GIS and Remote Sensing Approach for Ho Chi Minh City Vietnam.
concept to an internationally accepted framework. Environ. Toxicol. Chem. 38 (6), Heliyon 5, 1773.
1152–1163. Su, K., Yu, Q., Yue, D.P., Zhang, Q.B., Yang, L., Liu, Z.L., Niu, T., Sun, X.T., 2019.
Hou, D.Y., 2022. Expediting climate-smart soils management. Soil Use Manag. 38, 1. Simulation of a forest-grass ecological network in a typical desert oasis based on
Hou, D., Ok, Y.S., 2019. Speed up mapping of soil pollution. Nature 566, 455. multiple scenes. Ecol. Model. 413, 108834.
Joyce, L.A., Blate, G.M., McNulty, S.G., Millar, C.I., Moser, S., Neilson, R.P., Peterson, D. Sun, L.Y., Miao, C.L., Yang, L., 2017. Ecological-economic efficiency evaluation of green
L., 2009. Managing for multiple resources under climate change: national forests. technology innovation in strategic emerging industries based on entropy weighted
Environ. Manage 44, 1022–1032. https://doi.org/10.1007/s00267-009-9324-6. TOPSIS method. Ecol. Indic. 73, 554–558.
Kleinstreuer, N.C., Sullivan, K., Allen, D., Edwards, S., Mendrick, D.L., Embry, M., Suter II, G.W., 1993. Ecological Risk Assessment. Lewis Publishers, Chelsea, MI, USA.
Matheson, J., Rowlands, J.C., Munn, S., Maull, E., Casey, W., 2016. Adverse outcome Suter II, G.W., Cormier, S.M., 2011. Why and how to combine evidence in environmental
pathways: From research to regulation scientific workshop report. Regul. Toxicol. assessments: Weighing evidence and building cases. Sci. Total Environ. 409,
Pharmacol. 76, 39–50. 1406–1417.
Kramer, V.J., Etterson, M.A., Hecker, M., Murphy, C.A., Roesijadi, G., Spade, D.J., Turner, K.G., Anderson, S., Gonzales-Chang, M., Costanza, R., et al., 2016. A review of
Spromberg, J.A., Wang, M., Ankley, G.T., 2011. Adverse outcome pathways and methods, data, and models to assess changes in the value of ecosystem services from
ecological risk assessment: Bridging to population-level effects. Environ. Toxicol. land degradation and restoration. Ecol. Modell. 319, 190–207.
Chem. 30, 64–76. USEPA, 2010. ProUCL version 4.00.05 Technical Guide. EPA/600/R-07/041. U.S.
LaLone, C.A., Villeneuve, D.L., Wu-Smart, J., Milsk, R.Y., Sappington, K., Garber, K.V., Environmental Protection Agency, Office of Research and Development, Washington
Housenger, J., Ankley, G.T., 2017. Weight of evidence evaluation of a network of DC, USA.
adverse outcome pathways linking activation of the nicotinic acetylcholine receptor Wang, Y.X., Yu, X.H., Zhao, B.J., Xiong, X., Li, Y., Zhang, M., 2022. Evaluation of
in honey bees to colony death. Sci. Total Environ. 584–585, 751–775. ecological carrying capacity in Yangtze River Economic Belt and analysis of its
Laurance, W.F., Andrade, A.S., Magrach, A., Camargo, J.L.C., Campbell, M., Fearnside, P. spatial pattern evolution. Ecol. Ind. 144, 109535.
M., Edwards, W., Valsko, J.J., Lovejoy, T.E., Laurance, S.G., 2014. Apparent Zhang, X.C., Chen, M., Guo, K., Liu, Y., Liu, Y., Cai, W.N., Wu, H., Chen, Z.Y., Chen, Y.Y.,
environmental synergism drives the dynamics of Amazonian forest fragments. Zhang, J.G., 2021. Regional land eco-security evaluation for the mining city of daye
Ecology 95, 3018–3026. in China using the GIS-based grey TOPSIS method. Land. 10, 118.
Lee, J.W., Won, E.J., Raisuddin, S., Lee, J.S., 2015. Significance of adverse outcome Zhang, M.M., Tan, S.K., Pan, Z.C., Hao, D.Q., Zhang, X.S., Chen, Z.H., 2022. The spatial
pathways in biomarker-based environmental risk assessment in aquatic organisms. spillover effect and nonlinear relationship analysis between land resource
J. Environ. Sci. 35, 115–127. misallocation and environmental pollution: evidence from China. J. Environ.
Leonarda, J.A., Nelmsa, M., Craig, E., Perron, M., Pope-Varsalonab, H., Dobreniecki, S., Manage. 321, 115873.
Lowit, A., Tan, Y.M., 2019. A weight of evidence approach to investigate potential Zhang, H.M., Xiong, L.F., Li, L.S., Zhang, S.F., 2018. Political incentives, transformation
common mechanisms in pesticide groups to support cumulative risk assessment: a efficiency and resource-exhausted cities. J. Clean. Prod. 196, 1418–1428.
case study with dinitroaniline pesticides. Regul. Toxicol. Pharm. 107, 104419. Zhang, Y.Y., Yang, Y.J., Chen, Z.X., Zhang, S.L., 2020. Multi-criteria assessment of the
Li, Z., Yuan, M., Hu, M., Wang, Y., Xia, B., 2019. Evaluation of ecological security and resilience of ecological function areas in China with a focus on ecological
influencing factors analysis based on robustness analysis and the BP-DEMALTE restoration. Ecol. Ind. 119, 106862.
model: a case study of the Pearl River Delta urban agglomeration. Ecol. Indic. 101,
595–602. https://doi.org/10.1016/j.ecolind.2019.01.067.

11