Using state and trend analysis to assess ecological security for the vulnerable agricultural ecosystems of Pengyang County in the loess hilly region of China

Abstract

In this study, the pressure–state–response (PSR) framework and analytic hierarchy process (AHP) approaches were integrated for index selection and determination of associated weights among ecological, social, or economic indices. Comprehensive methods, comprising a fuzzy model, state evaluation, and trend evaluation were developed to assess the ecological security of specific agricultural ecosystems. The fuzzy model was used to rank ecological security by linking the membership function and a set of four levels, namely insecure (I), lightly secure (LS), moderately secure (MS), and secure (S). State and trend evaluation approaches were used to assess the ecological security of an agricultural ecosystem in Pengyang County in the southern Ningxia Autonomous Region, China. The results show that ecological security in Pengyang County was insecure (I) from 1986 to 1995 and lightly secure (LS) from 2000 to 2004. Ecological security has generally improved recently, and the constraining factors for the present ecological security are soil erosion, water shortage, excess population growth, low forest cover, and irrational application of chemical fertilizers. Finally, advice for decision makers is proposed in order to achieve sustainability.

Keywords:

• ecological security
• state and trend evaluation
• PSR framework
• AHP approach

Introduction

Eco-environmental security has received considerable attention from the international community. The concept of ecological security was proposed by the International Institute for Applied Systems Analysis (IIASA) in 1989, and is focused on the security of semi-natural or natural ecosystems, i.e., the integrity and health of a specific ecosystem (Huang et al. 2007). The term ‘environmental security’ was proposed for evaluation of threats to peace and security that are associated with environmental issues, and to develop management strategies to maintain life-support systems, production capacity, and the potential of ecological systems (Dale et al. 2004). Both concepts focus on interactions between economic development and the ecological environment, and both are based on the presence of a sound ecosystem, on sustainable exploitation of natural resources, and on harmony between humans and nature. Eco-environmental security provides a means to respond effectively to changing environmental conditions, which may affect political, economic, social, or environmental stability of a country.

Agricultural ecosystems, as human–nature interactive systems, are profoundly influenced and transformed by the cumulative impacts of human activities, especially in fragile ecosystems, such as the loess hilly region in China (Rapport and Singh 2006). The Loess Plateau has suffered serious soil erosion over the last 2000 years because of increasing human activity and intensive grain production. Environmental problems in the region include: (a) soil erosion, water shortage, and low land productivity; (b) increasing demand and loss of arable land resources due to rapid population growth and high economic activity, especially in minority areas where the population growth rate is about 1.5%; and (c) complex, fragmented and steep terrain leading to difficulties in land exploitation (Chen et al. 2001). Such environmental issues create pressures on the human-dominated ecosystem and cause degradation of the agricultural ecosystem.

The pressure–state–response (PSR) framework is based on the notion of causality and is a logical and comprehensive tool to examine environmental issues from an anthropocentric perspective (OECD 1993, 1996, 1999; Hughey et al. 2004). It can reveal linear relationships between humans and the environment and present information to end-users in a causal fashion by differentiating causes, effects, and human responses required to control the extent of anthropogenic impacts on nature (Wolfslehner and Vacik in press). We therefore used the PSR system to examine Pengyang County in the southern Ningxia Autonomous Region of China, and constructed an approach that integrates the PSR framework with the Analytic Hierarchy Process (AHP) to include ecological, social, and economic indices. Our aim was to establish a regional system to assess ecological security and predict development trends quantitatively and effectively, using methods such as the fuzzy mathematics approach, state evaluation, and trend evaluation, in order to address deteriorating ecological security. We hope that the results from this case study will be helpful to policy-makers in the assessment and sustainability of ecological security.

Study area

The study area, Pengyang County, is a typical county with a semi-arid climate and hilly loess landscape on the Loess Plateau (Figure 1). It is situated between 35°41′–36°17′N and 106°32′–106°58′E, and covers an area of 2528.7 km². The climate shows clear seasonal variations. The mean annual precipitation is about 520 mm, ranging from 350 to 550 mm. The mean annual temperature is 7.2°C, and the frost-free duration is 170 days per year. There are floods in summer and droughts in other seasons. The land surfaces, mostly at 1248–2483 m asl, are highly dissected by deeply incised gullies. Severe ecological issues, such as soil erosion and soil and water loss, are threatening rural ecosystems and constrain crop-pastoral activities at local and landscape scales. To realize ecological agriculture targets, some progress has been made in the last 20 years in soil and water conservation by using small catchments as the base unit, and the implementation of the national Grain for Green project. Environmental pressure is thus gradually decreasing.

Figure 1. Location and terrain of the study area.

Methods

Development of an index system for assessing ecological security

The index system was built on the basis of the well-known PSR structure. It included three boxes that are functionally linked and represent Pressure, State, and Response. Human-induced Pressures influence State parameters by creating functional and economic benefits, and further affect the (societal) Response variables. In contrast, Pressures use and exhaust resources both qualitatively and quantitatively (State), and task input, decisions, and actions from Responses to restrain them. State and Response are linked by exchange of information and data on one side, and by direct state-changing measures on the other side (Figure 2).

Figure 2. Adapted pressure–state–response framework for regional ecosystem assessment of ecological security.

The objective of studying an agricultural ecosystem is to describe the state of the ecosystems that compose it. According to the PSR framework, the ultimate objective can be broken down into several levels by using the AHP method of multi-objective decision-making. The first level is called the ‘target level’, for which we chose the pressures that result from nature and humans (A₁), the ecosystem status (A₂), and the responses of human society (A₃). The second level (B) is called the ‘factor level’, which includes the main factors that explain the upper-level indices, and can also be deemed a sub-target layer. The third level (C) is called the ‘index level’ and includes all indices that can be measured directly and that serve as the foundation for ecological security evaluation of Pengyang County.

Index selection

This paper follows the principles of integrity, simplicity, dynamic response, geographical accuracy, and data availability in index selection. We selected three categories of factors that are useful in assessing ecological security (environmental pressure, environmental state, and social response) and that comprehensively reflect the characteristics and situations of the regional ecosystem:

• Indices of environmental Pressure describe pressures on the environment originating from human activities, including pressures from population and environment (e.g., population density and growth rate, arable land area, fertiliser input, natural disasters).

• Indices of environmental condition (State) were designed to describe the status of the environment and the quality and quantity of land resources (e.g., soil quality, land-use structure).

• Indices of societal Response show the degree to which society is responding to environmental changes and concerns. This could be the number and form of measures taken, efforts for implementing such measures, or the effectiveness of the measures.

A set of 16 indicators for land use was elaborated and compared, including economic, ecological, and social indicators (Table 1).

The index system and thresholds for land ecological security in the loess hilly region

Objective level (A)	Factor level (B)	Indicator level (C)	Weight (w)	Threshold value (Z)
Ecosystem pressure A1	Population B1	Population density (/km^2) C1	0.042	100
		Population growth rate (‰) C2	0.032	6
		Arable land per capita (hm^2per capita) C3	0.037	0.3
	Environment B2	Chemical fertilizer per unit arable land area (kg/hm^2) C4	0.010	300
		Pesticide per unit arable land (kg/hm^2) C5	0.010	30
		Topology C6	0.037	0.3
		Percentage of natural disaster area (%) C7	0.037	10
Ecosystem state A2	Soil quality B3	Grain production per unit area (kg/hm^2) C8	0.051	2500
		Soil erosion modulus (t/km^2 a) C9	0.109	1000
		Percentage area for soil and water loss (%) C10	0.097	15
	Land-use structure B4	Forest cover rate (%) C11	0.108	39
		Percentage area for slope cropland over 25° (%) C12	0.090	1
Human response A3	Action to reduce effect, Improvement B5	Effective irrigated area rate (%) C13	0.085	30
		Invest in environmental management to GDP (%) C14	0.102	1.4
		Income of tertiary sectors to GDP (%) C15	0.068	30
		Net income per capita (Yuan) C16	0.085	2000

Determination of the weights of the indices

We used the AHP approach to decompose the factors that affect the ecosystem security of Pengyang County. According to relevant literature, we constructed a judgment matrix using the square root method (Liu 2004). We compared the relative importance among indicators and determined the weight of each indicator (Table 1).

Determination of threshold value of indices

We hypothesized that there may be a threshold value for each indicator at which ecological security starts to become assured. For instance, the soil erosion modulus is an important index to indicate soil and water loss. The threshold value is regarded as the largest permitted amount of soil loss. If the value is exceeded, the ecosystem in this area cannot be regarded as secure for soil and water loss. The threshold values were determined previously (Zhang et al. 2002; Liu et al. 2004; Gao and Han 2005; Tang and Zhu 2006) based on the following principles:

1. National or industrial standards. The standards for each indicator usually differ among regions. The national standards refer to relevant national environmental quality standards, e.g., the standard for safe use of agricultural chemicals (GB4285-89). Industrial standards are the regulations and rules issued by the environmental sectors. We also refer to the regulations issued by local governments to facilitate soil and water conservation.

2. Empirical judgment based on the literature. These threshold values were determined by citing existing results and conclusions from publications. For instance, although the largest permitted soil loss differs among areas, investigations in the loess hilly region revealed that the value is 1000 t/(km² a) (Wu and Zhao 1999). Consequently, we chose 1000 t/(km² a) as the threshold value for the soil erosion modulus (C₉) in our study area. Considering the integrative functions like economic, ecological, and social functions, we used 39% as the threshold value for forest cover, instead of the mean value 16.6% for the whole country given by Wu and Yang (1998). This value represents the least area required to maintain regional ecological equilibrium.

3. Environment background values were used in specific regions where standards or empirical judgments were not available.

Measures of ecological security

Ecological security State evaluation

1. Measures of single indicator ecological security. A comprehensive evaluation method was employed to determine the single indicator ecological security index. This was based on contextual principals, normalization of values of all indices in level C [0, 1] from different units and types. The single indicator ecological security index employed was expressed as follows:

   (1)
   (2)
   where x_i is the actual value of indicator i; z_i denotes the threshold value of indicator i; and r_i denotes the single indicator ecological security index of indicator i. Larger indices representing greater safety (x_i ) were chosen as a positive index, while indices that refer to lower safety (x_i ) are associated with a negative index.

2. Measure of comprehensive multi-indicator ecological security.

   (3)
   Where S is the multi-indicator ecological security index; r_i is the single indicator ecological security index; and w_i is the weight of indicator i.

3. Determination of ecological security. Based on the ambiguous and uncertain quantified attributes of security, the fuzzy method was employed to determine ecological security of agricultural ecosystems (Verbruggen and Zimmermann 1999; Shi et al. 2006). Four ecological security levels were designated: secure (S), moderately secure (MS), lightly secure (LS), and insecure (I). The membership function coinciding with the relevant levels was expressed in the following equations:

   (4)
   (5)
   (6)
   (7)
   where u is the ecological security index, either a single indicator ecological security index (r_i ) or a multi-indicator ecological security index (S), and F denotes membership coinciding with the given security level. Consequently, we can identify the ecological security levels for single-indicator or multi-indicator ecological security indices. The ecological security level of any indicator is determined by the largest membership value from Equations (4(4)) (5) (6) to (7(7)) for the corresponding membership. For instance, supposing the ecological security index of indicator 1 satisfies the equation r ₁ = 0.3, we can then determine the indicator at the level of insecure (I). In the same way, supposing the multi-indicator ecological indictor satisfies S(t) = 0.55, we can determine that it is at the level of lightly secure (LS).

Measuring the dynamic comprehensive trend evaluation of ecological security

The ecological security state of agricultural ecosystems was characterized based on their provisionality and dynamics. In this study, mathematical methods were used to fit the trend of security state changes. Suppose the temporal resolution is Δt, the change in ecological security is S(t), and the change in ecological security state in an interval of Δt is measured by the slope of change curve as K(Δt). Consequently, K(Δt) plays an important role in evaluating the security trend of the ecosystem, which reflects a change in the degree of ecological security in a given interval Δt. K(Δt) > 0 means the state of a given ecosystem advances toward a high level of ecological security; if K(Δt) < 0, there is a negative trend of change in ecological security.

This study used an integrative method that was static at certain times and also included dynamic changes in ecological security within the ecosystem. As discussed above, S(t) represents a state variable, K(Δt) denotes a trend variable; bi-variate analysis [S(t), K(Δt)] was then employed to assess the comprehensive ecological security of an ecosystem where:

1. If S(t) is at the level of secure (S) or moderately secure (MS), and K(Δt) satisfies K(Δt) > 0, then the given ecosystem is determined as being secure (S).

2. If S(t) is at the level of secure (S) or moderately secure (MS), and K(Δt) satisfies K(Δt) < 0, then the given ecosystem is determined as being insecure (I).

3. If S(t) is at the level of insecure (I) or lightly secure (LS), and K(Δt) satisfies K(Δt)> 0, then the given ecosystem is determined as being insecure and to be changing positively.

4. if S(t) is at the level of insecure (I) or lightly secure (LS), and K(Δt) satisfies K(Δt) < 0, then the given ecosystem is determined as being insecure and to be changing negatively.

Data processing

Survey and statistical data for 1986, 1990, 1995, 2000, and 2004 from the Ningxia Autonomous Statistical Bureau were employed to assess the state and trend of ecological security in Pengyang County. First, calculations according to Equations (1)(1), ( 2(2)), and ( 3(3)) were conducted to assess the ecological security state and to produce a multi-indicator ecological security index (S) for 1986, 1990, 1995, and 2004, namely S ₁₉₈₆ = 0.273, S ₁₉₉₀ = 0.353, S ₁₉₉₅ = 0.387, S ₂₀₀₀ = 0.434, S ₂₀₀₄ = 0.573. Second, the associated levels of ecological security were determined by the largest membership value after the multi-indicator ecological security indices were input into Equations (4)(4) (5) (6)– (7)(7) and calculated. Third, the fitted trend curve and associated function were then obtained (Figure 3) by means of the trend evaluation approach described above. Finally, to assess the current ecological security situation, we obtained the ecological security indices and security level of each single indicator for 2004 (Figure 4).

Figure 3. State and trend of regional land eco-security.

Figure 4. Status of eco-security indices in 2004.

Results

State and trend of ecological security from 1986 to 2004

According to the processed data for ecological security state evaluation (Figure 3), multi-indicator ecological security indices (S) were at the insecure level (I) in 1986, 1990, and 1995, and the lightly secure (LS) level in 2000 and 2004. Trend analysis results presented K(Δt) > 0. Thus, we conclude that the ecological security of Pengyang County is low, but improved from 1986 to 2004.

The low ecological security of the ecosystem from the 1980s to 1990s is attributed to overexploitation. Wood harvesting and other activities during the ‘Cultural Revolution’ significantly degraded the environment in this region during the late 1960s. After the second land reform in 1981, land policy change contributed to a steady increase in the wooded area. At that time, most trees were assigned to farmers, and a new policy, ‘who plants, owns’, was adopted. From the 1980s to the 1990s, soil and water loss and low vegetation cover became the primary eco-environmental issues and, consequently, land productivity was low, together with a lack of available resources and poor quality of life for residents. Since the end of the 1990s, soil and water conservation programs have been carried out continuously, with 83 small watersheds used as the base unit in the whole county. Hence, rural eco-construction has been progressing toward sustainability. After the initiation of the Grain for Green project in 2000, soil and water loss was efficiently controlled and the land-use pattern was optimized to ensure a higher vegetation cover ratio, which increased land productivity and quality of life. Although the agricultural ecosystem was still at an insecure level until 2004, due to the historical low level of the basic state, the trend toward a higher level of ecological security is increasing.

Main eco-environmental issues and constraints

Assessment of the single indicator ecological security index plays an important role in discovering the main environmental constraints for ecological security. The main ecological issues and constraints were explicitly reflected in the processed single indicator ecological security indices and associated membership gradient in 2004 (Figure 4). The soil erosion modulus (C₉), percentage area of soil and water loss (C₁₀), and effective irrigated area rate (C₁₃) were insecure (I), which indicates that soil and water loss and drought are the primary constraints. At the same time, the population growth rate (C₂) and chemical fertilizer use per unit arable land area (C₄) were at low levels, which indicate that the main environmental stressor resulted from population growth and overexploitation of land. Furthermore, the forest cover rate (C₁₁) was also at the level of insecure (I). The above results show that the main constraining factors were population growth, soil and water loss, lack of water, low forest cover, and irrational use of chemical fertilizers. These emerging issues need to be solved to achieve ecological security.

Measures to improve ecological security

Decrease population pressure and enhance environmental regulations

Restraining human activities is the most basic means of preventing further eco-environmental degradation in Pengyang County. An ‘ecological relocation program’ should be implemented by supplying better places for residents to settle in order to mitigate the pressure from population growth. The government should also implement other relevant laws or regulations to promote the construction and protection of the ecological environment.

Strengthen control of soil erosion

Soil and water loss has been a prime constraint for a long time. Although soil and water losses have been controlled to some extent through soil and water conservation programs, the soil erosion modulus (C₉), percentage of soil and water area (C₁₀), and effective irrigation rate (C₁₃) are still at the level of insecure. Thus, control of soil erosion must be a continuous long-term goal. Some effective measures, such as water-saving irrigation systems and planting drought-tolerant species, should be adopted to enhance soil and water conservation.

Expand forest and grassland area and improve management

Increasing vegetation cover is key to conserve soil and water and improve the state of ecological security. Forest planting should be appropriate to the geographical location. Effective maintenance and sustainability of the Grain for Green project is important for forest and grassland management.

Advocate appropriate fertilization in place of over-use of chemical fertilizers

Using an appropriate fertilizer makes land more productive, but farmers have tended to use chemical fertilizers. Excess chemical fertilizer use has led to deterioration of soils and the vulnerability of the whole agricultural ecosystem in Pengyang County. Therefore, inducing farmers to use suitable fertilizers is important to maintain and promote ecological security of the region.

Discussion and conclusions

Integrative evaluation approaches were used to investigate the ecological security of vulnerable agricultural ecosystems in Pengyang County. The results indicate that: (a) ecological security in Pengyang County was insecure from 1986 to 1995 and lightly secure from 2000 to 2004, which suggests that the present ecological security state is still low; (b) ecological security has improved in recent years; and (c) the constraining factors for ecological security are soil and water loss, water shortage, excess population growth, chemical fertilizer application, and low forest cover. These results may help decision-makers to better understand the status and dynamics of ecological security in the study area. Outcomes of this analysis will be useful to ensure future sustainability of ecological security.

This study combined the PSR framework with AHP analyses to build an index system to assess the ecological security of agricultural ecosystems in the loess hilly region. The methods of index selection and threshold value determination could be further improved by incorporating more data and quantification. This study, combining state and trend evaluation, has provided a comprehensive dynamic assessment approach for regional ecological security and decision-making. The approaches used objectively represented the ecological security state and revealed changes in trends of ecological security and causal relations among environment pressures, the state of the environment–human interface, and the human response. The results from this case study in Pengyang County comprehensively reflect the true situation of ecological security in the region. We suggest that combining remote sensing and geographic information systems to establish a spatial-temporal model will further improve assessment and monitoring of agricultural ecosystems for ecological security.

Acknowledgement

This project was supported by the National Natural Science Foundation of China (Contract No. 40571091).