Provincial-Level Carbon-Reduction Potential for Agricultural Irrigation in China

Abstract

Globally, agricultural irrigation accounts for the majority of freshwater use and 15% of annual agricultural greenhouse gas emissions, highlighting its critical mitigation potential amid climate change. While localized Chinese studies have analyzed the water–energy–carbon nexus, nationwide assessments of irrigation carbon-reduction potential, integrating crop water requirements, water use, and energy consumption, remain limited due to scarce longitudinal panel data. This study fills this gap by evaluating provincial-level potentials in China (2004–2020) using national/provincial statistical data on crop areas, irrigation water, energy use, and climate parameters. Findings reveal pronounced spatial–temporal variations: Henan, Heilongjiang, and Shandong exhibit the highest crop water demands (driven by rice/maize/wheat), while Heilongjiang, Jiangsu, and Guangdong show substantial water-saving opportunities. Xinjiang has the largest amount of irrigation-related carbon emissions, whereas the northeastern provinces offer the greatest reduction potential. A positive correlation between irrigation-carbon efficiency and groundwater utilization underscores the need for improved groundwater management. By linking crop water requirements to emission reductions through a nationally representative dataset, this study provides empirical evidence for region-specific strategies to enhance water-use efficiency and reduce irrigation’s environmental footprint. The findings inform policymakers on balancing agricultural productivity with sustainability goals, addressing both local water scarcity and global decarbonization imperatives.

1. Introduction

Globally, agricultural irrigation dominates the vast majority of the freshwater withdrawals and consumptions [1]. Meanwhile, irrigation contributes to 15% of the annual greenhouse gas emission in agricultural operations [2], while agriculture itself comprises 12% of the overall emissions [3]. In the context of global climate change and the urgent need to reduce greenhouse gas emissions, the carbon-reduction potential in agricultural irrigation cannot be overlooked.

As one of the nations with the largest agriculture scale and population globally, China comprises 22% of the world’s irrigation-equipped areas [4], and contributes 16% of the global CO₂ emissions from irrigation [2]. Over the past two decades, China’s central government has consistently placed a high priority on agricultural irrigation in its policy framework [5,6,7]. Consequently, from 2000 to 2015 [4], China witnessed the most significant expansion in irrigation area among all countries, increasing by 12.8 Mha, compared to the global total irrigated areas of 329 Mha in 2015. Moreover, China demonstrates pronounced spatial heterogeneity with respect to water resources, crop production systems, and climatic conditions, a pattern that can be effectively revealed through provincial-level analyses [5,8]. Therefore, it is extremely importance to assess the co-beneficial potential for water conservation and carbon emission reduction in China’s agricultural irrigation sector.

Existing analyses [9,10,11,12] of the water-energy-carbon nexus in China’s irrigated agriculture have documented critical insights into spatio-temporal dynamics, yet the majority of these investigations have been predominantly focused on localized regions with granular survey data. Nationwide, the discussion regarding the carbon emissions of the agricultural sector has gained increasing attention [5,13,14,15], yet research specifically addressing irrigation-related emissions remains comparatively limited, particularly given the paucity of comprehensive panel data at the national scale

Assessment of carbon emissions from agricultural irrigation generally encompasses direct energy consumption [16,17] associated with pumping, transporting, and spraying irrigated water, as well as the indirect energy use [18] embedded in the manufacturing of pumping and irrigation infrastructure. Additionally, groundwater degassing has been reported to contribute 6 Mt CO₂ annually, accounting for a portion of the global total of 222 Mt CO₂ [2], a component that warrants explicit consideration. While the potential for carbon mitigation is strongly contingent on irrigation methods [11], quantifying these effects at the national level presents significant methodological challenges. In this study, the measurement of carbon reduction potential—grounded in the definition of direct energy consumption—relies on irrigation water consumption reductions, estimated by calculating the discrepancy between current irrigation volumes and the deficit between crop water requirements and natural precipitation.

This study makes three principal contributions. First, it provides a meticulous calculation and analysis of annual crop water requirements for each Chinese province. Second, by integrating crop growth stage dynamics, it estimates the minimum annual irrigation volumes necessary for each province, thereby delineating water conservation potentials. Third, leveraging these water-saving potentials, it quantifies annual carbon emission reduction potentials for different provinces, accounting for heterogeneous irrigation technologies. The findings offer empirical foundations for formulating evidence-based recommendations to central and local policymakers, enabling the design of targeted strategies to alleviate water scarcity and mitigate the environmental footprint of agricultural irrigation.

2. Data and Methods

This section describes the methodological framework for quantifying crop water requirements, water-saving potentials, and carbon-reduction potentials. Annual assessments are conducted at the provincial level to characterize the spatial and temporal variability of these metrics across China.

2.1. Measuring Annual Crop Water Requirement

This section outlines the methodological approach for quantifying crop water requirements, water-saving potentials, and carbon-reduction potentials. Annual provincial-level assessments are performed to characterize the spatiotemporal dynamics of these indicators across China.

In agricultural systems, the crop water requirement per unit cultivated area is operationalized through crop evapotranspiration ($E{T}_c$), a metric that integrates both soil surface evaporation and plant transpiration processes throughout the crop growth cycle. Characterizing $E{T}_c$ is fundamental for agricultural water management, as it enables adaptive irrigation scheduling to meet crop water demands efficiently, thereby optimizing resource utilization and enhancing yield stability. The total annual water volume required for all crops ($V_{ET,total}$, km³ yr⁻¹) is calculated by aggregating the water requirements ($V_{ET,i}$, km³ a⁻¹) of individual crops as follows:

$$V_{ET,total}={\displaystyle \sum _{i=1}^N{V}_{ET,i}}={10}^{-8}\times {\displaystyle \sum _{i=1}^{N}P{A}_i\times E{T}_{c,i}}$$

(1)

where the subscript $i$ denotes the $i$th crop in Equation (1); the number ${10}^{-8}$ is used for unit transformation; $P{A}_i$ denotes the annual planting area for the $i$th crop (ha a⁻¹); $E{T}_{c,i}$ denotes the annual crop evapotranspiration for the $i$th crop (mm a⁻¹), which can be calculated by daily $E{T}_c$ from the planting day to the harvesting day as in the following:

$$E{T}_c={\displaystyle \sum _{J=1}^{365/366}\left(K_{c,J}\times E{T}_{o,J}\right)}$$

(2)

where $J$ denotes the number of days in a year between 1 (1 January) and 365 or 366 (31 December); $K_{c,J}$ denotes crop coefficient at day $J$; $E{T}_{o,J}$ denotes the daily crop evapotranspiration at day $J$ (mm d⁻¹), calculated using the FAO-recommended Penman–Monteith formula [19,20]. Figure 1 illustrates temporal variations in $K_{c,J}$ for wheat and maize as representative crops. Detailed computational procedures are provided in Appendix A.

2.2. Measuring Annual Water-Saving Potential

The water-saving potential ($WSP$) here refers to the maximum amount of irrigation water use ($IWU$) that can be reduced, which is determined by the difference between crop water requirements and effective precipitation. The total amount of crop water requirement has been quantified in the preceding subsection. The effective precipitation ($EP$) is normally defined as the aggregate precipitation that directly contributes to satisfying the evapotranspiration demands of a crop canopy, subsequently reducing the necessary amount of irrigation water [21,22]. In this context, a widely used method [9,23,24] in China was employed to determine the effective precipitation during one growth stage.

$$EP=\max \left(\min \left(AP,E{T}_c\right),0\right)$$

(3)

where $AP$ denotes the actual precipitation. It should be noted that the current total amount of irrigation water use is presumed to be entirely allocated for crop irrigation. Since the agricultural water data for individual crops are unavailable, it is necessary to calculate the annual cumulative effective precipitation for all crops during their respective growing seasons, namely the total annual effective precipitation ($V_{EP,total}$, km³ a⁻¹), which shall be consistent with the total annual crop water requirement ($V_{ET,total}$, km³ a⁻¹). Similar to Equation (1),the total annual effective precipitation can be calculated by

$$V_{EP,total}={\displaystyle \sum _{i=1}^N{V}_{EP,i}}={10}^{-8}\times {\displaystyle \sum _{i=1}^{N}P{A}_i\times E{P}_i}$$

(4)

where $E{P}_i$ denotes the annual effective precipitation for the $i$^th crop (mm a⁻¹). Therefore, the volume of potential irrigation water use ($PIWU$) considering $EP$ can be written as

$$V_{PIWU,total}=V_{ET,total}-V_{EP,total}$$

(5)

where $V_{PIWU,total}$ denotes the total volume of potential annual irrigation water use (km³ a⁻¹). Thus, the annual water-saving potential can be represented as

$$V_{WSP}=V_{IWU}-{\displaystyle \frac{1}{{\eta }_w}}{V}_{PIWU,total}$$

(6)

where $V_{WSP}$ denotes the total volume of annual water-saving potential (km³ a⁻¹);$V_{IWU}$ denotes the total volume of annual irrigation water use (km³ a⁻¹), which can be easily accessed via yearbooks. Considering the yearly variations in $EP$ and $ET$; ${\eta }_w$ denotes the efficiency of irrigation, which depends on the irrigation type. It should be noted that not all planting areas are irrigated. Therefore, Equation (6) can be modified as

$$V_{WSP}=V_{IWU}-{\displaystyle \frac{1}{{\eta }_w}}{\displaystyle \frac{I{A}_{total}}{P{A}_{total}}}{V}_{PIWU,total}$$

(7)

where $IA$ denotes annual irrigated area (km² a⁻¹).

Based on the data availability of China, the irrigation can be divided into non-water-saving ($NWS$) types and water-saving ($WS$) types, where the water-saving types can be further divided into sprinkler ($Spri$), micro-irrigation ($Micro$), low-pressure pipe-irrigation ($Pipe$), and other water-saving irrigation ($OWS$). Non-water-saving practices include flood irrigation, furrow irrigation, border strip irrigation, and so forth.

Table 1. Water efficiency (${\eta }_w$), working pressure ($H_{work}$), and head losses ($H_{loss}$) for each irrigation type.

Irrigation Type	Subscript	${\mathit{\eta }}_{\mathit{w}}$	${\mathit{H}}_{\mathit{w}\mathit{o}\mathit{r}\mathit{k}}$ (m)	${\mathit{H}}_{\mathit{l}\mathit{o}\mathit{s}\mathit{s}}$ (m)
Sprinkler	Spri	0.80	30.60	7.04
Micro-irrigation	Micro	0.85	10.20	7.04
Pipe irrigation (low-pressure)	Pipe	0.80	4.18	7.04
Other (water-saving)	OWS	0.50 (surface) 0.80 (groundwater)	0.00	7.04
Non-water-saving	NWS	Equation (8)	0.00	7.04

also shows the water efficiency of each irrigation type (${\eta }_w$), which measures how much water would be conveyed from the source to the field. In this study, ${\eta }_w$ of water-saving types are determined by the minimum requirement in China’s technical standards [25]. For the non-water-saving irrigation ($NWS$), the irrigation efficiency is assumed as

$${\eta }_{w,NWS}={\displaystyle \frac{I{A}_{NWS}}{I{A}_{total}/{\eta }_{w,total}-I{A}_{Spri}/{\eta }_{w,Spri}-I{A}_{Micro}/{\eta }_{w,Micro}-I{A}_{Pipe}/{\eta }_{w,Pipe}-I{A}_{OWS}/{\eta }_{w,OWS}}}$$

(8)

where ${\eta }_{w,total}$ denotes the overall irrigation efficiency. According to Table 1, ${\eta }_{w,OWS}$ is different for surface water ($SW$) and groundwater ($GW$). However, data on surface water and groundwater sources are only available for the total water use ($WU$). Thus, it is assumed that

$${\displaystyle \frac{V_{IWU,SW}}{V_{IWU,GW}}}={\displaystyle \frac{I{A}_{Spri,SW}}{I{A}_{Spri,GW}}}={\displaystyle \frac{I{A}_{Micro,SW}}{I{A}_{Micro,GW}}}={\displaystyle \frac{I{A}_{Pipe,SW}}{I{A}_{Pipe,GW}}}={\displaystyle \frac{I{A}_{OWS,SW}}{I{A}_{OWS,GW}}}={\displaystyle \frac{I{A}_{NWS,SW}}{I{A}_{NWS,GW}}}={\displaystyle \frac{V_{WU,SW}}{V_{WU,GW}}}$$

(9)

where $V_{WU}$ denotes the annual total water use, (km³ a⁻¹) $V_{WU}=V_{WU,SW}+V_{WU,GW}$. Therefore, $I{A}_{OWS}/{\eta }_{w,OWS}$ in Equation (8) can be replaced by the sum of $I{A}_{OWS,SW}/{\eta }_{w,OWS,SW}$ and $I{A}_{OWS,GW}/{\eta }_{w,OWS,GW}$.

Considering the irrigation efficiency of different irrigation types, Equation (7) shall be modified as

$$V_{WSP}=V_{IWU}-{\displaystyle \frac{V_{PIWU,total}}{P{A}_{total}}}{\displaystyle \sum _{irriType}\frac{I{A}_{irriType}}{{\eta }_{w,irriType}}}=V_{IWU,total}-{\displaystyle \frac{V_{PIWU,total}}{P{A}_{total}}}{\displaystyle \frac{I{A}_{total}}{{\eta }_{w,total}}}$$

(10)

where the subscript $irriType$ denotes the irrigation types, including all the five types in Table 1.

2.3. Measuring Annual Carbon-Reduction Potential

The carbon emission here refers to the combination of energy-related CO₂ emission and carbon emission from groundwater degassing [2], which can be written as

$$C_{irri}=C_e+C_{gd}$$

(11)

where $C_{irri}$ denotes the annual carbon emission due to irrigation (g CO₂ a⁻¹); $C_{gd}$ denotes the annual carbon emission due to groundwater degassing (g CO₂ a⁻¹), which can be written as

$$C_{gd}=V_{IWU,GW}{C}_{GW}$$

(12)

where $C_{GW}$ denotes the concentration of CO₂ in groundwater sources (g CO₂ km⁻³).

In Equation (11), $C_e$ denotes the annual carbon emission due to energy consumption during irrigation operations (g CO₂ a⁻¹), which can be written as

$$C_e=E_c{C}_{ef}$$

(13)

where $C_{ef}$ denotes the annual CO₂ emission factor of energy consumption (g CO₂/kWh a⁻¹), in this study set as 320.21 for energy use [2,26]; $E_c$ denotes the annual energy consumption during irrigation operations (kWh a⁻¹). The basic calculation formula of $E_c$ can be written as [2]

$$E_c={\displaystyle \frac{{10}^9{V}_{IWU}\rho g{H}_{total}}{{\eta }_e}}$$

(14)

where the unit of $V_{IWU}$ is km³ a⁻¹; $\rho$ denotes the water density, equal to 1000 kg m⁻³; $g$ denotes the gravity, equal to 9.81 m s⁻²; $H_{total}$ denotes the total pressure head (m) used in the irrigation systems, including the lift of groundwater ($H_{lift}$), working pressure ($H_{work}$), head losses ($H_{loss}$), $H_{total}=H_{lift}+H_{work}+H_{loss}$; ${\eta }_e$ denotes the energy efficiency of the irrigation system, set as 70% in this study [2], where it should be noted that energy efficiency ${\eta }_e$ is different from water efficiency ${\eta }_w$. $H_{total}$ would be affected by different irrigation types and different water sources. $H_{lift}$ refers to the water table depth, only for groundwater sources. Both $H_{work}$ and $H_{loss}$ would be affected by different irrigation types, as shown in Table 1 [16,27,28].

The corresponding water use of a certain irrigation type can be written as

$$V_{irriType}={\displaystyle \frac{I{A}_{irriType}/{\eta }_{w,irriType}}{I{A}_{total}/{\eta }_{w,total}}}{V}_{total}$$

(15)

In order to include the effects of irrigation types in the calculation of energy consumption, Equation (14) shall be modified as

$$E_c={\displaystyle \frac{{10}^9\rho g}{{\eta }_e}}{\displaystyle \sum _{irriType}{\left(V_{IWU}{H}_{total}\right)}_{irriType}}$$

(16)

Therefore, the carbon-reduction potential ($CRP$), which also represents the maximum possible reduction in energy use as well as groundwater during irrigation operations, can be calculated as

$$C_{CRP,total}={\displaystyle \frac{V_{WSP}}{V_{IWU}}}{C}_{irri,total}$$

(17)

2.4. Datasets

The data used in the above measurement, covering all the provinces in mainland China, can be mainly classified into the following types.

2.4.1. Meteorological Data

Meteorological data, obtained from the Global Hourly-Integrated Surface Database (ISD) at NCEI (https://www.ncei.noaa.gov/pub/data/noaa/isd-lite/, accessed on 14 May 2024), including weather station hourly observations such as $T$ (air temperature), $T_{dew}$ (dewpoint temperature), $u_2$ (wind speed rate), $S_{kc}$ (sky condition total coverage), and $AP$ (precipitation), as well as the station basic information such as $z$ (altitude) and $\phi$ (latitude). Due to the possibility that data from certain stations may not span the entire year, stations with at least 340 days of coverage annually were selected for each province, as illustrated in Figure 2.

Figure 3 further presents the count of weather observation stations in each province from 2004 to 2020, highlighting the consistent stability in the number of stations across the specific timeframe. The darker colors in Figure 3 indicate higher values, as shown by the corresponding numerical labels.

The hourly raw data may have gaps due to various issues, requiring specific adjustment to ensure its suitability for further calculation. Firstly, hourly data were converted into daily data by averaging the hourly data for each day, at each station within one province. Secondly, days with no data were estimated using inverse-distance-weighted (IDW) interpolation for each year, at each station within one province. In the IDW interpolation here, daytime served as the distance metric, utilizing data from the 10 nearest days as the interpolation stencil. Thirdly, daily data and location data from each station within one province were averaged into a single datum for the subsequent calculation of $E{T}_o$ and $EP$, as shown in Appendix A.

2.4.2. Statistics Data

Statistics data refer to the annual data, including $P{A}_i$ (planting areas for each crop), $V_{IWU,total}$ (volume of irrigation water use), $V_{WU}$ (volume of total water use), $V_{WU,SW}$ (volume of surface water use), $V_{WU,GW}$ (volume of groundwater use), $I{A}_{Total}$ (total irrigated area), $I{A}_{Spri}$ (irrigated area by sprinkler), $I{A}_{Micro}$ (irrigated area by micro-irrigation), $I{A}_{Pipe}$ (irrigated area by pipe irrigation), $I{A}_{OWS}$ (irrigated area by other water-saving irrigation), $I{A}_{NWS}$ (irrigated area by non-water-saving irrigation), and ${\eta }_{w,total}$ (overall irrigation efficiency) for each province from 2004 to 2020. These annual provincial data can be obtained from China’s National Bureau of Statistics (https://data.stats.gov.cn/, accessed on 11 June 2025), China Water Resources Bulletin, Statistic Bulletin on China Water Activities, and China Rural Statistical Yearbook.

2.4.3. Groundwater Table Data

Another important data used in the carbon emission calculation is $H_{lift}$, the lift of groundwater to surface for irrigation, also referring to groundwater table depth (

Table 2. Distance between groundwater table and the surface used for the lift of groundwater.

Province	${\mathit{H}}_{\mathit{l}\mathit{i}\mathit{f}\mathit{t}}$ (m)	Province	${\mathit{H}}_{\mathit{l}\mathit{i}\mathit{f}\mathit{t}}$ (m)	Province	${\mathit{H}}_{\mathit{l}\mathit{i}\mathit{f}\mathit{t}}$ (m)	Province	${\mathit{H}}_{\mathit{l}\mathit{i}\mathit{f}\mathit{t}}$ (m)
Anhui	18.09	Hainan	21.65	Jilin	12.43	Shanxi	41.22
Beijing	38.83	Hebei	43.78	Liaoning	20.65	Sichuan	84.67
Chongqing	69.52	Heilongjiang	8.28	Neimenggu	16.51	Tianjin	8.31
Fujian	52.50	Henan	22.90	Ningxia	41.99	Xinjiang	35.14
Gansu	46.01	Hubei	43.78	Qinghai	34.28	Xizang	51.95
Guangdong	29.63	Hunan	42.04	Shaanxi	63.27	Yunnan	80.79
Guanxi	45.34	Jiangsu	3.47	Shandong	13.72	Zhejiang	47.78
Guizhou	68.14	Jiangxi	33.37	Shanghai	0.66

Source from http://thredds-gfnl.usc.es/thredds/catalog/GLOBALWTDFTP/annualmeans/catalog.html, accessed on 14 May 2024.

). The groundwater table is assumed to be the same all the time, and for each province, $H_{lift}$ was calculated by averaging the global groundwater GIS data [29] (Figure 4).

3. Results

3.1. Crop Production in China

Based on the availability of data, 14 types of crops, including OtherCrops, were selected for this study to comprehensively cover the total irrigation water use in China. Figure 5 illustrates the total planting area and highlights the top five provinces with largest planting areas for each crop in 2020. It can be observed that maize, rice, and wheat, rank as the top three crops with largest planting areas, occupying 26.0%, 17.8%, and 14.2% of the total planting area across the country, respectively. The combined area of these top three crops accounts for more than half of the overall planting area nationwide.

According to Figure 5, the production of certain crops is primarily concentrated within one or several provinces. For example, 83.2% of cotton is planted in Xinjiang, while 53.5% of beans are planted in Heilongjiang and Neimenggu, and 53.3% of wheat is planted in Henan, Shandong, and Anhui. However, some crops, such as maize, rice, and vegetables, exhibit a more even distribution across the country and dominate in planting areas. Figure 6 shows the proportions of crop planting area in each province, where the size of the pie corresponds to the relative size of the planting area. It reveals that the top three provinces with largest planting are Henan, Heilongjiang, and Shandong. Furthermore, most of the maize (yellow parts in the pies) is planted in North China, while most of the wheat (orange parts in pies) is planted in the central part of China, and most of the rice (purple parts in pies) is planted in South China. In general, the majority of the crops are planted in the eastern region of China.

3.2. Evapotranspiration and Precipitation

Figure 7 shows the provincial-level distribution of mean annual reference evapotranspiration ($E{T}_o$) across the country during the period from 2004 to 2020. In comparison to Figure 6, it becomes evident that crops generally prefer lower $E{T}_o$ overall. It can be observed that $E{T}_o$ is higher in North China and the very south of China, where radiation levels are elevated or humidity levels are reduced. The smaller subfigures depict the temporal evolution of $E{T}_o$ from 2004 to 2020. Despite the presence of fluctuations in the changing trend of $E{T}_o$ over the years, the variance remains relatively small when compared to the overall value. Therefore, the mean annual $E{T}_o$ between 2004 and 2020, as colored in Figure 7, would be subsequently utilized for estimating the $E{T}_c$ of each crop.

Figure 8 illustrates the annual average precipitation of each province in China over the period from 2004 to 2020. The spatial distribution trend of precipitation is quite straightforward, with annual precipitation increasing from Northwest China to Southeast China. Specifically, Guangdong province has the highest annual average precipitation of 1843.4 (±279.7) mm/a, whereas Xinjiang province has the lowest annual average precipitation of 128.5 (±25.9) mm/a. It should be noted that both the evapotranspiration presented in Figure 7 and the precipitation presented in Figure 8 refer to the accumulated values over a whole year. It can also be observed that the precipitation has a higher year-to-year variation compared to the evapotranspiration. Based on the annual average data, most provinces in Northwest China exhibit lower precipitation levels than the reference evapotranspiration, which may, however, vary in specific years.

3.3. Annual Crop Water Requirement

As mentioned in Section 2.1, the estimation of the precise crop water requirement within a province should include as many crops as possible. Figure 9 shows the provincial-level distribution of annual crop water requirement ($V_{ET,total}$) and its proportions for different crops in 2020. Based on the year-to-year results, it can be observed that Henan province has the largest $V_{ET,total}$ of 48.7 (±2.8) km³/a, accounting for 9.1 (±0.3)% of the total $V_{ET,total}$ in China, followed by Heilongjiang with 39.5 (±5.4) km³/a, Shandong with 36.1 (±2.1) km³/a, Anhui with 31.8 (±2.4) km³/a, Hunan with 28.8 (±2.7) km³/a, Hebei with 28.6 (±1.4) km³/a, and so forth. However, as shown in Figure 9, Xinjiang province stands out as a special case, with its $V_{ET,total}$ increasing by 72.5% from 2004 to 2020, accounting for 5.9% of the total $V_{ET,total}$ in China in 2020, ranking 4th. The variance trend of $V_{ET,total}$ in Xinjiang is consistent with its total crop planting area, which increased by 80.0% from 2004 to 2020, accounting for 3.7% of that in China in 2020.

According to the results of 2020, rice, maize, and wheat are the top three crops in terms of $V_{ET,total}$, accounting for 24.0%, 20.3%, and 17.5% of all the crops, respectively. In fact, based on the year-to-year results, the $V_{ET,total}$ of rice, maize, and wheat occupies more than 2/3 of the total $V_{ET,total}$ in China during the period from 2004 to 2020.

$V_{ET,total}$ is calculated by the effective evapotranspiration, while $V_{PIWU}$ is derived from the difference between effective evapotranspiration and effective precipitation. The potential irrigation water use intensity ($V_{PIWU}/P{A}_{total}$), as depicted in Figure 10, is computed for comparison with the $E{T}_o$ in Figure 7 and the precipitation in Figure 8. $P{A}_{total}$ here represents the total planting area. It should be noted that $V_{PIWU}/P{A}_{total}$ is associated with the crop proportions but remains independent of agricultural irrigation practices. As shown in Figure 10, Xinjiang (421.7 mm/a), Tianjin (266.0 mm/a), and Ningxia (220.7 mm/a) are the top three provinces with the highest $V_{PIWU}/P{A}_{total}$ in 2020. It can also be observed that most provinces in South China exhibit a $V_{PIWU}/P{A}_{total}$ lower than 50 mm/a.

3.4. Water-Saving Potential

Irrigation water use intensity ($V_{IWU}/I{A}_{total}$) represents the annual volume amount of water used for each irrigation area. Due to the efforts of China’s government in the past two decades, $V_{IWU}/I{A}_{total}$ for the whole of China has consistently decreased by 20.6% to 521.6 mm/a from 2004 to 2020. Figure 11 shows the provincial-level distribution of $V_{IWU}/I{A}_{total}$ in 2020. It can be observed that Guangdong has the highest $V_{IWU}/I{A}_{total}$ of 1187.2 mm/a in 2020, which decreased by 35.2% from 2014 to 2020. This is followed by Hainan (1143.0 mm/a), Guangxi (1079.7 mm/a), Ningxia (1060.6 mm/a), Xinjiang (1014.0 mm/a), Xizang (968.8 mm/a), and so forth. Conversely, Beijing has the lowest $V_{IWU}/I{A}_{total}$ of 150.0 mm/a, which decreased by 65.2% from 2014 to 2020. This is followed by Henan (226.1 mm/a), Hebei (240.9 mm/a), Shandong (253.1 mm/a), and so forth. Overall, $V_{IWU}/I{A}_{total}$ is much lower in North China Plain than other regions in China.

Based on Equation (7), the water-saving potential $V_{WSP}$ is derived by the difference between $V_{IWU}$ and $V_{PIWU}$. According to the year-to-year results, the annual average $V_{WSP}$ for the whole of China is 564.7 (±16.0) km³/a during the period from 2004 to 2020. As shown in Figure 12, in the year of 2020, Heilongjiang has the highest $V_{WSP}$ of 27.4 km³/a, which is followed by Jiangsu (24.7 km³/a), Guangdong (20.7 km³/a), Hunan (19.6 km³/a), Guangxi (18.3 km³/a), Jiangxi (16.2 km³/a), and so forth.

3.5. Carbon-Reduction Potential in Irrigation

$C_{irri}$ represents the annual carbon emission due to irrigation. The $C_{irri}$ for the whole of China decreased by 3.2% to 10.4 Mt CO₂/a during the period from 2004 to 2020. Figure 13 shows the provincial-level distribution of $C_{irri}$ in 2020. It can be observed that Xinjiang province has the highest $C_{irri}$ of 1.99 Mt CO₂/a in 2020. This is followed by Heilongjiang (1.38 Mt CO₂/a), Neimenggu (0.77 Mt CO₂/a), Hebei (0.69 Mt CO₂/a), Henan (0.67 Mt CO₂/a), Shandong (0.55 Mt CO₂/a), and so forth. Overall, $C_{irri}$ turns out to be much higher in North China. According to the year-to-year results, the top five provinces experiencing an increase in $C_{irri}$ from 2004 to 2020 are Xinjiang (by 90.2%), Heilongjiang (by 63.3%), Xizang (by 61.3%), Chongqing (by 32.1%), and Anhui (by 30.0%). These are the only provinces that exhibited an increase in $C_{irri}$ over this period. Conversely, the top three provinces with a decrease in $C_{irri}$ are Beijing (by 86.2%), Hebei (by 51.3%), and Tianjin (by 50.0%).

According to Equation (17), $C_{CRP}$ is calculated by multiplying $C_{irri}$ by the ratio of water-saving potential to irrigation water use. Over the period from 2004 to 2020, the $C_{CRP}$ for the whole of China exhibits a variation, with a mean value of 7.28 Mt CO₂/a and a standard deviation of 0.72 Mt CO₂/a. No clear trend of increase or decrease is observed. Figure 14 shows the provincial-level distribution of $C_{irri}$ in 2020. It can be seen that Heilongjiang province has the highest $C_{CRP}$ of 1.37 Mt CO₂/a, accounting for 28.2% of the total $C_{CRP}$ for China. This is followed by Xinjiang (0.54 Mt CO₂/a), Neimenggu (0.45 Mt CO₂/a), Liaoning (0.37 Mt CO₂/a), Shandong (0.37 Mt CO₂/a), Jilin (0.33 Mt CO₂/a), and so forth. The carbon-reduction potential is primarily concentrated in the provinces of Northeast China, namely Heilongjiang, Jilin, and Liaoning, as well as Neimenggu, which account for nearly half of the total value in China.

4. Discussion

Based on the economic and climatic factors, provinces in mainland China are typically categorized into four regions [30]: eastern region, central region, western region, and northeastern region, as shown in

Table 3. Regions of provinces in mainland China.

Regions	Provinces	GDP Per Capita		$\mathit{I}{\mathit{A}}_{\mathit{W}\mathit{S}}/\mathit{I}{\mathit{A}}_{\mathit{t}\mathit{o}\mathit{t}\mathit{a}\mathit{l}}$
		2004	2020	2004	2020
Eastern	Beijing, Tianjin, Hebei, Jiangsu, Zhejiang, Shanghai, Fujian, Shandong, Guangdong, Hainan	CNY 19.3k	CNY 93.1k	42.9%	67.7%
Central	Shanxi, Anhui, Jiangxi, Henan, Hubei, Hunan	CNY 8.8k	CNY 60.5k	21.2%	30.9%
Western	Neimenggu, Guangxi, Sichuan, Guizhou, Yunnan, Xizang, Shaanxi, Gansu, Ningxia, Xinjiang, Chongqing, Qinghai	CNY 7.9k	CNY 55.6k	47.5%	71.6%
Northeastern	Heilongjiang, Jilin, Liaoning	CNY 12.2k	CNY 51.8k	36.1%	41.1%

. Among all these regions, the eastern region stands out as the most developed, accounting for 52.0% of China’s GDP and 40.0% of its population in 2020, yet it only comprises 21.6% of China’s planting area. The western region and central region occupy the majority of China’s planting area, with respective percentages of 33.6% and 29.3%. Over the past two decades, the Chinese government has consistently promoted water-saving irrigation [8,30], particularly in the western region, owing to its low economic level, arid climate, and extensive land available for planting. Through these continuous efforts, the percentage of water-saving irrigation area $I{A}_{WS}/I{A}_{total}$ in the western region has increased from 47.6% to 71.6% between 2004 and 2020.

4.1. Analysis of Irrigation-Carbon-Emission Intensity

Figure 15 shows the changes in irrigation-carbon-emission intensity ($C_{irri}/I{A}_{total}$) and irrigation-water-use intensity ($V_{IWU}/I{A}_{total}$) for the four regions between 2004 and 2020. It can be observed that the increase in $I{A}_{WS}/I{A}_{total}$ over the past two decades has directly resulted in the decline of the water used for irrigation, as well as a reduction in the carbon-emission associated with irrigation. However, Xinjiang province, which has the highest $C_{irri}$ in China, experienced a notable increase of 20.8% in $C_{irri}/I{A}_{total}$ from 2004 to 2020. The increase can be primarily attributed to the rapid adoption of water-saving irrigation technologies in Xinjiang. From 2004 to 2020, $I{A}_{WS}/I{A}_{total}$ in Xinjiang rose from 60.83% to 88.56%, especially with the micro-irrigation area undergoing a 12.66-fold expansion over the period. On the other hand, Heilongjiang province, which has the second highest $C_{irri}$ in China, experienced a remarkable decrease of 39.6% in $C_{irri}/I{A}_{total}$ from 2004 to 2020. Although the absolute decline of $V_{IWU}/I{A}_{total}$ in Xinjiang (457.0 mm/a) is slightly larger than that in Heilongjiang (365.3 mm/a), the relative change in Xinjiang (−31.1%) is smaller than that in Heilongjiang (−44.7%).

4.2. Analysis of Irrigation-Carbon-Emission Efficiency

In order to further analyze the relationship between carbon-emission and water-use, the concept of irrigation-carbon-emission efficiency is defined as $C_{irri}/V_{IWU}$ (g CO₂/m³), as shown in Figure 16, which illustrates the year-to-year trend of $C_{irri}/V_{IWU}$ from 2004 to 2020. It can be observed that $C_{irri}/V_{IWU}$ has increased in 9 out of the 31 provinces during this period. Specifically, Xinjiang province has the largest percentage increase in $C_{irri}/V_{IWU}$ with 75.2%, while Heilongjiang shows an increase percentage of 9.2%. Further investigation into the provincial year-to-year data revealed that $C_{irri}/V_{IWU}$ has a strong positive linear relationship with $V_{IWU,GW}/V_{IWU}$ (Pearson correlation coefficient = 0.981), which represents the ratio of groundwater utilized for irrigation purposes.

This finding underscores the critical role of groundwater management in mitigating irrigation-related emissions. For instance, Xinjiang, with a groundwater table depth of 35.14 m (Table 2), relies extensively on energy-intensive pumping for its micro-irrigation systems, contributing to its status as China’s top irrigation carbon emitter (1.99 Mt CO₂/a in 2020, Figure 13). Targeted policy interventions are warranted to address this nexus:

(1) Renewable energy integration: Provinces with deep groundwater tables (e.g., Xinjiang, Shanxi) should prioritize deploying solar- or wind-powered pumps to replace fossil fuel-driven irrigation systems. This shift could reduce energy-related emissions while addressing the high lift requirements of deep groundwater.

(2) Crop structure optimization: High-emission regions like Xinjiang could transition to drought-resistant crops (e.g., optimizing cotton cultivation) to reduce water and energy demand, while northeastern provinces (e.g., Heilongjiang) could expand low-carbon crop portfolios.

4.3. Water-Saving and Carbon-Reduction in the Regions of China in 2020

Table 4. Potentials of water-saving and carbon-reduction in the regions of China in 2020.

Regions	${\mathit{V}}_{\mathit{W}\mathit{S}\mathit{P}}/\mathit{I}\mathit{A}$ (103 m3/ha)	${\mathit{V}}_{\mathit{W}\mathit{S}\mathit{P}}$ (km3)	${\mathit{C}}_{\mathit{C}\mathit{R}\mathit{P}}/\mathit{I}\mathit{A}$ (kg CO2/ha)	${\mathit{C}}_{\mathit{C}\mathit{R}\mathit{P}}$ (Mt CO2)	${\mathit{C}}_{\mathit{i}\mathit{r}\mathit{r}\mathit{i}}/{\mathit{V}}_{\mathit{I}\mathit{W}\mathit{U}}$ (g CO2/m3)	${\mathit{V}}_{\mathit{I}\mathit{W}\mathit{U},\mathit{G}\mathit{W}}/{\mathit{V}}_{\mathit{I}\mathit{W}\mathit{U}}$ (%)
China	4.12	288.1	104.1	7.21	28.9	15.4
Eastern	4.19	80.7	75.0	1.44	21.9	10.3
Central	3.51	69.8	63.3	1.26	22.0	12.6
Western	4.51	92.0	119.6	2.44	31.4	15.7
Northeastern	4.39	42.6	213.0	2.07	48.5	38.5

shows the analysis of water-saving and carbon-reduction potentials across various regions of China in 2020, which identified distinct variations in key metrics related to agricultural irrigation practices. The eastern region, despite having the highest $V_{WSP}/IA$ at 4.19 m³/ha, exhibited relatively low carbon emission intensity due to $C_{CRP}/IA$ at 75.0 kg CO₂/ha, suggesting room for water efficiency improvements without significantly impacting carbon emissions. In contrast, the northeastern region stood out with the highest $C_{CRP}/IA$ at 213.0 kg CO₂/ha, as well as the highest $V_{IWU,GW}/V_{IWU}$ at 38.5%, reflecting intensive agricultural activities and high energy consumption for irrigation. Despite moderate $V_{WSP}/IA$ at 4.39 m³/ha, this region presented the greatest potential for carbon reductions, contributing 28.7% of China’s total $C_{CRP}$. The western region, characterized by arid conditions and heavy reliance on irrigation, demonstrated high $V_{WSP}/IA$ (4.51 m³/ha), $C_{CRP}/IA$ (119.6 kg CO₂/ha), and $V_{WSP}$ of 92.0 km³, emphasizing the need for targeted interventions to mitigate its substantial water and carbon footprints. The central region, with moderate $V_{WSP}/IA$ and $C_{CRP}/IA$ values, also presented opportunities for water savings and carbon reductions, albeit to a lesser extent compared to the northeastern and western regions. Our findings underscore the importance of region-specific strategies to address the unique challenges and opportunities for sustainable agricultural development in China.

4.4. Study Limitations and Future Research Directions

While this study provides a comprehensive national-scale analysis of irrigation carbon emissions and water-saving potentials, several limitations should be acknowledged. First, the provincial-level data aggregation may obscure sub-provincial heterogeneities, such as county-level variations in irrigation practices or groundwater dynamics, which is a common challenge in large-scale studies. For instance, in India [31], similar limitations were noted where district-level variations in pump efficiency and groundwater depth affected carbon emission estimates. Second, the calculation of groundwater degassing emissions relies on a static CO₂ concentration parameter, which may vary across aquifers due to geological processes or human activities, as highlighted in Iran’s groundwater management studies [32,33]. Additionally, the energy efficiency of irrigation systems is a generalized value, potentially underestimating technical variations among provinces. For example, Pakistan’s diesel pump efficiency (20~30%) [34] differs significantly from China’s electric pump efficiency assumptions, leading to cross-regional incomparability.

At the global level, irrigation carbon mitigation strategies exhibit significant diversity. In India [31], groundwater irrigation accounts for 8~11% of total national carbon emissions, driven by subsidized electricity and deep tubewells, mirroring China’s reliance on groundwater in arid regions like Xinjiang. However, India’s higher proportion of diesel-powered pumps (5% of total emissions) [31] contrasts with China’s electrification-dominated systems, leading to distinct emission profiles. Iran [32,33] and Pakistan [34,35] have demonstrated that improved irrigation scheduling and water productivity can reduce groundwater extraction by 20~40%, strategies that could complement China’s water-saving irrigation projects. Spain’s case [36] further highlights the trade-off between desalinated seawater and carbon emissions, showing that energy-intensive water sources may increase GHG footprints by 50%, a challenge relevant to China’s coastal agricultural zones exploring non-conventional water sources.

Future research should address these gaps through multi-scale, cross-country comparisons to identify transferable solutions. For example, integrating high-resolution satellite data (as used in Spain’s GIS analysis [36]) with machine learning models could refine sub-provincial/regional estimates of irrigation carbon footprints, enabling comparisons between China’s North China Plain and India’s Punjab region. Additionally, scenario-based analyses simulating climate change impacts (e.g., altered precipitation in Iran’s Gamasiab Basin [32]) could inform adaptive policies in China’s semi-arid zones. Cross-disciplinary studies exploring the water–energy–food nexus, such as evaluating the carbon implications of China’s crop structure versus India’s cotton–wheat systems, would provide actionable insights for global decarbonization.

Furthermore, international collaborations are needed to harmonize methodologies for quantifying irrigation-related emissions, particularly in transboundary. Lessons from Pakistan’s participatory groundwater management and Iran’s SWAP model applications could enhance China’s groundwater governance in shared aquifer regions. By bridging these gaps, future studies can deepen our understanding of sustainable irrigation practices and support global climate mitigation targets.

5. Conclusions

This study quantifies provincial-level crop water requirements, water-saving potentials, and carbon-reduction capacities in China’s agricultural irrigation (2004–2020), revealing pronounced spatial–temporal disparities. Henan, Heilongjiang, and Shandong exhibit the highest crop water demands, driven by rice, maize, and wheat cultivation, while Xinjiang faces the largest irrigation-related carbon emissions (1.99 Mt CO₂/a in 2020) due to energy-intensive micro-irrigation expansion. Northeastern provinces (e.g., Heilongjiang) offer the greatest carbon-reduction potential, contributing 28.7% of the national total, with a strong correlation (Pearson = 0.981) observed between irrigation-carbon efficiency and groundwater utilization.

Nationwide, water-saving potential reached 288.1 km³ in 2020, underscoring the effectiveness of water-saving technologies, particularly in the western region (water-saving irrigation area increased from 47.6% to 71.6%). However, the study’s provincial-level aggregation may obscure sub-regional heterogeneities in groundwater dynamics and irrigation efficiency. Future research could leverage high-resolution remote sensing and machine learning to refine sub-provincial estimates, validate groundwater degassing parameters, and explore climate change impacts on irrigation-carbon dynamics.

This work advances our understanding of China’s agricultural water–energy–carbon nexus, demonstrating that significant opportunities exist to enhance irrigation efficiency and reduce carbon footprints. The findings emphasize the importance of integrating crop-specific water requirements, irrigation technologies, and regional hydrology to achieve sustainable agricultural development. Future research should further explore the long-term impacts of climate change on these dynamics and refine localized policy tools.