| Agricultural non-point source pollution is the main source of water pollution in China,and has become an indispensable factor affecting agricultural and rural economic development and rural ecological sustainability.Planting occupies for a large proportion of nitrogen and phosphorus discharged from agricultural sources.Control of farmland non-point source pollution is the indispensable way to achieve effective control of agricultural non-point source pollution.In the control of farmland non-point source pollution,there is an interactive game relationship between the governments and farmers.The conflict of interests between the farmers’ pursuit of maximizing their own economic benefits and the governments’ pursuit of maximizing the interests of whole society is an important reason for the weakness of farmland non-point source pollution control.Therefore,this paper discusses the control of farmland non-point source pollution based on evolutionary game theory.Firstly,a model of farmland non-point source pollution control was introduced as a case.Secondly,on the basis of summarizing the relevant domestic and international research results and practical observations,ideal assumptions are made to establish an evolutionary game model between local governments and farmers regarding the adoption of source reduction technologies(SRTs),as well as an evolutionary game model regarding the adoption of SRTs within the farmers considering the influence of information asymmetry and herd psychology under the continuous guidance of local governments.Finally,the evolutionary stability strategies(ESSs)and evolution process of stakeholders in the promotion of SRTs are analyzed,and the influence of key parameters on the model is discussed through sensitivity analysis.The results of the study not only contribute to the construction of endogenous mechanisms for long-term control of agricultural non-point source pollution,but also provide scientific references for the government to formulate relevant policies.The main conclusions of this paper are as follows:(1)The "whole process-whole period-whole cycle" model(referred to as the "three whole" model)is introduced,which is suitable for the comprehensive control of non-point pollution in grain vegetable rotation farmland in hilly areas.This model integrates key technologies in each control link according to local conditions,highlighting regional characteristics,and is low cost and easy to popularize.After the demonstration and application of the "three-whole" model in Quxi small watershed,the nitrogen and phosphorus pollution loads were reduced by more than 40%,and the main pesticide dosage(glyphosate,isoprothiolane and imidacloprid)was significantly reduced.However,through field observation and communication with farmers,it is found that technical means alone may be difficult to achieve long-term control of farmland non-point pollution.It is extremely important to involve multiple stakeholders in farmland non-point source pollution control from the perspective of management.(2)In the evolutionary game model of SRTs adoption between local governments and farmers,there are four possible scenarios and three different ESSs,and the constraints of the four scenarios are put forward.There is a "prisoner’s dilemma" in the promotion of SRTs.At this time,both local governments and farmers choose the more favorable strategy of not to participate in the control of farmland non-point source pollution.As a result,both sides have not obtained the optimal benefits,and the amount of subsidies only slightly changes the evolution rate,and does not affect the final ESS.To achieve the optimal situation of farmers’ spontaneous adoption of SRTs without local government guidance,the perceived indirect benefits of SRTs adoption by farmers need to be greater than the costs of SRTs adoption.Therefore,it is critical to improve the level of scientific knowledge and environmental awareness of farmers,but this will require long-term efforts to achieve qualitative change.At present,the continuous guidance of local governments is the key to solve the "prisoner’s dilemma" in the promotion of SRTs and ensure the long-term adoption of SRTs by farmers.When the cost of guidance is greater than the additional penalties imposed for not guiding,local governments tend not to guide.Therefore,it is important to establish a sound system of SRTs adoption,including reducing the cost of adoption by local governments and strengthening the regulation and penalties for not adoption.Under the continuous guidance of the local government,the subsidy will have a greater impact on the final ESS and need to be set reasonably to avoid the loss of farmers’ motivation to adopt SRTs due to too little subsidy.(3)In the evolutionary game model on SRTs adoption within the farmer groups,there are three different ESSs,and their constraints are put forward.On the premise that farmers who are more likely to accept SRTs(defined as farmer A)finally choose to adopt SRTs,other farmers may finally choose to adopt SRTs.Therefore,first of all,it is necessary to ensure that farmer A has sufficient willingness to adopt SRTs,at least to ensure that farmer A will not suffer too much loss of interests if farmer A adopts SRTs alone.Subsequently,under the herd psychology,local governments can enhance other farmers’ perceptions of the non-direct benefits of adopting SRTs through guidance or training,so that as many other farmers as possible follow farmer A to choose to adopt SRTs,in order to achieve a good promotion effect.The difference in the proportion of farmers’ initial strategy choices can directly affect the evolution speed to the steady state.In addition,under certain conditions,the system may have two opposite ESSs: all farmers adopt SRTs and none adopt SRTs,and slight changes in the initial strategy selection proportion and key parameter values may result in different final ESSs.Therefore,it is important to increase the proportion of farmers adopting SRTs as much as possible at the beginning of the SRTs promotion. |