Assessing the Environmental Impact of Oasis Agriculture in the Yarkant River Basin: A Comprehensive Study of Water Use, Carbon Footprint, and Decoupling Index

Abstract

Studying the relationship between grain planting and the environment is an important means to promote sustainable production. This study takes wheat, a typical grain crop in the Yarkant River oasis irrigation district, the fourth largest agricultural irrigation district in China, as an example to analyze the relationship and changing trends between wheat yield and water footprint (WF), and carbon footprint (CF) from 2001 to 2020. The study found that during the research period, wheat yield, ${WF}_{green,blue},{WF}_{grey}$, and CF showed a fluctuating but significantly upward trend. Decoupling analysis indicates that the overall decoupling trend between wheat yield and water footprint and carbon footprint is not obvious. This suggests that the rapid development of wheat production in the Yarkant River Oasis has also led to significant water resource consumption, pollution, and greenhouse gas emissions. Among the three sub–irrigation districts, the Shache sub–irrigation district has the best decoupling state, reflecting that the increase in wheat yield in Shache did not lead to more water resource consumption and pollution which is may due to its abundant water resources and agriculture development. Further analysis found that the use of nitrogen fertilizers and irrigation electricity have contributed to water resource pressure and greenhouse gas emissions. This study reveals that there are significant environmental risks in the current wheat planting in the Yarkant River oasis irrigation district, but it also points out the direction for green development in the irrigation district.

1. Introduction

Over the past century, Earth’s temperature has been continuously rising. By the end of this century, the global average temperature could rise by 1.8 °C to 4.0 °C. Global climate change, characterized mainly by warming, may lead to the collapse of many ecosystems [1,2]. Against the backdrop of climate change, the water resource crisis, exacerbated by water shortages, is further deteriorating ecosystems [3,4]. Agriculture is a major contributor to greenhouse gas emissions and water resource usage. Agricultural production accounts for 84% of the global nitrous oxide emissions and more than 70% of the global water usage [5,6]. For a long time, agricultural development has led to resource depletion and environmental pollution, accelerating global warming and increasing water resource pressure [7]. Crop ecosystems are also adversely affected by this feedback loop, posing a threat to food security. To actively respond to climate change, ensure food security, and achieve sustainable agricultural development, irrational agricultural practices need to be changed. Therefore, it is essential to specifically understand the current state and relationship of the environmental impact of regional crop production, which should be a priority concern for sustainable development decision making.

To understand the potential impact of human activities on natural resource consumption and the environment, the concept of “footprint” was introduced [8]. The water footprint, for instance, is an indicator used to assess the amount of water required by human activities and the resulting water pollution, and it has been widely applied in water resource management [9]. The carbon footprint, on the other hand, is an indicator used to measure the impact of an activity on climate change, typically expressed as the mass of CO₂ equivalent generated by the 100–year global warming potential (GWP) [10]. Both carbon and water footprints have been extensively used in crop production analysis. The current research on crop production footprints generally focuses on revealing the spatial and temporal variations in footprints; comparing crop production differences between regions, analyzing the influencing factors and drivers of crop production, and evaluating resource utilization efficiency and optimization, these studies provide a theoretical foundation for ensuring green agricultural development [11,12,13,14,15]. However, compared to the analysis of a single footprint indicator, selecting and combining multiple footprint indicators can more comprehensively represent resource consumption and environmental pressure from different aspects. Despite this, research combining multiple footprints is still relatively scarce compared to studies on single footprints [16]. Combining water and carbon footprints in crop production forms a footprint family, allowing for a multi–faceted assessment of the environmental impact and pressure caused by regional crop production. This approach can, to some extent, compensate for the limitations of the footprint research domain.

In its 2011 report “Decoupling Natural Resource Use and Environmental Impacts from Economic Growth,” the United Nations Environment Programme (UNEP) introduced the concept of decoupling natural resource use and environmental impacts from economic growth [17]. The aim is to address global environmental issues and improve human well–being. Since decoupling analysis can dynamically reflect the interactions and changing characteristics between variables over a certain period, it has been widely applied in studies related to economic growth, energy use, and environmental change [18,19]. Under the “Triple Bottom Line” of economic, environmental, and social sustainability, conflicts may arise between “strong sustainability” (i.e., achieving economic benefits without environmental costs) and “weak sustainability”. The states defined by decoupling indicators can quantitatively measure this information and translate it into an easily understandable form [20]. Using decoupling analysis as a supplement to studies on crop water and carbon footprints, and constructing a comprehensive index of crop production, water resource consumption, and greenhouse gas emissions can improve the crop production environmental impact assessment system. It can also evaluate the sustainability of crop production and provide a more comprehensive scientific basis for regional agricultural sustainable development decision making.

China is a major agricultural country, feeding 22% of the world’s population with only 7% of the world’s arable land, making it a significant force in maintaining global food security [21]. However, the agricultural sector in China consumes 61.5% of the nation’s total water usage, and agricultural greenhouse gas emissions account for 24% of the country’s total emissions [22]. The resource and environmental pressures brought by agriculture may increase [23,24]. China, fully considering the shared future of the global community, actively promotes the global green environmental protection process and has proposed the long–term “dual carbon” strategy. Therefore, understanding whether crop production comes at an excessively high environmental cost is crucial for mitigating climate change, rationally utilizing water resources, promoting ecological restoration, and ensuring the long–term stable development of agricultural production.

The Yarkant River Basin oasis is located in the arid region of northwestern China and is the fourth largest agricultural irrigation area in the country. This irrigation area is not only an important part of China’s “granary” but also a crucial node in the “Belt and Road Initiative”. The agricultural development in this region plays a significant role in food security and regional economic growth. However, the impacts of climate change and the driving forces of economic development add more pressure and uncertainty to agricultural development in the ecologically fragile arid regions of northwest China [25]. At the same time, there is currently limited research on the sustainability of crop production in the oasis irrigation areas of northwest China, and the relationship between crop production development and environmental impact changes in these regions remains unclear. Therefore, we selected the Yarkant River oasis irrigation area as the target study area. We aim to assess the current status of wheat production, a typical crop in this region, from the perspectives of water resource consumption and greenhouse gas emissions. Additionally, we evaluate the decoupling degree between crop production and environmental impact. We hope this study can provide effective references for future water saving, emission reduction, and sustainable development in the study area and other regions with similar conditions.

Specifically, this study will evaluate the water footprint and carbon footprint of wheat production in the Yarkant River oasis over the past 20 years. By applying the decoupling theory, we will explore the relationship between the development of wheat production and changes in environmental resources. Through analyzing the main driving forces and influencing factors affecting these footprints, we aim to provide scientifically based evidence and recommendations for sustainable development and agricultural ecological civilization construction in the region. This will be carried out at the scale of the basin irrigation area, considering the “Triple Bottom Line” of economic, environmental, and social aspects.

2. Materials and Methods

2.1. Overview of the Study Area

The study area is the oasis irrigation district of the Yarkant River Basin (hereinafter referred to as the “study area”), located in the southwestern part of Xinjiang Uygur Autonomous Region, on the northwestern frontier of China. It lies at the western edge of the Tarim Basin, geographically positioned between 74°28′–80°54′ E and 34°50′–40°31′ N. The study area is bordered by the Karakoram Mountains to the south, Yuepuhu to the north, the Taklamakan Desert to the east, and the Heizi Gobi and Buguli Desert to the west. This region features a warm temperate continental climate, marked by scarce precipitation, intense evapotranspiration, and perennial arid conditions, with an average annual precipitation of 47.6 mm and an evaporation rate of 2.23 × 10⁴ mm per year. The geographical location of the study area is shown in Figure 1.

Situated in the mid–to–lower reaches of the Yarkant River Basin’s alluvial plain, the study area is the largest irrigation district in Xinjiang and the fourth largest in China, with an irrigation area of 4.4 × 10⁵ hectares. Administratively, it includes counties such as Shache, Bachu, and Maigaiti. The region enjoys long sunshine hours and abundant light and heat resources, providing favorable conditions for agricultural production, making it a prime base for high–quality cotton and grain production in China. Agriculture, particularly crop farming, is the dominant and pillar industry of the local economy. Wheat is the most widely planted and highest–yielding grain crop in the region. In recent years, driven by population growth and economic development, agricultural production in the study area has expanded significantly, increasingly impacting the environment and water resources. This has posed even greater challenges to the already fragile ecosystem.

2.2. Data Sources

This study covers the period from 2001 to 2020. The meteorological data, including monthly average precipitation, sunshine hours, relative humidity, wind speed, and temperature for Shache, Bachu, and Maigaiti, were sourced from the China Meteorological Data Service Center http://data.cma.cn (accessed on 10 June 2024). Crop data, such as the wheat growth cycle and crop coefficients, were obtained from the Food and Agriculture Organization (FAO) https://www.fao.org/ (accessed on 10 June 2024). Statistical data related to wheat production were sourced from the “Xinjiang Statistical Yearbook”.

2.3. Research Methods

2.3.1. Footprint Indicator Evaluation

Water Footprint of Wheat Production

The water footprint of wheat production refers to the total amount of water used in the wheat production process. It is divided into blue water footprint, green water footprint, and gray water footprint based on the type of water resource usage. The blue water footprint refers to the total amount of surface– and groundwater resources consumed during the growth and development of wheat, representing the water used for artificial irrigation. The green water footprint is the amount of water consumed through transpiration or absorbed by the roots of wheat, reflecting effective precipitation. The gray water footprint is the amount of water required to dilute specific pollutants to harmless levels, indicating the degree of water resource pollution during the production process [26]. The specific calculation process is as follows:

$$WF=W{F}_{blue}+W{F}_{green}+W{F}_{grey}$$

(1)

In the formula,

WF is the total water footprint of wheat, measured in cubic meters (m³);

WF_blue is the blue water footprint of wheat, measured in cubic meters (m³);

WF_green is the green water footprint of wheat, measured in cubic meters (m³);

WF_grey is the gray water footprint of wheat, measured in cubic meters (m³).

(1)

Blue Water Footprint and Green Water Footprint of Wheat Production

$$E{T}_C=K_{C}E{T}_0$$

(2)

In the formula,

ET_C is the evapotranspiration during the wheat growth period, measured in millimeters (mm);

K_C is the crop coefficient;

ET₀ is the reference crop evapotranspiration, measured in millimeters (mm), which is simulated and calculated using CROPWAT 8.0.

$$W{F}_{green}=10\times {\displaystyle \sum _{d=1}^{\lg p}E{T}_{green}}\times A$$

(3)

$$ETgreen=min(E{T}_C,P_{eff})$$

(4)

In the formula,

10 is the unit conversion factor between mm and m³/hm²;

A is the planting area, measured in hectares (hm²);

lgp is the length of the wheat growth period, measured in days;

ET_green is the green water evapotranspiration of wheat, measured in millimeters (mm);

P_eff is the effective rainfall, measured in millimeters (mm), which is simulated and calculated using CROPWAT 8.0.

Other variables are as previously described.

$$W{F}_{blue}=10\times {\displaystyle \sum _{d=1}^{\lg p}E{T}_{blue}}\times A$$

(5)

$$E{T}_{blue}=max(0,E{T}_C-P{e}_{ff})$$

(6)

In the formula,

ET_blue is the blue water evapotranspiration of wheat, measured in millimeters (mm).

The other variables are as previously described.

(2)

Grey Water Footprint of Wheat Production

Leaching or runoff of nutrients from farmland is a major cause of non-point source pollution in surface- and groundwater. Nitrogen is one of the most important nutrients in agricultural ecosystems. In China’s arid regions, more than 65% of the nitrogen applied to crop cultivation is in pure form, and only about 35% of the nitrogen is absorbed and utilized by crops [27]. Therefore, this study quantifies the gray water footprint associated with nitrogen use:

$$W{F}_{grey}=[(\alpha ·F_{SN})/(C_{max}-C_{nat})]\times A$$

(7)

In the formula,

α is the runoff leaching rate of nitrogen, with a value of 0.1 based on experimental data on nitrogen leaching rates [28];

F_SN is the amount of nitrogen fertilizer applied per unit area in pure form, measured in kilograms per hectare (kg/hm²);

C_max is the maximum acceptable concentration of nitrogen in the water body, with a value of 0.01 kg/m³ according to the standard concentration limit for nitrates (measured as nitrogen) in the “Environmental Quality Standards for Surface Water” (GB3838-2002) [29];

C_nat is the natural background concentration of nitrogen.

The other variables are as previously described.

Carbon Footprint of Wheat Production

The carbon footprint of wheat production is an indicator used to measure the potential climate forcing caused by the total greenhouse gas emissions during the wheat production process. It mainly includes direct emissions related to agricultural inputs and indirect emissions caused by NO₂ during cultivation. The calculation process is as follows [30]:

$$CF=C{F}_{N_{2}O}+C{F}_{INPUT}$$

(8)

In the formula,

CF is the total carbon footprint of wheat, measured in kilograms of CO₂ equivalent (kg CO₂-eq);

CF_N2O is the carbon footprint caused by N₂O emissions, measured in kilograms of CO₂ equivalent (kg CO₂-eq);

CF_INPUT is the carbon footprint caused by agricultural inputs, measured in kilograms of CO₂ equivalent (kg CO₂-eq).

The other variables are as previously described.

(1)

Calculation of Carbon Footprint Caused by N₂O Emissions

$$C{F}_{N_{2}O}=(C{F}_{direct,N_{2}O}+C{F}_{indirect,N_{2}O})\times A$$

(9)

$$C{F}_{direct,N_{2}O}=F_{SN}\times E{F}_D\times {\displaystyle \raisebox{1ex}{$44$}\!\left/ \!\raisebox{-1ex}{$28$}\right.}\times GW{P}_{N_{2}O}$$

(10)

$$C{F}_{indirect,N_{2}O}=F_{SN}\times [FRA{C}_L\times E{F}_L\times FRA{C}_S\times E{F}_S]\times {\displaystyle \raisebox{1ex}{$44$}\!\left/ \!\raisebox{-1ex}{$28$}\right.}\times GW{P}_{N_{2}O}$$

(11)

In the formula,

CF_{direct, N2O} is the carbon footprint caused by direct soil emissions per unit area, measured in kilograms of CO₂ equivalent per hectare (kg CO₂-eq/hm²);

CF_{indirect, N2O} is the carbon footprint caused by indirect soil emissions per unit area, measured in kilograms of CO₂ equivalent per hectare (kg CO₂-eq/hm²);

F_SN is the amount of nitrogen fertilizer applied per unit area in pure form, measured in kilograms per hectare (kg/hm²);

44/28 is the conversion factor from nitrogen (N) to nitrous oxide (N₂O);

GWP_N2O is the 100-year global warming potential of N₂O;

According to the “Low Carbon Development and Provincial Greenhouse Gas Inventory Compilation Training Materials” [31], EF_D is the emission factor for N₂O emissions from nitrogen inputs to soil, with a value of 0.0056;

FRAC_L is the fraction of nitrogen lost through leaching and runoff, with a value of 20%;

EF_L is the emission factor for N₂O emissions from nitrogen leaching and runoff, with a value of 0.0075;

FRAC_S is the fraction of nitrogen volatilized from fertilizer applied to farmland, with a value of 10%;

EF_S is the emission factor for N₂O emissions from nitrogen volatilization and deposition, with a value of 0.01.

The other variables are as previously described.

(2)

Calculation of Carbon Footprint Caused by Agricultural Inputs

$$C{F}_{INPUT}={\displaystyle \sum _{i=1}^n{N}_i}\times E_{carbon,i}\times A$$

(12)

where

N_i is the amount of the ith type of agricultural input per unit area, measured in kilograms per hectare (kg/hm²);

E_carbon,i is the carbon emission factor for the ith type of agricultural input, detailed in

Table 1. Emission parameters of agricultural inputs for wheat production.

Agricultural Inputs	Carbon Emission Factor
wheat seeds	0.58 kg CO2–eq kg−1
diesel fuel	0.89 kg CO2–eq kg−1
irrigation in northwest China consumes electricity	0.97 kg CO2–eq 3.6 MJ−1
nitrogenous fertilizer	1.53 kg CO2–eq kg−1
Phosphate fertilizer	1.63 kg CO2–eq kg−1

;

A is the planting area, measured in hectares (hm²), and the other variables are as previously described [32,33].

2.3.2. Tapio Decoupling Analysis Indicators

In this study, wheat yield is selected as the indicator to represent the development of wheat production, while the water footprint and carbon footprint are used as the indicators of environmental impact. The Tapio decoupling model is employed to analyze the decoupling between wheat production development and changes in environmental impact, as follows [34,35]:

$$D_{WF-Y}={\displaystyle \frac{\%\Delta WF}{\%\Delta {\rm Y}}}={\displaystyle \frac{\raisebox{1ex}{$W{F}_j$}\!\left/ \!\raisebox{-1ex}{$W{F}_{j-1}-1$}\right.}{\raisebox{1ex}{${{\rm Y}}_j$}\!\left/ \!\raisebox{-1ex}{${{\rm Y}}_{j-1}-1$}\right.}}$$

(13)

$$D_{CF-Y}={\displaystyle \frac{\%\Delta CF}{\%\Delta {\rm Y}}}={\displaystyle \frac{\raisebox{1ex}{$C{F}_j$}\!\left/ \!\raisebox{-1ex}{$C{F}_{j-1}-1$}\right.}{\raisebox{1ex}{${{\rm Y}}_j$}\!\left/ \!\raisebox{-1ex}{${{\rm Y}}_{j-1}-1$}\right.}}$$

(14)

In the formula,

D_WF-Y and D_CF-Y are the decoupling indices between wheat yield and water footprint, and between wheat yield and carbon footprint, respectively;

%ΔY, %ΔWF, and %ΔCF are the rates of change in wheat yield, water footprint, and carbon footprint, respectively, between the current period and the base period;

Y is the yield;

j is the current time period, and (j − 1) is the base time period. In this study, the time interval is one year.

The other variables are as previously described. Based on previous research and the actual situation of this study, the decoupling relationships are divided into six categories [7,36], with each decoupling state explained in

Table 2. The relationship and degree of decoupling between wheat yield and footprint.

Decoupling Status	The Relationship Between WF and Y	The Relationship Between CF and Y
strong decoupling	$∆WF\le 0, ∆Y>0, ∆D_{WF-Y}\le 0$	$∆CF\le 0, ∆Y>0, D_{CF-Y}\le 0$
weak decoupling	$∆WF>0, ∆Y>0, 0<∆D_{WF-Y}<1$	$∆CF>0, ∆Y>0, 0<D_{CF-Y}<1$
expansive coupling	$∆WF>0, ∆Y>0, ∆D_{WF-Y}\ge 1$	$∆CF>0, ∆Y>0, ∆D_{CF-Y}\ge 1$
recessive coupling	$∆WF<0, ∆Y<0, ∆D_{WF-Y}\ge 1$	$∆CF<0, ∆Y<0, ∆D_{CF-Y}\ge 1$
weak negative decoupling	$∆WF<0, ∆Y<0, 0<∆D_{WF-Y}<1$	$∆CF<0, ∆Y<0, 0<∆D_{CF-Y}<1$
strong negative decoupling	$∆WF\ge 0, ∆Y<0, ∆D_{WF-Y}\le 0$	$∆CF\ge 0, ∆Y<0, ∆D_{CF-Y}\le 0$

.

In decoupling analysis, the stability coefficient can be used to evaluate the stability of the decoupling state. The lower its value, the lower the volatility of the decoupling:

$$SD={\displaystyle \frac{1}{n-1}}{\displaystyle {\sum }_j^n\left|\frac{x_{j+1}-x_j}{x_j}\right|}$$

(15)

In the formula,

SD is the stability coefficient;

x_j+1 and x_j represent the decoupling indices of the subsequent and preceding stages, respectively;

n is the number of samples of the decoupling index.

3. Results

3.1. Evaluation of Crop Yield and Footprint Indicators from 2001 to 2020

The changes in wheat yield, ${WF}_{green,blue}$, ${WF}_{grey},$ and CF in the Yarkant River plain oasis from 2001 to 2020 are shown in Figure 2. It can be clearly seen from the figure that wheat production in the Yarkant River plain oasis experienced fluctuating growth during the study period. The annual average increases in yield, ${WF}_{green,blue}$, ${WF}_{grey},$ and CF were 7.18 × 10³ tons, 2.78 × 10⁶ m³, 4.17 × 10⁶ m³, and 7.72 × 10⁶ kg CO₂ eq, respectively. In 2016, both the yield and environmental footprints reached their maximum values within the study period, with yield and ${WF}_{green,blue}$, ${WF}_{grey},$ and CF being 5.15 × 10⁵ tons, 5.77 × 10⁸ m³, 2.84 × 10⁸ m³, and 3.03 × 10⁸ kg CO₂ eq, respectively. The results of linear regression tests indicate that the wheat yield and footprint indicators in the Yarkant River plain oasis showed a significant upward trend from 2001 to 2020 (p < 0.05). Clearly, wheat production experienced rapid development over the 20 years, causing varying degrees of impact on water resources, the water environment, and greenhouse gas emissions.

3.2. Decoupling of Crop Yield and Footprint Indicators

3.2.1. Decoupling of Overall Irrigation District Crop Yield and Footprint Indicators

Over the past 20 years, crop yield and water and carbon footprints have served as indicators of the economic and environmental benefits of crop production. Based on the Tapio decoupling theory, we further analyzed the relationship between the wheat production footprint and yield in the oasis irrigation district of the Yarkant River Basin (

Table 3. Decoupling of wheat yield from footprint.

The Decoupling Relationship Between Yield and ${\mathit{W}\mathit{F}}_{\mathit{g}\mathit{r}\mathit{e}\mathit{e}\mathit{n},\mathit{b}\mathit{l}\mathit{u}\mathit{e}}$
Period	$\mathit{\%}\Delta \mathit{Y}$	$\mathit{\%}\Delta {\mathit{W}\mathit{F}}_{\mathit{g}\mathit{r}\mathit{e}\mathit{e}\mathit{n},\mathit{b}\mathit{l}\mathit{u}\mathit{e}}$	$\mathit{D}{\mathit{W}\mathit{F}}_{\mathit{g}\mathit{r}\mathit{e}\mathit{e}\mathit{n},\mathit{b}\mathit{l}\mathit{u}\mathit{e}}\mathit{Y}$	Degrees of Decoupling/Coupling
2001–2002	0.153	0.138	0.905	weak decoupling
2002–2003	−0.022	−0.137	6.164	recessive decoupling
2003–2004	−0.042	−0.048	1.129	recessive decoupling
2004–2005	0.125	0.069	0.548	weak decoupling
2005–2006	0.085	0.072	0.843	weak decoupling
2006–2007	−0.006	0.086	−14.493	strong coupling
2007–2008	0.028	−0.046	−1.657	strong decoupling
2008–2009	0.172	0.316	1.838	expansive coupling
2009–2010	0.102	−0.029	−0.287	strong decoupling
2010–2011	−0.119	−0.020	0.166	weak coupling
2011–2012	0.080	−0.023	−0.283	strong decoupling
2012–2013	0.018	0.037	2.108	expansive coupling
2013–2014	0.003	0.043	15.260	expansive coupling
2014–2015	0.116	0.077	0.662	weak decoupling
2015–2016	0.012	0.077	6.202	expansive coupling
2016–2017	−0.053	−0.136	2.566	recessive decoupling
2017–2018	−0.241	−0.094	0.391	weak coupling
2018–2019	0.096	−0.104	−1.073	strong decoupling
2019–2020	0.002	−0.033	−14.870	strong decoupling
The decoupling relationship between yield and ${WF}_{grey}$
Period	$\mathbf{\%}\Delta \mathit{Y}$	$\mathbf{\%}\Delta {\mathit{W}\mathit{F}}_{\mathit{g}\mathit{r}\mathit{e}\mathit{y}}$	$\mathit{D}{\mathit{W}\mathit{F}}_{\mathit{g}\mathit{r}\mathit{e}\mathit{y}}\mathit{Y}$	Degrees of Decoupling/Coupling
2001–2002	0.153	0.100	0.652	weak decoupling
2002–2003	−0.022	−0.064	2.904	recessive decoupling
2003–2004	−0.042	−0.009	0.218	weak coupling
2004–2005	0.125	0.084	0.670	weak decoupling
2005–2006	0.085	−0.010	−0.120	strong decoupling
2006–2007	−0.006	0.055	−9.183	strong coupling
2007–2008	0.028	0.039	1.417	expansive coupling
2008–2009	0.172	0.153	0.891	weak decoupling
2009–2010	0.102	0.110	1.081	expansive coupling
2010–2011	−0.119	−0.115	0.960	weak coupling
2011–2012	0.080	0.070	0.876	weak decoupling
2012–2013	0.018	0.052	2.970	expansive coupling
2013–2014	0.003	0.002	0.585	weak decoupling
2014–2015	0.116	0.087	0.750	weak decoupling
2015–2016	0.012	0.032	2.591	expansive coupling
2016–2017	−0.053	−0.053	0.992	weak coupling
2017–2018	−0.241	−0.203	0.840	weak coupling
2018–2019	0.096	0.144	1.491	expansive coupling
2019–2020	0.002	−0.003	−1.348	strong decoupling
The decoupling relationship between yield and CF
Period	$\mathbf{\%}\Delta \mathit{Y}$	$\mathbf{\%}\Delta \mathit{C}\mathit{F}$	$\mathit{D}\mathit{C}\mathit{F}\mathit{Y}$	Degrees of Decoupling/Coupling
2001–2002	0.153	0.034	0.223	weak decoupling
2002–2003	−0.022	−0.024	1.061	recessive decoupling
2003–2004	−0.042	−0.036	0.854	weak coupling
2004–2005	0.125	0.088	0.703	weak decoupling
2005–2006	0.085	0.021	0.253	weak decoupling
2006–2007	−0.006	0.047	−7.807	strong coupling
2007–2008	0.028	0.075	2.709	expansive coupling
2008–2009	0.172	0.189	1.101	expansive coupling
2009–2010	0.102	0.118	1.163	expansive coupling
2010–2011	−0.119	−0.029	0.240	weak coupling
2011–2012	0.080	0.108	1.351	expansive coupling
2012–2013	0.018	0.099	5.603	expansive coupling
2013–2014	0.003	−0.116	−41.513	strong decoupling
2014–2015	0.116	0.466	4.017	expansive coupling
2015–2016	0.012	0.010	0.818	weak decoupling
2016–2017	−0.053	−0.073	1.374	recessive decoupling
2017–2018	−0.241	−0.160	0.662	weak coupling
2018–2019	0.096	−0.001	−0.007	strong decoupling
2019–2020	0.002	0.135	60.106	expansive coupling

). This analysis aims to obtain a dynamic understanding of the relationship between the demands of crop industry development and ecological resource usage. These results can provide a basis for future agricultural production decision making in the development of oasis areas in arid regions.

In terms of water resource consumption, during the study period, strong decoupling DWF_green,blueY occurred most frequently, with five instances. This decoupling state, which mainly appeared after 2007 (2007–2008, 2009–2010, 2011–2012, and 2018–2020), indicates a decrease in water resource consumption while crop yield increases, representing the optimal state for water use in wheat production. Expansive coupling, a transitional state indicating the almost synchronous growth of water consumption and yield, occurred four times (2008–2009, 2012–2014, and 2015–2016). This state warns of a potential higher risk of water scarcity due to wheat production. Weak decoupling, which occurred four times (2001–2002, 2004–2006, and 2014–2015), indicates that yield increases come at the cost of higher water consumption, although the rate of yield increase is faster than that of water consumption. If this state persists long–term, it could lead to excessive water use in agricultural production.

The decoupling state during the periods of yield increase suggests that while wheat production poses some risk of water scarcity, there is also potential to control excessive water consumption. Conversely, during the periods of yield decline, recessive decoupling occurred three times, the most frequent state during the study period. This transitional state, indicating almost synchronous reductions in water consumption and yield, occurred twice in the early years of the study period (2002–2004) and once in the later period (2016–2017). Strong coupling, which occurred once (2006–2007), represents an undesirable growth state where water consumption increases while yield decreases, reflecting the worst scenario for both economic and ecological aspects. Weak coupling occurred twice (2010–2011 and 2017–2018), symbolizing reductions in both yield and water consumption, with yield decreasing at a faster rate.

Comparing the results between the periods of yield increase and decrease reveals that increased yield does not necessarily lead to greater water pressure, nor does decreased yield always mean reduced water issues. Improving water use efficiency in wheat production is a key focus for future development.

The decoupling results between the development of wheat production and the gray water footprint, which represents water pollution, indicate that during the periods of yield increase, the most frequent decoupling state was weak decoupling, occurring six times during the study period (2001–2002, 2004–2005, 2008–2009, 2011–2012, and 2013–2015). This indicates that the rate of yield increase was greater than the rate of water quality deterioration. While this production model may be acceptable in the short term, it is not conducive to water ecological health and sustainable development in the long term.

Expansive coupling occurred five times (2007–2008, 2009–2010, 2012–2013, 2015–2016, and 2018–2019). During the periods 2007–2010 and 2011–2016, the decoupling between wheat yield and gray water footprint alternated between weak decoupling and expansive coupling. The most favorable decoupling state for production and water environmental benefits, strong decoupling, occurred only twice during the study period (2005–2006 and 2019–2020). This suggests that the water environmental cost of wheat production is relatively high, and the negative impact on the water environment due to the pursuit of higher yields in regional agricultural production needs to be addressed.

During periods of yield decline, strong coupling and the transitional state of recessive decoupling each occurred once, in 2006–2007 and 2002–2003, respectively. Weak coupling was the predominant state, occurring in 2003–2004, 2010–2011, and 2016–2018. During these periods, water pollution improved, but the rate of improvement was slower than the rate of yield reduction.

Although the evaluation of WF_grey considered only the pollution caused by nitrogen fertilizer application during wheat production, the results of DWF_geryY still indicate that the water quality cost of wheat production development during the study period was relatively high, creating significant pressure on the water environment.

In the decoupling results between wheat production and greenhouse gas emissions, expansive coupling was the most frequent, occurring seven times (2007–2010, 2011–2013, 2014–2015, and 2019–2020), almost all in the mid– to late stages of the study period. During this time, the rapid development of wheat production and the almost synchronous increase in greenhouse gas emissions serve as a warning signal that wheat production is increasingly contributing to environmental and climate change.

During the periods of yield increase, weak decoupling occurred four times (2001–2002, 2004–2006, and 2015–2016). Compared to the rate of yield growth during the same periods, the rate of greenhouse gas emissions from regional wheat production increased even more. The strong sustainability state, or strong decoupling, occurred twice (2013–2014 and 2018–2019). The infrequent occurrence of decoupling in DCFY during the periods of yield increase, with most instances being weak decoupling, indicates the poor sustainability of wheat production during these periods and a significant contribution to climate change.

During the periods of yield decline, weak coupling occurred three times (2003–2004, 2010–2011, and 2017–2018), and strong coupling occurred once (2006–2007). This indicates that even during the periods of yield reduction, wheat production still poses a threat to the climate and environment. Therefore, the issue of excessive emissions caused by wheat production in the study area needs to be addressed.

Further examination of the stability of the three decoupling indicators reveals that during the period from 2001 to 2020, the stability coefficients for DWF_green,blueY, DWF_greyY, and DCFY were 4.491, 5.544, and 478.873, respectively. The decoupling trends all lack stability, especially for CF, which is highly unstable. This indicates a potential risk of high greenhouse gas emissions from future wheat production. If left unchecked, future wheat production is likely to cause significant water scarcity and exacerbate negative impacts on the ecosystem.

Overall, the environmental issues caused by wheat production over the past 20 years have been quite severe. Increased yields have led to greater water environmental degradation and environmental impact, while yield reductions have not effectively alleviated these impacts. The uncertainty in the changes in the three decoupling indicators throughout the entire study period suggests that future wheat production in the Yarkant River plain oasis should place greater emphasis on environmental protection while pursuing production development.

However, the occurrence of strong decoupling also demonstrates that sustainable development and environmentally friendly agriculture are achievable in the study area. Moving forward, it will be important to regulate production to ensure balanced development with ecological considerations.

3.2.2. Decoupling of Sub–Irrigation District Crop Yield and Footprint Indicators

Trend and Phase Analysis of Sub–Irrigation District Crop Yield and Footprint Indicators

To more specifically and thoroughly analyze the development of wheat production in the Yarkant River oasis and its environmental impact, this study conducted a trend analysis of wheat yield and footprint indicators in three sub–irrigation districts (Shache, Bachu, and Markit) based on the year 2001. The results are shown in Figure 3. It is clear that the impact and changes in wheat production development on different environmental aspects vary among the three sub–irrigation districts.

In the Shache irrigation district, wheat yield increased by 49% over 20 years, WF_green,blue decreased by 6%, while WF_grey and CF increased by 30% and 100%, respectively.

In the Bachu irrigation district, yield increased by 51%, with WF_green,blue, WF_grey, and CF increasing by 52%, 70%, and 110%, respectively.

In the Markit irrigation district, yield increased by 53%, with WF_green,blue, WF_grey, and CF increasing by 63%, 79%, and 175%, respectively.

From these results, it can be seen that except for a slight decrease in water resource consumption in Shache, all the sub–irrigation districts experienced varying degrees of increase in water pollution and greenhouse gas emissions. The environmental changes in the Shache sub–irrigation district were the best among the three, which may be related to its geographical location and the degree of agricultural modernization.

The Shache sub–irrigation district is located in the central part of the Yarkant River Basin and is the first sub–irrigation district after the Yarkant River flows out of the mountains. It has runoff within its boundaries and relatively abundant surface water resources, providing conditions for the development of oasis agriculture that relies almost entirely on irrigation. Additionally, the Shache sub–irrigation district has a longer history of agricultural development, a relatively well–established wheat planting system, and to some extent, has controlled the waste of resources and energy, making its agriculture more environmentally friendly.

In contrast, the Bachu and Markit sub–irrigation districts are located farther from the mountain outlet and do not have runoff within their boundaries. Their geographical and water resource conditions are less favorable compared to the Shache sub–irrigation district. In the pursuit of agricultural development, the consumption and loss of irrigation water have increased, along with greater water pollution.

Although the development levels of wheat production in the three sub–irrigation districts vary, the trend changes in yield development can be roughly divided into three stages: 2001–2010 (slow development period), 2010–2016 (rapid development period), and 2016–2020 (fluctuating development period). The average annual base change rate during the rapid development period was 173%, which is 1.341 times that of the slow development period and 1.607 times that of the fluctuating development period.

Phase–Wise Decoupling Analysis of Sub–Irrigation District Crop Yield and Footprint Indicators

Further, based on the segmentation of the development stages of wheat in each sub–irrigation district and the average values of the yield and footprint indicators for each stage, a decoupling analysis was conducted. The results are shown in

Table 4. The results of $D{\overline{WF}}_{green,blue}\overline{Y}$ and the decoupling status of each sub–irrigation area.

	Period	$\mathit{\%}\Delta \overline{\mathit{Y}}$	$\mathit{\%}\Delta {\overline{\mathit{W}\mathit{F}}}_{\mathit{g}\mathit{r}\mathit{e}\mathit{e}\mathit{n},\mathit{b}\mathit{l}\mathit{u}\mathit{e}}$	$\mathit{D}{\overline{\mathit{W}\mathit{F}}}_{\mathit{g}\mathit{r}\mathit{e}\mathit{e}\mathit{n},\mathit{b}\mathit{l}\mathit{u}\mathit{e}}\overline{\mathit{Y}}$	Degrees of Decoupling/Coupling
Shache	2001–2010	0.070	0.051	0.719	weak decoupling
	2010–2016	0.009	0.015	1.718	expansive coupling
	2016–2020	−0.050	−0.124	2.471	recessive decoupling
Bachur	2001–2010	0.068	0.059	0.868	weak decoupling
	2010–2016	0.020	0.046	2.275	expansive coupling
	2016–2020	−0.007	0.016	−2.490	strong coupling
Markit	2001–2010	0.054	0.027	0.506	weak decoupling
	2010–2016	0.057	0.097	1.712	expansive coupling
	2016–2020	−0.034	−0.035	1.021	recessive decoupling

. describes the dynamic relationship between water resource consumption and wheat yield in the three sub–irrigation districts at different stages of wheat production development. The results indicate that during the slow development period, all three sub–irrigation districts exhibited weak decoupling, where the growth rate of wheat yield was greater than that of WF_green,blue. This suggests that although wheat production had an impact on water resources during the slow development period, it was still within a controllable range.

During the rapid development period of wheat production, all three sub–irrigation districts showed expansive coupling, indicating that the rapid development of wheat production in the Yarkant River oasis led to greater water resource consumption. In the fluctuating development period, Shache and Markit exhibited synchronous reductions in water resource consumption and yield, while the more water–scarce Bachu experienced strong coupling. In Bachu, the decrease in wheat yield did not lead to a reduction in water resource consumption, continuing to pose a threat to water resources.

Table 5. The results of $D{\overline{WF}}_{grey}\overline{Y}$ and the decoupling status of each sub–irrigation area.

	Period	$\mathbf{\%}\Delta \overline{\mathit{Y}}$	$\mathbf{\%}\Delta {\overline{\mathit{W}\mathit{F}}}_{\mathit{g}\mathit{r}\mathit{e}\mathit{y}}$	$\mathit{D}{\overline{\mathit{W}\mathit{F}}}_{\mathit{g}\mathit{r}\mathit{e}\mathit{y}}\overline{\mathit{Y}}$	Degrees of Decoupling/Coupling
Shache	2001–2010	0.070	0.719	0.678	weak decoupling
	2010–2016	0.009	1.718	0.983	weak decoupling
	2016–2020	−0.050	2.471	0.540	weak coupling
Bachur	2001–2010	0.068	0.069	1.011	expansive coupling
	2010–2016	0.020	0.029	1.438	expansive coupling
	2016–2020	−0.007	0.006	−0.929	strong coupling
Markit	2001–2010	0.054	0.050	0.924	weak decoupling
	2010–2016	0.057	0.067	1.173	expansive coupling
	2016–2020	−0.034	−0.017	0.508	weak coupling

shows the decoupling relationship between wheat production development and WF_grey in the three sub–irrigation districts. The negative impact of wheat production on the water environment in Bachu was greater than in the other two sub–irrigation districts. Even during the slow development period, Bachu exhibited expansive coupling, a state that persisted into the rapid development period. During the fluctuating development period, the decrease in yield did not improve the water environment, with water pollution continuing to worsen. This indicates that wheat production in this region has caused irreversible negative impacts on the water environment.

In Shache and Markit, during the slow development period, both districts exchanged water environment degradation for yield growth that exceeded the rate of degradation. However, Shache maintained this “low water environment pollution for high yield” state into the rapid development period, while Markit transitioned to a state of almost synchronous growth in yield and water pollution. During the fluctuating period, the yield reduction rates in these two sub–irrigation districts were faster than the recovery rates of the water environment, providing limited ecological benefits.

Table 6. The results of $D\overline{CF}\overline{Y}$ and the decoupling status of each sub–irrigation area.

	Period	$\mathbf{\%}\Delta \overline{\mathit{Y}}$	$\mathbf{\%}\Delta \overline{\mathit{C}\mathit{F}}$	$\mathit{D}\overline{\mathit{C}\mathit{F}}\overline{\mathit{Y}}$	Degrees of Decoupling/Coupling
Shache	2001–2010	0.070	0.049	0.692	weak decoupling
	2010–2016	0.009	0.083	9.251	expansive coupling
	2016–2020	−0.050	−0.020	0.389	weak coupling
Bachur	2001–2010	0.068	0.011	1.626	expansive coupling
	2010–2016	0.020	0.250	1.236	expansive coupling
	2016–2020	−0.007	0.143	−2.182	strong coupling
Markit	2001–2010	0.054	0.042	0.789	weak decoupling
	2010–2016	0.057	0.218	3.844	expansive coupling
	2016–2020	−0.034	−0.018	0.517	weak coupling

breaks down the situation of wheat production and greenhouse gas emissions in the three sub–irrigation districts at different stages. The overall emission situation in Bachu was worse than in the other two sub–irrigation districts. Bachu experienced a transition from expansive coupling to strong coupling, where not only did yield increases bring corresponding greenhouse gas emissions, but yield reductions also failed to reduce emissions.

In terms of the decoupling indicator for carbon footprint and yield $D\overline{CF}\overline{Y}$, Shache and Markit followed a similar pattern, transitioning from weak decoupling to expansive coupling, and then to weak coupling. During the development process, emissions and yield failed to achieve a balance, with all three sub–irrigation districts facing the issue of excessive greenhouse gas emissions.

Overall, the decoupling results at different stages indicate that the environmental impact of wheat yield increases in the study area is significant in terms of water resources consumption and greenhouse gas emissions. However, yield reductions do not immediately lead to environmental recovery and may still impose stress on the environment, causing both ecological and economic losses. Among the three sub–irrigation districts, Bachu is in the worst situation. Currently, wheat production has already led to excessive resource consumption in this area, which is detrimental to sustainable ecological and economic development.

Reasonably controlling the excessive water resource consumption, water environment pollution, and greenhouse gas emissions caused by agricultural production is a critical issue that the wheat production system in the Yarkant River plain oasis needs to address.

4. Discussion

This study evaluated the relationship between wheat production and its environmental impact in terms of water resources and greenhouse gas emissions in the oasis irrigation districts of the arid northwest region. The findings revealed that from 2001 to 2020, there existed a close coupling relationship between wheat production and water resource consumption, water environment pollution, and greenhouse gas emissions. This reflects the high consumption and low efficiency issues, as well as the risks of unsustainability in the wheat production process.

It is noteworthy that decoupling relationships between the yield and footprint indicators did occur. Although the trend in strong decoupling was not prominent, it indicates that there is still potential to eliminate decoupling barriers and achieve green agricultural development. Therefore, we focused on and analyzed the components and driving factors of the footprint indicators, aiming to propose strategies to improve crop production, reduce water resource consumption, and mitigate climate change.

4.1. Wheat Production Water, Carbon Footprint Composition

Agricultural production in arid and semi–arid regions requires more irrigation water and agricultural inputs, which consequently leads to higher WF and CF. From the composition structure of ${WF}_{green}$, ${WF}_{blue}$, and ${WF}_{grey}$ in wheat production in the Yarkant River plain oasis (Figure 4), it can be seen that ${WF}_{blue}$ has the largest proportion, while ${WF}_{green}$ has the smallest. This is a result of the natural characteristics of the continental arid climate in the Yarkant River oasis and the region’s reliance on irrigation agriculture.

Previous studies have paid less attention to ${WF}_{grey}$ in this region, but this study found that approximately one–third of the water resource wastage during the wheat growth period in the irrigation district is due to pollution. This indicates that the water environment impact caused by human interference during the agricultural replacement of natural oases in this region cannot be underestimated.

Figure 5 shows the detailed composition and changes in CF during wheat production. As the production progresses, the proportion of CF from irrigation electricity $C{F}_{irrigation electricity}$ gradually increases, even surpassing ${CF}_{direct,N_{2}O}$. This suggests that climate conditions, the application of technology, and the growing demand for food have led to an increase in the irrigated area for wheat, resulting in a larger proportion of ${WF}_{blue}$ and a gradually increasing proportion of $C{F}_{irrigation electricity}$.

However, overall, the components of CF directly or indirectly caused by nitrogen ($C{F}_{direct,N_{2}O},C{F}_{indirect,N_{2}O},\mathrm{a}\mathrm{n}\mathrm{d}C{F}_{input,phosphate fertilizer}$) remain the main constituents of CF. The application of nitrogen fertilizer continues to be the largest contributing factor to greenhouse gas emissions from wheat production in this study.

4.2. Impact of Wheat Planting Area on Water and Carbon Footprints

Climate conditions and agricultural inputs are the primary intrinsic factors affecting the environmental footprint of crop production. From the calculation methods of crop WF and CF, the planting area (PA) is the direct factor determining their changes. Therefore, we used the rates of change in WF, CF, and PA for wheat production as variables for Spearman’s non–parametric correlation analysis. The purpose of this analysis was to avoid collinearity in the results and to understand the synchronous effects of changes in planting area and footprints. The results are shown in

Table 7. Correlation analysis between wheat planting area and the change rate of footprint indicators.

	$\mathit{\%}∆\mathit{W}\mathit{F}$	$\mathit{\%}∆\mathit{C}\mathit{F}$
$\%∆PA$	0.765	0.616

.

At p < 0.001, the rates of change in WF, CF, and PA were all significantly positively correlated. This indicates that the expansion of the wheat planting area in the Yarkant River plain oasis led to an increase in footprints, meaning that the growth of PA is a key driving factor for the increase in water resource usage and environmental impact. In the study area, agricultural planting relies almost entirely on irrigation. The expansion of planting areas inevitably leads to increased evapotranspiration, irrigation water usage, and agricultural input, thereby exacerbating the impact of agriculture on the ecological environment.

4.3. Policy Drivers and Green Agriculture Recommendations for Changes in Wheat Production Footprint Indicators

Grain demand, policy–driven factors, and regional economic development needs are the driving forces behind the expansion of wheat production in the study area. Due to special natural conditions and development needs, wheat production in the Yarkant River oasis requires more electricity for irrigation and more fertilizer input to maintain soil fertility, which will affect and change the ecological environment.

In the study, nitrogen fertilizer application is the largest contributing factor to water resource pollution and greenhouse gas emissions. Therefore, controlling nitrogen fertilizer application and optimizing fertilizer management are necessary measures to reduce regional environmental impacts. It is also recommended to reasonably plan cultivated areas, adjust planting structures, and promote crop production to concentrate in advantageous areas, which can effectively improve the utilization rate of agricultural resources and irrigation efficiency. Developing solar–powered irrigation systems and biomass power generation according to the local conditions can also help save water and reduce emissions, transitioning to high–quality mechanized agriculture.

Improving the accuracy and relevance of future research on the decoupling of agricultural production and environmental footprint is also a key focus of our next step. Currently, the Logarithmic Mean Divisia Index (LMDI) is widely used in exploring the factors influencing carbon footprint and carbon emissions. Combining the Tapio decoupling elasticity index with the LMDI model can obtain the driving factors affecting the decoupling state. We also consider using machine learning methods, such as support vector machines, BP neural network models, and random forests, to predict future agricultural production–related data to improve the accuracy of the results. Additionally, we will collect more detailed and higher–precision data to increase the accuracy and relevance of the models.

5. Conclusions

Based on the calculation of grain yield data and water footprint and carbon footprint in the Yarkant River oasis over the past 20 years, this study evaluates the water resource consumption and greenhouse gas emissions caused by wheat production in the Yarkant River oasis. On this basis, using decoupling methods, it analyzes the changes and dynamic relationships between wheat production development and environmental impact from multiple aspects. By analyzing the components of footprint indicators and the driving factors of changes, it aims to provide effective suggestions for the green development of agriculture in the basin. The results show the following:

(1)

From 2001 to 2020, the annual average growth of wheat production in the Yarkant River Plain oasis was 7.18 × 10³ tons, with annual average growths of 2.78 × 10⁶ m³, 4.17 × 10⁶ m³, and 7.72 × 10⁶ kg CO₂ eq for ${WF}_{green,blue},{WF}_{grey}, \mathrm{a}\mathrm{n}\mathrm{d} CF,$ respectively. During the study period, wheat planting was in a period of rapid development, causing significant consumption and impact on water resources, water environment, and greenhouse gas emissions.

(2)

From 2001 to 2020, the decoupling trend in wheat production ${WF}_{green,blue},{WF}_{grey}, \mathrm{a}\mathrm{n}\mathrm{d} CF$ in the Yarkant River Oasis was mainly weak decoupling and connection, with unstable decoupling states. This indicates that wheat production caused significant resource occupation and environmental impact. However, the strong decoupling state from 2018 to 2020 also shows that wheat production has the potential to reduce environmental consumption with the support of modern agricultural technology, presenting both risks and opportunities for achieving green sustainable development in regional wheat production.

(3)

Overall, there are differences in the development process of wheat yield and environmental changes among the three sub–irrigation districts (Shache, Maigaiti, and Bachu) in the Yarkant River. The decoupling trend between the production and footprint indicators differs among the sub–irrigation districts. The Shache sub–irrigation district, located near the outlet of the basin with more surface water resources, has the best environmental change situation among the three sub–irrigation districts, while the Bachu sub–irrigation district, farthest from the basin’s water source, has the greatest environmental impact caused by wheat production. This indicates that resource advantages are an important foundation for the green development of agricultural planting.

(4)

Wheat production in the Yarkant River plain oasis has issues of low efficiency and high consumption, especially the significant impact of nitrogen fertilizer and irrigation electricity on the water environment and carbon emissions. Therefore, it is recommended to improve technology and increase resource utilization efficiency.