Opportunities or Risks: Economic Impacts of Climate Change on Crop Structure Adjustment in Ecologically Vulnerable Regions in China

Abstract

Global warming by 2 °C or above will frequently see weather beyond the critical tolerance threshold for agricultural extreme high temperatures. If so, people will have to more accurately evaluate the opportunities and risks posed by future climate change while adjusting the structure of agricultural production. However, accurate assessment results of the impacts of climate change on crop yield are absent in the current studies on the impact of climate change on the agricultural economy. To address this gap, this paper sets forth a comprehensive evaluation method using a crop model coupled with a computable general equilibrium model. According to research findings, future climate change may continue with the trend of the continued decline of grain planting areas and the continued growth of cash crop planting areas in ecologically vulnerable regions. This will make grain security more difficult. On one hand, perennial crop planting broadens the eco-space for future economic and social development in ecologically vulnerable regions. Therefore, attention should be paid to grain security. The cash crop planting area should not be excessively expanded. On the other hand, it is advised to plant perennial crops in those areas which are unsuitable for planting cash crops.

1. Introduction

According to Climate Change 2021: The Physical Science Basis, the report of the Intergovernmental Panel on Climate Change (IPCC) Working Group I, global warming (i.e., the average temperature of the Earth’s surface air and oceans) by 2 °C or above will frequently see weather beyond the critical tolerance threshold for agricultural extreme high temperatures. If so, people will have to more accurately evaluate the opportunities and risks posed by climate change while adjusting the structure of agricultural production. Li Yingchun et al. [1] apply the multi-indicator method to evaluate the agricultural vulnerability of China at a regional scale in the 2010s and the 2040s under the A1B Climate Scenario. As found in SRESA1B, the provinces of Guizhou, Guangxi, and Yunnan are among the most vulnerable regions in China, which suffer from high exposure and/or sensitivity to climate change, relatively low capacity for economic development and adaptability, and the increasing diversion of farmland to other uses, thereby possibly exacerbating such vulnerability. The studies on these regions reveal the common characteristics of their vulnerable agricultural ecological environment: widely and severely distributed rocky desertification, frequent meteorological disasters, scarce agricultural land, and weak agricultural anti-interference capacity [2,3]. Moreover, statistical data indicate that the agricultural production industry in Guizhou, Yunnan, and Guangxi was continuously dominated by non-grain production of cultivated land from 1980 to 2020. As a result of this change, the proportions of the grain crop planting area in the cultivated area of Guizhou, Yunnan, and Guangxi decreased from 85%, 89.7%, and 81.1% to 50.3%, 59.6%, and 45.9%, whereas the proportions of cash crops conversely increased by 34.7%, 30.1 and 35.2%, respectively. These above facts raise a question for changes in the structure of crop production in ecologically vulnerable regions: What opportunities and risks will arise from future impacts of climate change on crop production in ecologically vulnerable regions? The research outcome of this question will furnish insights into balancing grain production and agricultural economic development in ecologically vulnerable regions under the impact of climate change. The following research methods should be adopted to answer this question: Firstly, the impacts of climate change on crop yield in ecologically vulnerable regions are simulated, which can be based on a crop model to discuss the impacts of climate change on crop production, with reference to the existing domestic and foreign studies. Secondly, an economic model is built to describe the intermodal relationship between crop yield change and the agricultural economy in ecologically vulnerable regions. The complex connections between crop yield change and other economic systems are taken into account. Such an economic model can be based on the computable general equilibrium (CGE) model.

However, accurate assessment results of the impact of climate change on crop yield are absent in the current studies on the impact of climate change on the agricultural economy. To address this gap, this paper, taking Guizhou Province as an example, sets forth a comprehensive evaluation method to give play to a crop model coupled with a computable general equilibrium model. So far, the impacts of climate change on the regional agricultural economy have been analyzed from the perspective of “Climate Change—Crop Yield”. Specifically, in Step 1, a crop model is built to simulate the impacts of climate change on crop yield in the ecologically vulnerable Guizhou Province by giving full consideration to the complex carbon and nitrogen cycling systems (such as climate, soil, and farmland management system). In Step 2, the regional dynamic CGE model is built and the productivity conversion factor is introduced into the CGE model. In Step 3, the impacts of climate change on the agricultural economy in ecologically vulnerable Guizhou Province are evaluated so as to construct an analytical framework of “Economic Shock on -Crop Yield-in the Context of Climate Change”.

The rest of this paper consists of the following parts: Section 2 explains the research methodology and data sources. Section 3 gives a simulation of the impacts of climate change on the agricultural economy of Guizhou Province. Section 4 contains the analysis of climate change on the agricultural economy of ecologically vulnerable regions, and Section 5 ends with the research conclusions.

2. Method and Materials

2.1. Overview of the Study Region and Its Crop Production

2.1.1. Study Region

Guizhou Province is located in the hinterland of Southwest China (103°36′ E~109°35′ E, 24°37′ N~29°13′ N), where the Yangtze River and Pearl River, two major river systems, flow through. With an average elevation of about 1100 m, Guizhou Province is the East Asian center of the world’s three major karst regions. The area of carbonate outcrops accounts for about 73% of the provincial land area [4], and mountainous and hilly land represents 92.5% of the provincial area. Therefore, Guizhou Province is famous for its unique landscape of “eight mountains, one river and one piece of land”. Guizhou Province features a subtropical monsoon climate, where a complex, diverse, warm, and humid climate dominates. Due to atmospheric circulation, topography, and other factors, the climate of Guizhou Province is diverse, as a saying goes: “The four seasons prevail in a mountain, and different weather conditions are demonstrated in two places at a distance of ten li (Chinese unit of length)”. According to the General Annals of Guizhou Province, local weather is changeable in one day. Guizhou Province has an average annual temperature of 10–20 °C, ≥annual accumulated temperature of 4000–5000 °C under the precondition of at least 10 °C, average temperature of generally 3–6 °C in the coldest month (January is higher than that in other areas of the same latitude), and an average temperature of generally 22–25 °C in the hottest month (July), which is typically mild in winter and cool in summer. Water resources abound in Guizhou Province on the whole, with annual average precipitation ranging from 1100 mm to 1500 mm. The rainy season is distinct. Spatial distribution decreases from the southern and northeast parts to the northern and western parts. There are a large number of rivers in Guizhou Province, with a total of 1059 rivers with a single drainage area of over 50 km. The annual average runoff was 106.2 billion cubic meters in 2018. However, Guizhou is disadvantaged by climate instability, with severe agricultural disasters, such as drought, autumn wind, freezing, and hail. The soil resource is scarce and the quality of agricultural soil resources is low in Guizhou Province. According to statistical data from Qin Song et al. [5], yellow soil, limestone soil, and red loam were the most in the provincial soil area of 15.91331 million hm², accounting for 46.4%, 17.5%, and 9.7%, respectively. The cultivated soil resources merely totaled 3.7315 million hm², which were mainly in the form of sloping fields totaling 1.7773 million hm². However, dam fields with superior comprehensive agricultural conditions totaled only 446,200 hm². In 2019, the forest coverage rate of the whole province reached 59.95%, and the vegetation types were mainly woodland, shrub, grassland, permanent wetland, crop/natural vegetation mosaic, etc.

2.1.2. Overview of Crop Production in the Study Region

Since China’s reform and opening-up, Guizhou has witnessed the expanded agricultural sown area, increasing yield, and improved yield per unit area and labor productivity. The growing capital input of chemical fertilizers, plastic mulches, and pesticides has not only improved the yield of agricultural products but also polluted the ecological environment and undermined farmland quality. Meanwhile, the agricultural structure continuously changes in Guizhou Province. As depicted in Figure 1, from 1988 to 2020, the proportions of the cultivated area for rice, wheat, and corn in the provincial cultivated area dropped continuously. In particular, the proportion of the cultivated area for rice dropped from 21.9% to 12.3%, and the cultivated area shrank from 146,000 hm² to 82,000 hm². The proportion of cash crops kept rising, in particular, the proportion of vegetables sharply rose from 4.8% to 25.6%. Correspondingly, the growth rate of grain yield remained moderate, and only potato yield bottomed with a growth rate of 93.7%. The yield of cash crops increased significantly, and especially tea leaf yield grew by 879%.

In the same period, the average disaster rate and disaster rate of crops reached 33.6% and 17.9% in Guizhou Province, respectively. As shown in Figure 2, since the reform and opening-up, natural disasters hit more and more agricultural areas in Guizhou Province, which even reached a staged peak in 1992, followed by a slight drop. However, in 2010–2011, Guizhou Province suffered from the most extensive serious natural disasters, and then the affected area remarkably decreased. In some years, the depth of natural disasters would sharply intensify, which will pose uncertain risks to the economic development of Guizhou Province.

2.2. Research Methodology

Compared with other model methods, the crop model coupled with the CGE model adopted in this paper more accurately reveals the impacts of climate change on crop yield. On this basis, the economic shocks from such impacts can be simulated in a more satisfactory manner.

Numerous studies on the economics of climate change have been undertaken in the wake of a paper by Nordhaus in 1977. The subsequent large number of studies give birth to a lot of research methods about climate change economy. As pointed out by Li Tong et al. [6], the CGE model describes the overall equilibrium state of the economic system. Quantity and the endogenous price interact and adjust, thereby resulting in the optimal allocation of resources. Therefore, the CGE model emerges as a powerful tool in climate policy research. One of the key issues in the studies on the impacts of the agricultural CGE model on climate change lies in the way to integrate climate change and its impacts on economic activities in a reasonable way within the framework of the CGE model. The differences in understanding of the impact mechanism of climate change lead to different structures of the agricultural CGE model and different model assessment results. In this case, many curtailing problems will eventually come out in the proposed agricultural policies to cope with climate change. Under the review of the ideas of previous studies, some studies took meteorological elements as a production input, for example, Erol et al. [7] and Roberto [8]. However, it is difficult to find appropriate values to comprehensively describe meteorological elements and shed light on the complex process of climate change to exert impacts on the agricultural economy. At present, most studies regard meteorological factors as agricultural production constraints and introduce a productivity conversion factor embedded in the production function of the model to describe the impacts of climate change on agricultural output efficiency, thereby raising a solution to the above problems.

2.2.1. Analysis of CGE Modeling for Climate Change as Agricultural Production Constraints

In the agricultural CGE modeling for climate change, the idea of taking meteorological factors as agricultural production constraints refers to taking the impacts of climate change on the agricultural economy as an exogenous biochemical method and characterizing the impacts of climate change on the agricultural economy by changing the production efficiency of the production function. The meteorological factor account is not set in the SAM table, and no income is obtained in the capital flow structure of the model. The production function of the standard model is expressed as follows:

$$min{\displaystyle \sum }_{i=1}^n{P}_i{X}_i$$

(1)

$$s.t. QV{A}_i=A{\left[{\displaystyle \sum }_{i=1}^n{\delta }_i{X}_i^v\right]}^{-\frac{1}{v}}$$

(2)

The study, with meteorological factors as production constraints, multiplies Equation (2) by a productivity conversion factor R. Implementation model is the rate of change in agricultural yield under the impacts of climate change. The R-value, as a key variable, can be obtained in three ways: Firstly, the R-value is estimated through econometric models, as mentioned in the studies by Maria Sassi et al. [9] and Saeed [10]. Secondly, the R-value is defined directly in the simulation scenario of the CGE model, as mentioned in the studies by Xu Taotao et al. [11]. The productivity conversion factor R is not attributed to Hicks-neutral technical progress (in this case, it may be Harrod-neutral or Solow-neutral, with the substitution of labor and capital), and Equation (2) is modified as follows:

$$QV{A}_i=A{\left[{\displaystyle \sum }_{i=1}^n{\delta }_i{\left[R·X_i\right]}_i^v\right]}^{-\frac{1}{v}}$$

(3)

According to the studies made by Li Ximing et al. [12] Thomas et al. [13] and Hang Delin et al. [14], the R-value is similarly defined in the literature under the CGE model simulation scenarios, but the productivity conversion factor R is attributed to Hicks-neutral technical progress, and Equation (2) is modified as follows:

$$QV{A}_i=R·A{\left[{\displaystyle \sum }_{i=1}^n{\delta }_i{X}_i{}_i^v\right]}^{-\frac{1}{v}}$$

(4)

The R-value is obtained in the third way, i.e., simulation, as mentioned in the studies by Suda Rshan et al. [15], Yuan Feng et al. [16], Zhao Zijian et al. [17], and YosRi [18]. They apply a crop model to simulate changes in crop yield under the impacts of climate change. The rate of change is defined by the R-value. Such studies embody several advantages. Firstly, the SAM table and CGE model are concise and clear. Secondly, this method can reveal the impacts of climate change on agricultural factor productivity more comprehensively. These studies begin with the calculation of the rate of change of crop yield per unit area under the impacts of climate change and then simulate the change in the economic system caused by the change in crop yield per unit area.

However, the above three research methods are disadvantageous due to some problems. The first problem lies in that the productivity conversion factor, introduced by these studies, is biased. There are three main reasons for this. Firstly, the rate of change of crop yield, estimated through the econometric models, may present a weak fitting degree, as crop yield is subjected to combined effects of the crops, soil, meteorological factors, and farmland management. The dependent variables are limited in the econometric models, and it is difficult to incorporate all these combined factors. If the study region of the econometric model goes beyond the study region of the agricultural CGE model, the accurate fitting degree of the rate of change of crop yield (R-value) would further reduce. Secondly, defining the productivity conversion factor R through the simulation scenario setting or the reference to define the productivity conversion factor R depends on the experience level of the research personnel. The results from the agricultural CGE model are relatively untrustworthy. Thirdly, the climate scenario setting is not all-inclusive. The climate change scenario setting discretely separates temperature, precipitation, and atmospheric CO₂ concentration, which deviates from the scientific fact that crop growth is subjected to temperature, precipitation, and atmospheric CO₂ concentration. The defects of the incomplete precipitation scenario setting and low atmospheric CO₂ concentration simulation also undermine the research results. The second problem of such studies is that the types of technical progress in agricultural production under the impacts of climate change are not properly identified. The type of production technology progress implies the approaches to integrating productivity conversion factor with the production function. One type is Solow-neutral technical progress or Harold-neutral technical progress (Equation (3)), and the other type is Hicks-neutral technical progress (Equation (4)), which presents the different research ideas and conclusions made by research personnel when climate change is used as a constraint condition for agricultural production and the CGE model is nested.

2.2.2. Improvement of Modeling Idea

• Multi-model coupling

In this paper, multi-model coupling assessment refers to the introduction of crop model simulation results (as the exogenous parameter R) into the agricultural CGE model, which is an exogenous biochemical processing method for the transmission of the crop model to the agricultural CGE model. Functions of the crop model for single or multiple crops into point simulation and regional simulations are adopted. Using the model features with full consideration to crops, soil, meteorological factors, and farmland management activities, according to the needs of agricultural CGE model research, the climate change scenarios are designed to carry out directional simulation for changes in crop yield of specific regions and specific crops under climate change, thereby solving the problem of poor degree of fitting of the productivity conversion factor R. This improvement can make up for the lack of the CGE model in describing the technical details of agricultural production, and embody the technical details of agricultural production in the change of productivity conversion factor. The multi-model coupling assessment process is as follows. Firstly, a crop model is used to simulate the impacts of climate change scenarios on crop yield in the study region, and such climate change scenarios comprehensively take into account temperature, precipitation, atmospheric CO₂ concentration, and other factors. Secondly, the rate of change in crop yield is introduced into the agricultural CGE model as an exogenous parameter. Finally, the changes in the regional economic system are simulated.

It should be noted that the multi-model coupling assessment herein is an exogenous biochemical method that adds the crop model into the CGE model, which is different from the integrated assessment model (IAM) of climate change. Wang Can et al. [19] point out that the integrated assessment model (IAM) is an exogenous biochemical processing method under the framework of the CGE model by adding physical climate equations or modules to the CGE model. Related research works include Rose et al. [20], Alvaro et al. [21], and Alvaro et al. [22]. Cline [23] and the World Bank [24] build larger and more complex integrated assessment models by adding climate modules and crop modules to the CGE model framework.

2.

Identification of the types of technical progress

It is important to identify the types of technical progress that productivity changes when meteorological factors are used as agricultural production constraints. As mentioned by David et al. (translated by Li Shantong et al. [25]), in the two-factor production function, neutral productivity growth can be realized through the efficiency parameter in the conditional Equation (4), while the Hicks-neutral productivity growth can only be realized through the R coefficient related to the labor factor in Equation (3). In fact, climate change influences the efficiency of all agricultural inputs rather than changing a certain factor of production efficiency. That is to say, the productivity conversion factor R is a Hicks-neutral technical progress. Thus, the introduced productivity conversion factor R should be multiplied by the production function efficiency A under the CGE model. Recent studies by Wang Lu et al. [26] also prove that agricultural output is more dependable on improvements in the efficiency of all inputs instead of a single production factor, because in agricultural production, “the intermediate input coefficient of household agricultural production have risen to 0.4, intermediate input coefficient of land and labor have fallen to 0.22 and 0.24, and intermediate input coefficient of capital has remained low at 0.02”.

3.

Optimization of climate simulation scenarios

Studies on the impacts of climate change on crop yield usually refer to climate simulation scenarios of IPCC assessment reports, which take meteorological elements of temperature and precipitation into account. In the course of small region simulation, climate scenarios of IPCC assessment reports can be cited with regard to temperature change, but the climate characteristics of the study region should be fully considered under the precipitation scenario. Because of the complexity of precipitation scenarios, it is difficult to control the bias of the rate of change of crop yield R if the climate forecast data in line with local climate characteristics cannot be used to simulate the precipitation scenario. Moreover, atmospheric CO₂ fertilizer efficiency should be taken into account under climate scenarios. Therefore, the design of climate simulation scenarios should comprehensively investigate temperature, precipitation, and atmospheric CO₂ concentration. In addition, the results of climate prediction in the study region should be fully used for reference.

2.2.3. Multi-Model Coupling Herein

This paper evaluates the impacts of climate change on the agricultural economy caused by the yield changes of rice, corn, and tea leaf, and adopts the DNDC model coupled with the CGE model to analyze the impacts of climate change on the agricultural economy in Guizhou Province. The DNDC crop model was developed by Professor Li Changsheng and his team at the Institute for the Study of Earth, Oceans, and Space at the University of New Hampshire, USA. It can be used to simulate crop yield, soil carbon sequestration, nitrate leaching, and carbon and nitrogen emissions in agroecosystems, which consists of seven geographic information system (GIS) files, a meteorological database, a soil database, and a crop database to enable the model to simulate the most sensitive factors of each crop type one by one for the smallest simulation units and sum the simulation results to have access to the regional simulation results. For the unavoidable errors in the regional simulation, the determination coefficient R² indicator is used to verify the degree of coincidence between the simulated yield value and the actual statistical value under the research idea set forth by Gao Maofang [27]. The DNDC simulation database of rice, corn, and tea leaf in Guizhou Province is constructed as the simulation object, with tea leaves seeing the largest yield change among cash crops in Guizhou Province.

The CGE model is outreached and developed on the basis of the Chinese agriculture general equilibrium model developed by the “National Agricultural Strategy Analysis and Decision Support System Open Laboratory”—the “Belt and Road” Agricultural Strategy Analysis Platform of the Chinese Academy of Agricultural Sciences. The GAMS software and PATH solver, developed by World Bank, are used to solve the model. The model deems climate change as an agricultural production constraint and assumes that climate change influences all agricultural inputs and thus changes the yield level of agricultural activities. Based on this precondition, a Hicks-neutral productivity conversion factor R is introduced into the production module, which is defined as the rate of change of crop yield per unit area simulated by the above DNDC model. The function expression of agricultural production activities is as follows:

$$min {\displaystyle \sum }_{i=1}^n\left(p_{Ki}{Q}_{Ki}+p_{Li}{Q}_{Li}+p_{Ei}{Q}_{Ei}+p_{INTi}{Q}_{INTi}\right)$$

(5)

$$\begin{array}{cc}\hfill s.t.Q_i=\left(1+R\right)& \\ & ·A_{3-i}{\displaystyle ⟦}{\alpha }_{KLEi}⟨A_{2i}{\left\{{\alpha }_{KLi}{\left[A_{1i}{\left({\alpha }_{Ki}{Q}_{Ki}^{-{\rho }_{1i}}+{\alpha }_{Li}{Q}_{Li}^{-{\rho }_{1i}}\right)}^{\frac{-1}{{\rho }_{1i}}}\right]}^{-{\rho }_{2i}}+{\alpha }_{Ei}{Q}_{Ei}^{-{\rho }_{2i}}\right\}}^{\frac{-1}{{\rho }_{2i}}}⟩{}^{-{\rho }_{3i}}\hfill \\ & +{\alpha }_{INTi}{Q}_{INTi}^{-{\rho }_{3i}}{\displaystyle ⟧}{}^{\frac{-1}{{\rho }_{3i}}}\hfill \end{array}$$

(6)

where $i$ denotes the $i$ sector; $A_{1i}$, $A_{2i}$, and $A_{3i}$ indicate the efficiency values of nested layers, respectively, $Q_{Ki}{Q}_{Li}{Q}_{Ei}$. $Q_{Ki}$, $Q_{Li}$, $Q_{Ei}$, and $p_{Ki}$, $p_{Li}$ and $p_{Ki}{p}_{Li}{p}_{Ei}$ describe capital, labor, land input, and their respective prices, while ${\alpha }_{Ki}$, ${\alpha }_{Li}$, and ${\alpha }_{Ei}$ refer to their share parameters, respectively. ${\alpha }_{KLi}$ and ${\alpha }_{KLEi}$ denote the capital–labor factor combination and capital–labor–land factor combination, respectively. ${\rho }_{1i}$, ${\rho }_{2i}$, and ${\rho }_{3i}$, respectively, represent the elasticity of substitution at each nested layer. The method of nested function, proposed by Zhang Xin [28], is used to solve the problem of the same $\rho$-value when multiple layers are nested, and such value refers to the research results made by Fan Xiaojing [29]. The constructed production structure consists of three layers: the first layer is the primary elements of capital and labor, which are combined into the capital–labor input combination with the CES function. The second layer is the primary elements of capital–labor and land, which are combined as factor inputs with the CES function. The third layer is factors and intermediate inputs, which are combined as production inputs with the CES function. The other consumption module, institution module, and trade module adopt the standard CGE module. The model adopts a dynamic recursive mechanism, which is closely related to endogenous economic growth theory according to Lou Feng [30]. Capital is endogenously determined by capital accumulation in the previous period and investment in the current period. Keynesian closure is used in the model.

2.3. Data Sources

2.3.1. Data Sources of DNDC Crop Model

The meteorological database of the DNDC crop model brings together daily historical data and predicted data on the maximum temperature, minimum temperature, and precipitation. The meteorological data are written into files according to the format requirements of the DNDC model. Given the integrity of meteorological data over the years, 34 meteorological stations, such as Weining Meteorological Station and Shuicheng Meteorological Station, are selected. Figure 3 shows the distribution of 34 meteorological stations in Guizhou Province. The historical meteorological data are obtained from the daily ground observations by 34 meteorological stations of Guizhou Provincial Meteorological Bureau from 2010 to 2020. For the predicted meteorological data, reference is made to the research findings of Zhang Jiaoyan et al. [31] who compare the climate simulation effects of Guizhou Province under different climate models. The daily meteorological predicted data of four typical concentration paths of the RCP scenario of Community Climate System Model Version 4 (CCSM4) under the CMIP5 Global Climate Model are used. Kriging Method is applied to interpolate 1° × 1° to the above 34 meteorological stations in Guizhou Province. Due to the estimated time frame of the study until 2050, the simulated future time periods are divided into 2021–2030 and 2031–2050 according to the time-division methods of IPCC AR5 and IPCC AR6.

The soil database contains soil organic carbon content (SOC), clay content, soil pH, and bulk density. Data are cited from Guizhou Soil Species Records [32], The New Edition of Guizhou Soil Species Records [33]. The soil in the basic simulation unit is assumed to be mean during the regional simulation, and the soil data of meteorological stations in the soil database represent the soil rational state of the counties where the soil is located. Given the deep-plowing habit (arable land depth of 30 cm) of crop cultivation in Guizhou Province, the data of soil tillage layer (30 cm above the ground) are used.

The farmland management data are obtained from the field surveys for 34 counties with the corresponding meteorological stations, such as those in Weining County and Panxian County. The survey data include sowing and harvesting time, soil tillage, basin irrigation, fertilizer use, straw residue management, crop parameters, etc. The survey data are distributed to grids according to different geographical locations. Rice and corn are ripe once a year, and tea leaf is perennial. The carbon content of crop fruits refers to model parameters. Model default values are used for other farmland management parameters. During the field survey, one township (town) in each county is selected to collect farmland management data, and then two villages in the township (town) are selected to collect the corresponding farmland management data. The two groups of data are compared to retain the data which comply with the local agricultural production situations.

2.3.2. Data Sources of the CGE model

The SAM table of the CGE model refers to the methods set forth by Huang Delin [34] and Ju Shaopeng et al. [35]. For Guizhou Province and other parts of China, the agricultural sector in the input–output table for the year 2017 is expanded and divided into 13 subsectors for rice, corn, tea leaves, and other cash crops. The industrial sector and service sector are combined into 17 subsectors. The residential sector falls into the urban resident subsector and rural resident subsectors. The control numbers of the macro-level SAM table are derived from the input-output table and various statistical yearbooks for the year 2017. According to the research needs, some control numbers are split or combined to construct the micro-level SAM table. Finally, the SAM table is balanced by the one-step one-calibration method. Most of the parameters of the model are obtained by calibration. The required exogenous given technical progress rate, population, and labor force growth rate refer to research results set forth by Jean FouRe et al. [36].

3. Results

3.1. Simulation Scenario Design

Firstly, it is necessary to set up the crop model to simulate the scenario. This paper envisages the future socioeconomic scenario and then takes into account the future carbon emission scenario further sets the future climate change scenario and fully considers the climate characteristics of small regions. In the future, the development path at the cost of high emissions will be abandoned by various countries. However, the development pressure and emission reduction externalities also make it difficult to achieve the ideal emission scenario. The most likely emission scenario will be a “moderate medium emission scenario”, which is “higher than the low-concentration path scenario and lower than the medium-concentration path scenario”. Therefore, the “low-concentration path scenario” and “medium-concentration path scenario” are considered in the temperature model scenario setting: low warming under a temperature rise of 0.6 °C, and moderate warming under a temperature rise of 1 °C. As for the precipitation, the regional characteristics of Guizhou Province are considered with reference to the previous research results. According to the monitoring data of the National Oceanic and Atmospheric Administration (NOAA) of the United States, the global atmospheric CO₂ concentration was 390 ppm in 2010 and 415 ppm in 2019. The future climate change simulation scenarios for Guizhou Province are shown in

Table 1. Future climate change simulation scenarios.

Meteorological Scenarios	Temperature Rise (°C)	Precipitation Change (%)		Increase in Atmospheric CO2 Concentration	References
	2021–2050	2021–2030	2031–2050
S1	0.6	−1	−7.5	2.5 ppm/year	Regional Climate Change Assessment Report of the Southwest China: 2020 (Compilation Committee of) [37], Zhang Jiaoyan et al. [38]
S2	0.6	−8.5	7.5
S3	1	−1	−7.5
S4	1	−8.5	7.5

.

Secondly, the CGE model is set to simulate the scenario. In this paper, the rate of change of crop yield per unit area from 2021 to 2050, simulated with the DNDC model, is used as an exogenous shock variable multiplied by the total factor productivity variable A of the agricultural production function, in an effort to simulate the economic changes caused by the shock from climate change on yields of rice, corn, and tea leaf in Guizhou Province during the next 30 years (2021–2050). In this paper, the economic impacts of climate change in the short and medium term are analyzed. According to the IPCC AR6 Report, with the same time division as that under the DNDC model, the simulation time period of the CGE model falls into the short and medium term, i.e., the short term focuses on interdecadal variation during 2021–2030, and the middle term focuses on interdecadal variation during 2031–2050. The initial year of the CGE model is 2017, the simulation time span is 2017–2050, and the result inspection and presentation time is 2021–2050. Simulation scenarios of economic shock under the set CGE model are shown in

Table 2. Simulation scenarios of shock for the agricultural economy.

Meteorological Scenarios	2021–2030			2031–2050
	Rice	Corn	Tea Leaf	Rice	Corn	Tea Leaf
S1	Rate of change per unit yield (%)			Rate of change per unit yield (%)
S2
S3
S4

.

3.2. Simulation Results

3.2.1. Impacts of Climate Change on Crop Yield in Guizhou Province

The yields of rice, corn, and tea leaf in Guizhou Province are accessed d by regional simulation under the DNDC model version 9.5. In this paper, the simulated yields of the model are compared with the statistical data of crop yield to analyze the applicability of the model in Guizhou Province. The fitted R² values of the simulated yield and statistical data of rice, corn, and tea leaf are 0.8633, 0.839, and 0.8269, respectively, and the fitting results are good. The simulation results of crop yield in Guizhou Province, affected by climate change, are shown in

Table 3. Interdecadal Variation of Yield per Unit for Rice, Corn, and Tea Leaf in Guizhou Province (%).

Meteorological Scenarios	Scenarios	2021–2030			2031–2050
		Rice	Corn	Tea Leaf	Rice	Corn	Tea Leaf
S1	Scenario 1	−0.6	−3.0	3.0	−3.6	−3.7	2.0
S2	Scenario 2	1.0	−3.0	3.0	−3.8	−3.7	5.4
S3	Scenario 3	0.5	−2.8	4.8	−2.9	−4.8	0.4
S4	Scenario 4	1.0	−3.1	4.8	−4.0	−4.5	3.6

. The yield simulated under the DNDC model is the carbon content of crop fruits. Since the carbon content of crop fruits is relatively stable, its rate of change can be regarded as the rate of change of crop yield per unit area.

According to the scenarios, from 2021 to 2050, the interdecadal temperature rise and atmospheric CO₂ concentration increase would make rice yield per unit increase and then decrease, the corn yield per unit decrease, and tea leaf yield per unit increase in Guizhou Province. Due to complex situations from the perspective of precipitation, when the temperature and atmospheric CO₂ concentration increase slightly, the decrease in precipitation will have few impacts on crop yield in Guizhou Province. When the temperature and atmospheric CO₂ concentration rise further, the increase in precipitation is more beneficial to crop yield than the decrease in precipitation in Guizhou Province. From 2021 to 2030, when the atmospheric CO₂ concentration increases to 446.5 ppm, the overall rice yield will slightly increase, the corn yield will decrease, and the tea leaf yield will increase under different temperature rise and precipitation scenarios. From 2031 to 2050, when the atmospheric CO₂ concentration increases to 496.5 ppm different warming and precipitation scenarios will reduce the rice yield and corn yield, while the tea leaf yield will still increase. This is different from the conclusions of studies on a larger spatial scale. For example, Xie Wei et al. [39], based on a study of China, propose that the yield per unit of major grain crops in China would decrease by 2.6% in case of temperature rise by every 1 °C and the grain yield would increase by 0.4% in case of precipitation rise by every 1%. Due to the effect of CO₂ on fertilizer efficiency, the yield per unit of grain crops would increase by about 16% on average. This discrepancy may be owed to the differences in climate, soil, and farmland management.

By crop species, the interdecadal temperature rise and atmospheric CO₂ concentration increase, and different precipitation scenarios would make rice yield per unit increase and then decrease. The annual precipitation is high in Guizhou Province, which mainly happens from April to September during the rice-growing period (Wu Junming et al., 2007). A decrease of 1% or 8.5% in precipitation has limited impacts on rice yield per unit in Guizhou Province. In addition, rice is a type of thermophilic crop. A slight increase in temperature and atmospheric CO₂ concentration is generally beneficial to rice production. However, with further increases in temperature and atmospheric CO₂ concentration, rice production would reduce in Guizhou Province, and the extent of reduction would be greater than that of the increase in the previous period. During the same period, with the increase in temperature and atmospheric CO₂ concentration, corn yield per unit in Guizhou Province would continue to decrease. With the further increase in temperature and atmospheric CO₂ concentration, the decrease in corn yield per unit in Guizhou Province would be greater, and climate change would have a significant negative impact on corn production in Guizhou Province. From 2021 to 2050, the increase in temperature and atmospheric CO₂ concentration would increase the tea leaf yield per unit in Guizhou Province, but the yield increase performance would differ in 2021–2030 and 2031–2050. Tea trees prefer high temperatures and humidity and are sensitive to coldness. The abundant annual precipitation in Guizhou Province implies that a decrease in precipitation by 1% or 8.5% would have few impacts on tea leaf yield. Therefore, a slight increase in interdecadal temperature and atmospheric CO₂ concentration during 2021–2030 would be beneficial to increase the tea leaf yield. From 2031 to 2050, after the further increase in temperature and atmospheric CO₂ concentration, precipitation has greater impacts on tea leaf yield than that in the previous interdecadal period. The increase in precipitation would be more beneficial to tea leaf yield than the decrease in precipitation.

3.2.2. Economic Impacts of Changes in Crop Yield on the Agricultural Sector

In this paper, the dynamic CGE model is used to simulate the shocks of changes in the crop yield on the agricultural economy of Guizhou Province. The impacts of climate change on the agricultural economy in Guizhou Province are shown in

Table 4. Impacts of Climate Change on the Agricultural Sector Economy (%).

Crops	Scenarios	Yield Value		Product Price		Net Outflows		Rural Residents’ Consumption Expenditure
		2021–2030	2031–2050	2021–2030	2031–2050	2021–2030	2031–2050	2021–2030	2031–2050
Rice	S1	−0.355	−0.377	0.972	1.232	−2.547	−4.176	−0.499	−0.569
	S2	0.118	−0.278	−0.136	0.552	0.369	−1.600	0.166	−0.425
	S3	0.047	−0.239	−0.072	0.443	0.188	−1.285	0.063	−0.364
	S4	0.112	−0.272	−0.136	0.603	0.368	−1.811	0.155	−0.416
Corn	S1	−0.126	−0.053	1.776	0.688	−8.248	−2.780	−0.650	−0.411
	S2	−0.076	−0.075	0.604	0.745	−1.934	−3.018	−0.411	−0.498
	S3	−0.072	−0.073	0.551	0.742	−1.733	−3.015	−0.396	−0.489
	S4	−0.077	−0.089	0.630	1.408	−2.040	−3.085	−0.427	−0.631
Tea Leaf	S1	0.150	0.410	−0.194	−0.350	0.523	1.185	0.337	1.002
	S2	0.274	1.088	−0.341	−0.576	0.943	2.598	0.598	2.713
	S3	0.487	0.521	−0.470	−0.442	1.422	1.536	1.066	1.291
	S4	0.488	1.151	−0.470	−0.574	1.423	2.666	1.069	2.872

. In this table, the net outflow refers to the difference between the value of the product manufactured in the province but outflowing outside the province and the value of the product manufactured outside the province but inflowing into the province according to the input–output table preparation method, which indicates the domestic product flow between provinces. Net outflow is equal to domestic outflows from Guizhou Province minus domestic inflows into Guizhou Province.

The economic shock of the agricultural sector, arising from the impacts of climate change on crop yield in Guizhou Province, varies by subsector. (1) Rice subsector: From 2021 to 2030, the climate change scenario S1, characterized by slight temperature rise and little precipitation, would increase the rice yield value by 0.355%, raise the price by 0.972%, and reduce the net outflow by 2.547%. In other words, the domestic inflows of rice into Guizhou would increase, and rural residents’ consumption expenditure would decrease. The other three climate change scenarios would increase the rice yield value, reduce rice prices, increase net outflows, and raise rural residents’ consumption expenditure. However, from 2031 to 2050, with the increase in temperature, precipitation, and atmospheric CO₂ concentration, all four climate change scenarios would decrease the rice yield value, raise rice prices, decrease net outflows, and reduce rural residents’ consumption expenditure. Scenario S1, characterized by a slight temperature rise and little precipitation, would decrease the rice yield value by 0.377%, raise rice prices by 1.232%, decrease net outflows by 4.176%, and decrease rural residents’ consumption expenditure by 0.569%. (2) Corn subsector: From 2021 to 2050, all four climate change scenarios would be adverse for the corn subsector economy. From 2021 to 2030, climate change scenario S1, characterized by a slight temperature rise and little precipitation, would decrease the corn yield value by 0.126%, raise corn prices by 1.776%, and decrease the net outflows by 8.248%. In other words, the domestic inflows of corn into Guizhou would increase, and rural residents’ consumption expenditure would decrease. With the aggravation of climate change, from 2031 to 2050, climate change scenario S4, characterized by the high temperature and frequent rain, would decrease the corn yield value by 0.089%, raise corn prices by 1.408%, decrease net outflows by 3.085%, and decrease rural residents’ consumption expenditure by 0.631%. (3) Tea leaf subsector: From 2021 to 2050, all four climate change scenarios would increase the tea leaf yield value, reduce tea leaf prices, increase the net outflows, and increase rural residents’ consumption expenditures. Among them, in 2021–2030 and 2031–2050, climate change scenario S4 would increase the tea leaf yield value by 0.488% and 1.151%, and decrease tea leaf prices by 0.47% and 0.574%, respectively. The increase in net outflow would go up from 1.423% to 2.666%. The growth rate of rural residents’ consumption expenditure would go up from 1.069% to 2.872%.

3.2.3. Impacts of Crop Yield Changes on Macroeconomy and Rural Household Economy

Table 5. Impacts of Climate Change on Macroeconomy and Rural Household Economy (%).

Scenarios	GDP		Rural Residents’ Labor Income		Rural Resident’s Savings
	2021–2030	2031–2050	2021–2030	2031–2050	2021–2030	2031–2050
S1	0.021	0.095	4.80	1.10	−0.50	−0.57
S2	0.007	0.026	−1.20	−2.10	0.17	−0.43
S3	0.010	0.017	−2.30	−0.40	0.06	−0.36
S4	0.011	0.063	−2.20	3.40	0.16	−0.42

shows the simulation results of impacts on the macroeconomy and the rural household economy as a result of future climate change on the crop yield value in Guizhou Province. In general, the impacts of interdecadal climate change on rice, corn, and tea leaf production in Guizhou Province from 2021 to 2050 will contribute to the increase in GDP growth. Under comparison of the four climate change scenarios in different periods, climate change scenario S1, characterized by slight temperature rise and little precipitation, would be the most conducive to economic growth and increase GDP by 0.021% after influencing crop growth, while climate change scenario S2, characterized by temperature rise and little precipitation, would increase GDP by 0.007% after influencing crop growth. In 2031–2050, climate change scenario S3, characterized by high temperature and little precipitation, would increase the GDP by 0.017%, while climate change scenario S1, characterized by slight temperature rise and little precipitation, would still be the most favorable to economic growth and increase the GDP by 0.095%.

Moderate climate change is useful to increase rural household income, while drastic climate change is not useful to increase rural household income. In 2021–2030, climate change scenario S1, characterized by temperature rise and little precipitation, would raise rural residents’ income by 4.8% and slightly decrease the savings by 0.5%. In 2031–2050, climate change scenario S1 would increase rural residents’ income by 1.1% and decrease the savings by 0.57%. Climate change scenarios S2 and S3 would decrease rural residents’ income in Guizhou Province from 2021 to 2050, but both scenarios would at first increase rural residents’ savings and then decrease them. Climate change scenario S4 would decrease rural residents’ income by 2.2% in 2021–2030 and increase them by 3.4% in 2031–2050. On the contrary, S4 would at first increase savings by 0.16% and then decrease them by 0.42%.

4. Discussion

4.1. Climate Change May Continue the Trend of Crop Structure Changes in Ecologically Vulnerable Regions

On the whole, the impacts of future climate change on the agricultural economy in Guizhou Province will be positive, but moderate climate change will be significantly more favorable for economic growth than dramatic climate change. Moreover, structural differences will exist in the economic impact of future climate change on the agriculture sector: future climate change will have negative impacts on the rice subsector and corn subsector and positive impacts on the tea leaf subsectors.

Although the impacts of climate change on the yields of these three types of crops in Guizhou Province lead to small changes in the agricultural sector economy and rural household economy, the direction of changes has implications for ecologically vulnerable regions. Firstly, future climate change will make the grain crop yield fluctuate up and down, which will aggravate the economic risk of grain planting. Coupled with the factors of low comparative benefits of grain planting, this may continue the decreasing trend of grain planting areas in ecologically vulnerable regions since the reform and opening-up. Secondly, future climate change will be useful to increase the yield of cash crops, which can partially hedge the future market risk of cash crops. If the future market price of cash crops is higher than the break-even point in the future, favorable climate resources will continue the trend of growing cash crops in ecologically vulnerable regions since the reform and opening-up. Scale expansion of perennial cash crops (such as tea trees) will mainly take place in mountainous land or hilly slope land, which are not suitable for food crop production. For this reason, the agricultural economy development of ecologically vulnerable regions should continue with such an eco-space layout: making scientific use of the climatic resources of ecologically vulnerable regions, dedicating limited plains and dam cultivated land to grain crop planting, and dedicating sloping fields to perennial crop planting.

4.2. Climate Change Leads to More Difficult Grain Support in Ecologically Vulnerable Regions

The increase in atmospheric CO₂ concentration, temperature rise, and complex and ever-changing precipitation may increase or decrease rice yield, thereby aggravating the uncertainty and instability of rice production in Guizhou Province. However, the increase in atmospheric CO₂ concentration and warming will decrease corn yield regardless of the changes in precipitation. If the corn price cannot rise above the break-even point in the future, the market risk will be combined with the natural risk. In this case, the decreasing trend of corn planting areas since the reform and opening-up will continue.

As for regional grain security, emphasis should be placed on a problem: future climate change will be more conducive to cash crop production and may induce ecologically vulnerable regions to continue the trend of crop structure change, i.e., the “sown area of grain crops decline but sown areas of cash crops rises”. Future climate change, in conjunction with crop structure change (“sown area of grain crops decline but sown areas of cash crops”) will reduce grain yield and call for more external imports. For example, from 1995 to 2017, the proportion of grain produced by rural households in Guizhou Province dropped from 80.44% to 42.71% of grain obtained through self-production, procurement, and loans. The proportion of purchased grain soared from 9.15% to 57.29% (Wei Xiaoyun et al. [40]). The increase in the proportion of grain procurement undoubtedly aggravates the market risk of grain supply guarantees for rural residents in ecologically vulnerable regions. Moreover, it should be emphasized that the scarce and low-quality cultivated land resources in ecologically vulnerable regions, the confined natural conditions for grain production, and the limited contribution of grain storage in farmland to the increase in total grain yield will increase the difficulty of grain security.

4.3. The Expansion of Perennial Crop Planting Area Will Expand the Eco-Space for the Future Economic and Social Development in Ecologically Vulnerable Regions

The expansion of perennial crop planting areas will improve the ecological environment in ecologically vulnerable regions. Hilly, mountainous lands and plateaus account for more than 80% of the land in Guizhou, Yunnan, and Guangxi Provinces, but the local hydrothermal conditions are favorable. Moreover, the three provinces are the key areas of corn structure adjustment for the “Sickle Bend” Region in China. After continuous fallowing or conversion of sloping cropland at a slope gradient of less than 25° and barren land for two seasons, with reference to the national policy of returning farmland to forest or grassland, tea trees and other perennial plants, characterized by wind break, sand fixation, water conservation, and protection of tilling layer, are planted, thereby gaining ground in mountainous agriculture and continuously improving the ecological environment. After perennial crops are planted, they interact with the underlying plants. On this basis, dry branches, fallen leaves, and humus fully ferment and merge into the soil and become the active components of solidified soil and increase the water-holding capacity of the soil. The developed root system can alleviate rain wash on the ground during rainfall periods. Eventually, this improves the vegetation coverage, increases the humidity of the air, and adjusts the local microclimate so as to improve the ecological environment. A more adequate eco-space will be available for the future economic and social development of ecologically vulnerable regions under China’s grand goal of a carbon dioxide emission peak before 2030 and carbon neutrality before 2060.

5. Conclusions

Since the agricultural risks caused by climate change are increasingly becoming an important factor in people’s decision-making, when the crop planting structure is adjusted in ecologically vulnerable regions, it is necessary to take into account economic effects and keep a balance between grain security and the ecosystem. Therefore, it is of great theoretical and practical significance to identify the risks and opportunities for ecologically vulnerable regions when adjusting the crop structure, with a view to better furnishing the information for the development of crop structure adjustment strategies.

In order to better identify these risks and opportunities, this paper analyzes the methods used in previous studies in detail and proposes applying a crop model coupled with an economic model for the purpose of analysis, evaluation, and prediction. In order to assess the impacts of future climate change on crop yield, a crop model is used for simulation evaluation in this paper. To assess the economic impacts of such yield impact, the CGE model is used for simulation evaluation in this paper. According to the main research findings, in the ecologically vulnerable regions, compared with food crops, cash crops will benefit more from future climate change in terms of yield, which may continue the trend of crop structure change in the ecologically vulnerable regions and aggravate the difficulty of grain security in the ecologically vulnerable regions. However, vice versa, this will expand the eco-space for the future economic and social development of ecologically vulnerable regions. Therefore, on the one hand, attention should be paid to grain security, while the planting area of cash crops should not be excessively expanded. On the other hand, it is advised to plant perennial cash crops in those areas which are unsuitable for planting cash crops.

It should be noted that in this paper, the time frame for simulation is only 30 years (2021–2050), which is a very short period of time for the physical evolution of the climate, with limited effects of climate change. Greater impacts on the agricultural economy of ecologically vulnerable regions will gradually come out in the future, for which further studies are needed.