How can agricultural non-point source pollution be controlled? Interdisciplinary co-governance is key
As the level of industrial and urban point source pollution control gradually improves, agricultural non-point source pollution has increasingly become the main contributing source of water environment pollution. Academic research on the control of agricultural non-point source pollution is mostly concentrated in the natural science field, generally showing the characteristics of emphasizing monitoring, accounting, and engineering measures, while neglecting socioeconomic driving factors and weakening policy design. But these issues, which are less explored in the natural science field, are precisely the main direction of social science research. How to give play to the different advantages of natural science and social science research methods in the study of agricultural non-point source pollution control, and achieve full-process analysis from driving factors to emission accounting, transport accounting, environmental impact analysis, and all the way to the effectiveness of policy intervention, is a quite challenging academic proposition.
Nutrient pollution discharge is key
At present, countries have many types of policies for controlling agricultural non-point source pollution. From the perspective of intervention methods and implementing subjects, they can mainly be divided into three categories: command-and-control, economic market-based, and voluntary agreement-based, among which the first two are dominant. Command-and-control instruments are mainly based on laws, regulations, and standards, and are promoted and implemented from the top down by the government through administrative orders. China's agricultural policy is dominated by growth and income increase, and there are not many laws and regulations on agricultural environmental management. Agricultural environmental industry norms and technologies often lack strong operability, and there are many non-mandatory standards, resulting in weak enforcement. Chinese farming households are small in scale, large in number, and scattered in distribution, making policy publicity, implementation, and supervision difficult. Therefore, the management cost of adopting command-and-control policies is too high. Because agricultural non-point source pollution is generated in the process of agricultural production and daily life, it has the characteristics of a wide pollution source area, high dispersion, difficulty in pipe network collection, and long distance from centralized treatment facilities. Apart from pollution sources from large-scale livestock farming and concentrated residential areas, it is very difficult to directly adopt engineering measures for treatment. Therefore, in China, using economic market-based instruments to guide farming households to consciously adopt environmentally friendly production methods based on economic rationality is more feasible.
From the perspective of non-point source pollutant types, calculation data from most river basins in China show that, according to equivalent standard pollution load accounting, among the three most representative agricultural non-point source pollutants - total nitrogen and total phosphorus, which characterize nutrient pollution, and chemical oxygen demand, which characterizes reducing organic matter pollution - nutrient pollution load accounts for the vast majority (more than 90%). Scientifically measuring and controlling nutrient pollution discharge is a key link that should be broken through in current research on agricultural non-point source pollution in China. Considering both policy applicability and pollutant characteristics, we urgently need a methodological platform that can effectively analyze the effectiveness of economic market-based control policies for nutrient pollution.
NEWS in natural science and MAREM in social science
Research on non-point source pollution in the natural science field began in the 1960s and has successively gone through stages of development including empirical statistical models, mechanistic models, and integrated models combined with GIS. At present, the Global NEWS (Nutrient Export from WaterSheds) model is a relatively novel international method used to simulate nutrient loads and export fluxes in river basins. NEWS can effectively describe the processes of nutrient loss, convergence, migration and transformation, retention, and export (to the sea), and analyze their environmental impacts.
In contrast, research in the social science field does not concern itself with the physical and chemical natural processes of pollutant transport, but instead focuses on discussing the root causes, namely how socioeconomic processes cause non-point source pollutant emissions and how to design policy measures to regulate such emissions. However, the farming household models commonly used to solve this problem have weaknesses such as difficulty in representing bounded rationality, lack of heterogeneity, and lack of support from historical environmental data. The MAREM model is a rural environmental management model developed based on experimental economics theory and using multi-agent modeling technology. By realizing "artificial villages" and "virtual farming households" in computers, as well as multiple types of agents such as government, markets, and enterprises, it simulates the responses of farming household groups to different agricultural policies under bounded rationality, as well as agricultural pollutant emissions.
Comparing the modeling approaches of NEWS and MAREM involving agricultural non-point source pollution, the two have very different emphases, but there are indeed areas where they can complement each other. From the generation of agricultural non-point source pollutants to the environmental consequences ultimately caused, scholars in the social science and natural science fields are respectively devoted to research on the first half and the second half of the process. A simple example: if we find that the water environment quality of a certain river basin urgently needs improvement and propose an environmental control target, then the solution given by the NEWS model may be a 30% reduction in chemical fertilizer application in the river basin. But how to achieve the reduction in chemical fertilizer application is something the NEWS model does not care about and cannot analyze. This requires the MAREM model to conduct upfront analysis and tell us what kind of policy design and implementation intensity can achieve this control target. Conversely, economic research methods are likewise hard-pressed to tell us how river basin environmental quality will actually change under different emission levels.
Full-process analysis of NEWS and MAREM
Can the two apparently very different methods, NEWS and MAREM, be effectively integrated to achieve full-process analysis of agricultural non-point source pollution (especially nutrient pollution)?
In the integrated application of the two types of models, one of the most difficult gaps to overcome is that their research scales are completely different. In terms of spatial scale, NEWS is a macro-scale model at the river basin level, while MAREM is a micro-scale model at the village level. In terms of pollution source scope, NEWS also includes industrial and urban point source pollution, while MAREM only discusses agricultural non-point source emissions, making data matching difficult. Therefore, the direction of combining the two must inevitably be a soft connection of input and output data, rather than integration into a unified model.
One simple idea for matching data is to expand the research scope of MAREM so that it has certain urban analysis functions. When the "artificial village" of MAREM is regarded as a grid cell, if multiple "artificial villages" are simultaneously realized and towns are designed surrounded by villages, then not only can urban-rural interaction analysis be achieved, but villages can also be grouped into multiple blocks with different planting and breeding characteristics. However, this expansion method is difficult to operate, and although it achieves unification of the pollution source scope, it cannot achieve matching of spatial scale. In fact, according to the principle that multi-agent models should not seek to completely reproduce the real world, scale expansion should not be relied upon to achieve the leap from individual behavior analysis to river basin spatial scale.
Another idea is to build a bridge connecting the two methods. System dynamics may be a good choice for this task. System dynamics models are called policy simulation laboratories. This model is composed of feedback loops among different elements, is extremely easy to expand, and the application of table functions enables it to better accept input data from other external models. We only need to establish a river basin-scale system dynamics model containing subsystems such as agriculture, industry, and urban areas, and refine the agricultural part according to the design approach of MAREM. The system dynamics model accepts MAREM's simulation results for various agricultural policies and provides river basin-scale output data to NEWS. This not only realizes the expansion of MAREM's research scale, but also overcomes the shortcoming that system dynamics models find it difficult to simulate new policies lacking relevant historical data.
The economic and ecological symbiosis of farming household behavior
Once the methodological problems are solved, return to policy design itself. Once environmental quality and control requirements have been clarified, what kind of policy should be selected to achieve this target, and what type of policy is MAREM more likely to simulate?
As mentioned above, economic market-based policies may better meet the requirements of China's agricultural development and agricultural non-point source pollution control. The basic premise of this type of policy design is that since the root cause of environmental problems is market failure, introducing market mechanisms or similar mechanisms for economic incentives can correct the deviation. This type of policy can be further divided into Coasean regulation and Pigouvian regulation. Pigouvian regulation mainly punishes negative externality behavior or rewards positive externality behavior through taxes, fees, and subsidies, thereby internalizing externalities. Coasean regulation, by clarifying property rights, incentivizes the emergence of emission trading to achieve the internalization of externalities.
Comparing the two types of economic market-based instruments, Coasean regulation requires turning emission rights into property rights and implementing quota trading, and is used more in industrial point sources. But in agricultural non-point sources, the group faced is too large and social relations are complex, so clarifying farming households' emission rights is very difficult to operationalize. Relatively speaking, the taxes, fees, and subsidies used in Pigouvian regulation are more practical. Among them, policies of taxing polluting agricultural production factors and granting subsidies to farming households that adopt environmentally friendly production methods have already been implemented in European countries. Moreover, taxes and subsidies are often used in combination, and the funds for subsidies can come from corresponding tax and fee revenues.
Agricultural non-point source pollution control is an issue of great practicality and urgency in China. Achieving the economic and ecological symbiosis of farming household behavior is the fundamental way out for solving agricultural pollution problems. The integrated method of economics and environmental science as intersecting disciplines may provide a feasible approach for exploring continuous research from policy design, farming household behavioral response, and pollutant emission and migration, all the way to regional environmental security analysis.