Water footprint of food production and consumption in China Open Access

Water and food are at the core of the United Nations 2030 Agenda for Sustainable Development (United Nations 2015). The lifeblood of humans is in the field; the lifeblood of the field is in the water. Agriculture carries a huge environmental cost because it is the largest water user. Food systems include various actors and their interrelated value-added activities related to production and consumption. Therefore, the comprehensive measurement and comparison of water resources consumption of food production and consumption is a hot issue to achieve sustainable food and water resources systems. Food production needs the supply of water resources, and food consumption also implies the water consumption. From local and global levels, water shortage is a key problem that may aggravate food security (Skaf et al. 2020). Food utilization is for the well-being of human nutrition, depends on water and sanitation conditions, and may also be damaged by the loss of drinking water (Wheeler & Braun 2013). What is more, irrigation water accounts for more than two-thirds of the world's total. In some countries (e.g., Bangladesh, Afghanistan, India, Pakistan) and regions (e.g., Xinjiang, China), about more than 90% of total water consumption (physical water) was irrigation water, most of which is used to produce food (Chen 2020).

Moreover, water is fundamental to maintain human life and the balance of ecosystems. However, due to global economic development, climate change, changing consumption patterns and population growth, the world's water scarcity is particularly serious. The World Water Development Report 2020 identifies water security as one of the lasting and profound crises facing the world in the coming decades. Global water withdrawal has increased six-fold in the past 100 years and continues to grow at a steady rate of about 1% a year (United Nations 2020). The effective solution to the problem of water shortage is related to human beings and the sustainable balance of the ecological environment. However, so far there has been no critical breakthrough in the way of solving the problem, which is a cause for concern. At present, people are still trying to find an appropriate mechanism that can provide effective help to solve the crisis. The issue of water resources is a complex social task that needs comprehensive and flexible application of various means which include economic, political, cultural and technological.

However, water resources are generally recognized as physical, which has been limited to some extent. Scholars proposed the new concepts ‘virtual water’ and ‘water footprint’ to understand water resources. In addition to visible physical water consumption, there is also invisible water consumption. ‘Virtual water’ was originally proposed by Tony Allan as a solution to water problems in the Middle East (Allan 1993) and refers to the amount of water needed to produce goods and services. In 2002, the ‘water footprint’ concept was put forward by Hoekstra & Hung (2002), which refers to the water resources needed for all the goods and services consumed by a country, a region or an individual under a certain living standard. They are also considered as effective tools to alleviate regional and global water crises. The main problem to be solved in this paper is to analyze the utilization of water resources caused by food production and consumption by a water footprint method and put forward sustainable countermeasures and suggestions.

Water footprint is a multi-faceted indicator that can reflect the quantity, quality and type of water resource. The water footprint measurement and evaluation are mainly carried out from different spatial scales and industrial sectors, such as global scale (Hoekstra & Chapagain 2007; Vanham & Bidoglio 2013), national scale (Dong et al. 2014), cities (Gao et al. 2014) and river basins (Vanham 2013; Pellicer-Martínez & Martínez-Paz 2018). As for different sectors, the research is mainly concentrated on the agricultural and industrial sectors, while the service sector is slightly weak (Zhang et al. 2021; Zhang & Tian 2022). The measurement in the agricultural sector includes two categories of crop and animal products, from production and consumption perspectives. Referring to the production point, there are both the study of multi-crop assemblage (Mekonnen & Hoekstra 2010a, 2010b; Hui et al. 2021) and the study of single crop, for example, rice (Chapagain & Hoekstra 2011; Marano & Filippi 2015) and corn (Bocchiola et al. 2013). Food consumption is also an important research perspective of water footprint, such as residents' food consumption and tourists' food consumption. In China, due to differences in dietary structure and family income, rural households have a higher water footprint of food consumption than urban households (Zhang et al. 2019). However, according to the trend, the food water footprint of urban residents' consumption shows a rising trend, while that of rural residents is on the decline (Zhang et al. 2019). The per capita food virtual water consumption in Gansu Province in China showed an increasing trend from 2000 to 2014, largely because the consumption of animal products increased in the total food consumption structure (Ju et al. 2018). However, comparative studies on the water footprint of food production and consumption need to be enriched and deepened, especially the diachronic study of multi-region and multi-food species; whether there is a balance between supply and demand is also worth studying.

Food production and consumption is a kind of human activity with certain environmental externality. In recent years, high grain yield in China has been essential for ensuring the national economy. However, the water resources consumption and the environmental load of cultivated land are also worrying. Agriculture has become the country's largest water consumption sector (Dong et al. 2014). Moreover, China's large population results in the increasing demand for food and has also stimulated the continuous growth of grain production, which further aggravates the shortage of water resources. Therefore, the water footprint of Chinese agricultural products has become a hot point for many researches. In terms of grain cultivation, the country's water consumption was about 689.04 billion m³ (Cao et al. 2015). In Beijing, the water footprint of crop production has decreased by 35.1% from 1978 to 2012 (Xu et al. 2015); while the virtual water of residents' food consumption in daily life was nearly five times that of the physical water consumption, reaching 1023.05 m³/cap (Wu et al. 2011). Unfortunately, the water use of China's agriculture is inefficient and the threat to water resources is growing. In addition, the input of agricultural resources is relatively high, but the efficiency is low, agricultural non-point source pollution is intensifying.

Water is essential for sustainable food production and consumption. Water footprint is an effective method to comprehensively measure water resource consumption of products, which is helpful for in-depth understanding of water resource problems. However, related research on Chinese agricultural products mainly focused on a few major crops or specific provinces or cities, and there is a lack of comparison and evolution research on food production and consumption. By comparing the statistical yearbook data, the study took 12 types of food into consideration which accounted for more than 70% of residents' food consumption, including seven crop and crop products (rice, wheat, corn, soybean, potato, vegetable, edible vegetable oil) and five animal products (pork, beef, mutton, poultry and eggs). The study estimated the water consumption of food production and consumption from 2001 to 2019, and put forward suggestions for sustainable production and consumption.

denotes the water footprint of food production; denotes the production of food j in province i (ton); denotes the unit water footprint of food j in province i (m³/ton) which is classified into green, blue and gray water (Mekonnen & Hoekstra 2010a).

denotes the water footprint of food consumption (urban or rural); denotes the per capita annual consumption of food j in province i (ton) (urban or rural); denotes the urban or rural population in province i.

A water stress indicator can be used to reflect the degree of water scarcity which is the ratio of annual availability to renewable water resources. When the water use intensity value is lower than 0.1, it indicates low water stress; higher than 0.1 and lower than 0.2 indicates medium water stress; higher than 0.2 and lower than 0.4 indicates medium-high water stress, higher than 0.4 indicates high water stress (Zhang & Tian 2022). In this study, the regional water stress indicator was evaluated from 2001 to 2019.

Spatial analysis can intuitively display the spatial distribution characteristics of each element. In this study, the spatial overlay analysis was used to overlay the water stress indicator and production and consumption water footprint of each province. By revealing the spatial pattern characteristics of the water footprint of production and consumption, as well as the spatial matching differences with provincial water resources pressure, it is helpful to analyze the impact on local water resources.

According to the formula, the data mainly consisted of three parts: the quantity of food production and consumption, unit water footprint of food, and the data of population. The study covers the period from the year 2001 to 2019 and covers 31 provincial administrative regions in China, excluding Hong Kong, Macao and Taiwan, China.

The quantity data of food production and consumption were from the China Statistical Yearbook, China Rural Statistical Yearbook and the statistical yearbooks of the provinces. The study used the data from 2001 to 2019. In the yearbooks, there are detailed records of production and household consumption. Taking consumption for example, there are records of consumption of different foods per person per year in towns and villages. However, there are also a small amount of data missing which can be calculated or estimated by using existing data. For example, Heilongjiang province lacks food consumption data of urban and rural residents from 2013 to 2014, which can be obtained by using the prediction worksheet function of EXCEL. Jiangxi, Hunan and Beijing lack the food consumption data of urban residents, so this study uses the average food consumption data of national urban residents instead.

The data of the unit water footprint of each food were mainly from Value of Water Research Report Series No. 47 and No. 48 (Mekonnen & Hoekstra 2010a, 2010b). Due to the unavailability of data, we cannot calculate the water footprint of 12 categories of agricultural products in all provincial administrative regions of China. The two reports are the most comprehensive calculations to date of the water footprint of various products in various countries around the world. The results of the report are based on data from 1996 to 2005. Since 2005, there have been many quantitative studies on the water footprint of agricultural products, but due to different parameter settings or methods, it is impossible to do a simple comparative study. Although the report has some limitations, it also provides some reference for policy makers, managers or researchers because it is an equal measurement. For China, Report No. 47 offers the green, blue and gray water footprints of various agricultural products in China's provinces. Besides, because the Chinese statistics do not involve a more detailed classification of vegetable and edible vegetable oils, the unit water footprint is based on the recorded average of 20 vegetable and 11 vegetable oils respectively. Report No. 48 provides the average value of animal products in China, and the water footprint of provincial animal products should be back-estimated based on its ratio to the national average. Assuming that drinking water and cleaning service water for animals in each province are the same, and the difference lies in the feed water footprint, the ratio of the feed water footprint (mainly including corn flour, bran and soybean cake) to the national average water footprint is taken as the ratio of the provincial animal product water footprint to the national average water footprint (Zhang et al. 2019). For example, the national average water footprint of beef is 13,688 m³/ton. After calculation, the ratio of the Anhui beef water footprint to the national average is 1.03, so the Anhui beef water footprint is 14,098 m³/ton.

The provincial urban and rural population data are derived from the online database National Data. This is the official online database owned by the National Bureau of Statistics of China.

This section analyzes and discusses the and from the points of spatial-temporal, water structure and food structure and puts forward suggestions for sustainable food production and consumption.

According to the formula of the water footprint of production and consumption, it can be found that reducing the water footprint of unit product is an important way to reduce the total water resources utilization. In particular, the production water footprint has been on the rise, and its impact on water resources is also increasing, so it is crucial to reduce the water footprint of unit product. It is necessary to improve irrigation efficiency and control the application of chemical fertilizer and pesticide to reduce the grey water footprint. In addition, provinces should constantly improve their capacity to supply water resources and ensure water security. For example, optimize and adjust the crop planting structure, reduce the planting area of high-water consuming crops, and strictly control the irrigation area of paddy fields downstream of urban and rural water supply reservoirs. We will deepen domestic water conservation and strictly control water use in high water consumption service industries.

In addition, the spatial distribution of China's water resources is extremely uneven, and soil and water resources are not completely matched with the development of national economy. Therefore, different regions are faced with different pressures of water resources, and the spatial distribution difference is significant, which is not optimistic on the whole. The water footprint of food production brings different water resources pressure to different regions, which must be analyzed in combination with the local water resources. A water stress indicator can reflect the degree of water resource scarcity in a region. In general, the water footprint of food production roughly matched the water stress level (Figure 2). The regions with high water stress also had high ⁠, such as Hebei, Henan, Heilongjiang, Jiangsu, and Shandong, mainly in the north. Similarly, Qinghai, Tibet, Guizhou were low water stress regions, while with low due to relatively poor agricultural production conditions. For ⁠, high and low value provinces have different water resources stress, and there seems to be no significant correlation between the two. For example, Guangdong, Shandong and Sichuan had higher ⁠, but medium-high, high and low water stress.

What is more, national water resources show a reverse flow configuration of real and virtual forms between the northern and southern regions in China, that is, on the one hand, water flows from water abundant regions to water deficient regions in the form of physical water (south-to-north water diversion); on the other hand, water flows from water deficient regions to water abundant regions in the form of virtual water (north grain transportation to south). The precious water resources, soil nutrients, light and temperature resources that are not rich in the north are transferred to the south where there is abundant water, abundant light and temperature resources and great high yield potential through the virtual water trade. The consequence of this abnormal subsidy and resource plunder and transfer not only severely hit the production enthusiasm of farmers in the north, but also caused serious agricultural water consumption in the north, aggravating the water resources crisis. This requires the establishment of a sound ecological compensation mechanism for water resources.

There is little difference in water resource structure between and ⁠. The green water footprint of 12 food productions was 70.07% of the total, blue water 9.28% and gray water 20.65% while the ratio of food consumed was 69.68, 11.76 and 18.56% respectively. Table 1 depicts the water structure of and in each province. There is great variation between provinces.

Green water is known as the ‘invisible water’ that crops take up from the soil at the root of plants. For rain-fed agriculture, it is extremely important. The low opportunity cost (Hess 2010) and negative environmental externalities (Aldaya et al. 2010) often result in people ignoring green water and considering it unimportant. However, as a part of crop water consumption, it can be used to evaluate the impact of agricultural production on water environment systems (Hess 2010). In China, the green water footprint occupies the highest proportion in both food production and food consumption, as depicted in Table 1. Green water consumption is far more than blue and gray water. It affects agricultural production and food security in China and the world. Therefore, to reduce the total ⁠, agricultural production in China should strengthen water management and protection, and minimize the green water evaporation loss.

By comparing the structure of water used for food production in China with that in the world, we can also see that China's gray water footprint was relatively high (Zhang et al. 2021). Gray water can reflect the degree of water pollution because it indicates the amount of fresh water required to absorb and assimilate a given pollutant load under the natural background concentrations and existing environmental water quality standards (Mekonnen & Hoekstra 2010a). In the past 20 years, with the increase of food production, the amount of chemical fertilizer applied in China has also been increasing rapidly. Excessive use of chemical fertilizers and pesticides in food production has made China's agriculture become the largest non-point source pollution industry, surpassing industry. Although the water footprint per food produced only considered nitrogen, phosphorus emissions were also not negligible (Mekonnen & Hoekstra 2010a). According to the Bulletin of the Second National Census of Pollution Sources in China, nitrogen and phosphorus pollutants have great influence on water environment quality, and agricultural pollution sources are important sources of nitrogen and phosphorus pollutants. Total nitrogen and phosphorus emissions from agricultural sources accounted for 46.5 and 67.2% of China's total emissions (Ministry of Ecology and Environment et al. 2020). In addition, the amount of chemical fertilizer applied per unit of cultivated land exceeded the world average by four times in China. However, the effective utilization rate of chemical fertilizer is only 35% (Li et al. 2018). How to decrease agricultural pollution to reduce the gray water footprint? For government regulation, agricultural subsidies can be used to encourage good behavior, while pollution taxes can be used to discourage bad behavior (Abadie et al. 2016). Production subsidies increase the amount of food, but they also increase agricultural pollution. In contrast, green subsidies help to reduce agricultural pollution. Besides, reducing the chemical fertilizers use and replacing them with organic fertilizers are also essential to reduce pollution. Enough attention should be paid to nitrogen and phosphorus discharge in the livestock and poultry industry, except agricultural planting. Finally, it is necessary to correct the cognitive bias of farmers in fertilization and continuously improve the level of scientific and efficient fertilization. At present, farmers overestimate the effect of fertilizer input on grain yield increase in China.

Blue water is the consumption of surface water and groundwater (Mekonnen & Hoekstra 2010a). It is mainly used for irrigation in food production. It is the lowest in most provinces. However, several provinces need to pay attention, such as Ningxia, Xinjiang, and Gansu, especially Xinjiang, where the blue water proportion in the is 34.88% while 37.88% in (Table 1). This is mainly due to climate, rainfall is scarce, and green water resources are relatively short. Therefore, these areas should be a more reasonable and efficient use of blue water resources. Moreover, the effective utilization coefficient of irrigation water in China is relatively low. In 2019, water consumption for agricultural irrigation was 238.76 billion m³, and the effective utilization coefficient was only 0.559 (Ministry of Water Resources of the People's Republic of China 2020). To this end, China has vigorously promoted water-saving irrigation to achieve maximum yields with minimum water consumption, such as drip irrigation and sprinkling irrigation. It is necessary to make reasonable design according to the water demand characteristics, growth stage, climate and soil conditions of the corresponding plants, and formulate corresponding irrigation systems, timely, appropriate and reasonable irrigation.

In this study, we took seven crop and crop products and five animal products into consideration. In general, the of crop products accounted for 71.27%, animal products accounted for 28.73%; the of crop products accounted for 62.26%, animal products accounted for 37.74%. Table 2 shows the and by food structure in all provinces. The food with the largest water footprint was grain, followed by vegetables, contributing 34.77 and 21.25% to the total respectively. Provincially specific, a few provinces do not have the highest water footprint of grain production, for example, edible vegetable oil production in Guangdong and Tianjin, and meat production in Qinghai and Yunnan contributed the largest, as well as vegetable production in Beijing, Fujian, Guangxi, Hainan, Hebei and Zhejiang. From the point of food consumed, the first is grain consumption, followed by pork, beef and mutton consumption, contributing 36.28 and 23.19 to the total respectively. There are still a few exceptions. For example, pork, beef and mutton consumption in Fujian, Guangdong, Guizhou, Shanghai and Zhejiang contributed the largest water footprint.

Assuming that all food consumption is produced locally, the difference between the and can be found for all provinces, only vegetables have no virtual water input from outside. The virtual water input of the other products is shown in Table 3. To be clear, this is an ideal hypothetical scenario based on production and consumption. In the real economic market, the trade of products is affected by many factors, such as political, economic and cultural factors. However, in the context of severe water shortage, the virtual water trade can be used as a new means of trade regulation, which plays an important role in realizing optimal allocation of water resources and alleviating water shortage. Moreover, it is important to note that China's grain production is used for a variety of purposes. In addition to household rations, it can also be used for industrial grain, forage grain, seed grain, as well as a certain proportion of ineffective loss in the links of sowing, harvesting, transportation, storage, processing and eating. Therefore, if considering all consumption channels, in addition to Beijing, Guangdong, Qinghai, Shanghai and Tibet, there are still many provinces that belong to the grain virtual water import zone.

On a national scale, the water footprint of each produced food showed an upward trend. However, as the unit water footprint of various products is different, its proportion in the total water footprint is constantly changing. For edible vegetable oil, for instance, its production water footprint has increased from 7.39% in 2001 to 17.65% in 2019, while that of meat (pork, beef, mutton) has decreased from 20.60 to 14.19%. The water footprint of edible vegetable oil production basically took 2008 as the boundary, stable growth from 2001 to 2008, and rapid growth after 2008. In 2008, the price of vegetable oil in China experienced unprecedented fluctuations, and the risk of oil enterprises was obviously higher than that in previous years. In 2008, the import volume of vegetable oil and oilseeds hit a new record again. In 2008, on the basis of increasing the reserves of imported vegetable oil, China acquired and stored rapeseed in the south rapeseed producing areas and soybeans in the north soybean producing areas, so as to enrich the market capacity of oil regulation. After overcoming many adverse influences such as rising prices of raw materials and abnormal fluctuations of the oil market at home and abroad, China's oil industry has achieved rapid development speed and better economic benefits based on scientific decision-making and prudent operation. In addition, the of grain has been dominant, with an increase but little change. On the contrast, the proportion of of grain products has been declined greatly, from 54.18% in 2001 to 30.26% in 2019. As China's grain output has been increasing year by year, it has greatly exceeded demand, and grain stocks have been rising. It is estimated that with the current diet structure, only 71% of rice and 52% of wheat production would be needed to fully meet ration requirements (Xiao et al. 2017). The proportion of vegetable has also fallen while the proportion of edible vegetable oil, poultry, egg and meat increased. On the whole, the food consumption structure tends to be diversified and balanced.

In both food production and consumption, soybean is a special product because it is the largest import. In addition to oil pressing, the main reason for importing so many soybeans are to use the soybean meal left after oil pressing as protein feed for livestock, poultry and aquaculture industry to meet the rising demand for meat, eggs and milk of residents. This has led to many virtual water imports (Dalin et al. 2012). It is estimated that the virtual water inflows from soybean imports accounted for about 33% of the total virtual water imports in terms of agricultural products (Zhang et al. 2016). A large number of soybean imports can alleviate the water stress in China from the perspective of virtual water. However, from the perspective of trade, the high dependence on the international soybean market is also prone to trade frictions. Soybean has a very important position in China's agriculture. It is not only the most important source of edible vegetable oil and protein, but also an important industrial raw material and livestock feed. Starting in 2019, China implemented a soybean revitalization plan to encourage farmers to grow soybeans. In addition, spring tillage production of soybean is very important for high yield of soybean. Soybean farmers should understand the key techniques of soybean planting and cultivation management in spring, and conscientiously implement the key techniques of cultivation management, such as variety selection, planting mode selection, soil cultivation, sowing, soil fertilizer and water management, disease and insect pest prevention and control.

Food and water security are vital for maintaining the stable development of economy and social stability. The study analyzed and in China from a multi-dimensional and conducted comparative study. In terms of time and space scale, this study analyzes the temporal changes and spatial patterns of and in provinces from 2001 to 2019. In terms of structure, the green, blue and gray water, seven kinds of crop and crop products and five animal products were calculated from water resource structure and food structure. However, there are some limitations at present. At first, for some missing data, we used the existing data for calculation or estimation, and the final data would be slightly different, but it would not affect the final results and conclusions. Second, the water footprint of unit products comes from reports 47 and 48, which does not highlight the inter annual change. We will pay attention to this limitation and strive to achieve more accurate and detailed research results of water footprint. In addition, the consumption of grain only calculated the consumption of residents' rations. In the future, the research on supply and demand balance should take into account industrial grain, feed grain, seed and loss. To conclude, this study clearly shows the and of the main food in China since 2001, which has important reference value for China's food policy and water resource management. In particular, at this stage, influenced by the global COVID-19, the import and export trade of products is blocked, and the food policy is necessary to make corresponding adjustments. How to achieve coupling coordination of food security, nutrition and health of diet and high water use efficiency is a matter of concern.

In the physical water sense, agriculture is by far the highest water user in China. When virtual water is taken into account, the total amount of agricultural water used is more than three times that of physical water. Therefore, sustainable food production from a water usage perspective is very necessary. First of all, we should improve and strengthen the water use efficiency in grain production, and strive to produce more agricultural products with less water resources, using scientific and technological means to improve crop yield per unit area, for example, artificial lighting, temperature and humidity control. Besides, to ensure the standardization of agricultural water use and high efficiency, it is necessary to build rural water conservancy facilities and do a good job in water storage and water saving. In irrigated agriculture, drip irrigation and micro-sprinkling irrigation need to be developed. Moreover, a comprehensive pricing reform for agricultural water would promote water saving, which has a certain influence on improving water efficiency and decreases the water footprint per unit of product (Chapagain et al. 2006).

Secondly, the application efficiency of chemical fertilizers and pesticides to reduce the gray water footprint should be improved. Farmers linked the amount of pesticide fertilizer application to agricultural output in the early stage, and considered that the widespread use of chemical fertilizer was an important driving factor for grain yield. However, the high-intensity application not only leads to the increase of grain production cost, but also intensifies the contradiction between food security and ecological security (Godfray et al. 2010; Foley et al. 2011). Excessive application of pesticides improves the resistance to diseases and insect pests, which in turn promotes the iterative upgrading of pesticide products, forcing the continuous increase of pesticide concentration and use frequency, thus leading to a vicious circle of the whole rural ecological chain. Therefore, we should carry out targeted training on fertilizer application technology for farmers, regularly test soil fertility and nitrogen, phosphorus, potassium and other elements, and timely feed back the basic elements of soil to farmers through technical training and guidance services, so as to promote soil testing and formulated fertilization, correct farmers' cognitive bias on fertilizer, and avoid overuse and overuse of fertilizer. At the same time, the propaganda of eco-friendly fertilizer should be strengthened, and the combined application of organic fertilizer and chemical fertilizer should be encouraged.

Thirdly, the planting structure of agricultural products and realizing rational allocation of water resources needs to be optimized. Under the indicators of water footprint and water stress level of each province, agricultural products with comparative advantages are planted, and the intensity of agricultural production in water-deficit areas by planting crops that use less water is reduced. For example, in Northwest China we should develop potato crops with low water consumption and increase investment in water-saving and heat-preserving technology and agricultural facility construction. In the North China Plain, we will expand the area with water-saving and drought-resistant crops, increase investment in agricultural technology and of farmer training. In the south regions of the Yangtze River, which is rich in water and heat resources, the construction of high-standard farmland facilities should be increased, and suitable crops such as rice, vegetables and fruits should be supported to meet the increasing and diversified needs of Chinese food consumption. In addition, we should make rational use of the international market and moderately increase the water-intensive products import. Within the carrying capacity of domestic resources and environment, production potential and demand for agricultural products, we will set reasonable targets for self-sufficiency and priorities for the import of agricultural products, make reasonable arrangements for the type and quantity of imports, keep the pace of imports at a proper pace, maintain the stability of the domestic market, and ease the pressure on domestic resources and the environment.

Population, diet structure and the quantity of food consumed are the most direct factors affecting the ⁠. In 2021, in order to further optimize the fertility policy, China implemented the policy of one couple having three children and supporting measures. It is expected that the population of China and each province will continue to grow in the future. Therefore, to reduce the ⁠, we must adjust the diet structure and the quantity of food consumed by firstly optimizing the diet structure from nutrition and water conservation perspectives. We should not blindly reduce consumption of animal products with a relatively high unit water footprint, which may lead to the imbalance of nutrient intake. Balancing nutrition and water resources under the guidance of The Chinese Dietary Guidelines is necessary. It advocates eating a variety of foods, including plenty of vegetables, fruits and soybeans with a relatively low water footprint per unit; eating moderate amount of poultry, eggs and lean meats with a relatively high water footprint per unit (Mekonnen & Hoekstra 2010a, 2010b). In the future, we will calculate the water footprint according to The Chinese Dietary Guidelines, and compare it with the of each province, and put forward specific suggestions for optimizing the diet structure.

Secondly, personal food consumption is also a very important factor, especially food waste (Reganold & Wachter 2016; Hasegawa et al. 2019). In the process of production and consumption, much food is lost or wasted. Globally, FAO estimates that the food waste is 1.6 billion tons of ‘primary product equivalent’, while the total waste of edible parts is 1.3 billion tons. From a blue water perspective, the volume of water wasted due to food waste is about 250 billion m³. What is more, food waste is balanced between upstream and downstream of the food supply chain. Upstream waste accounts for 54% of total waste and downstream accounts for 46% (FAO 2013). Taking 2010 for instance, about 19% of food in China was lost or wasted, equivalent to 135 billion m³ of water, and consumer waste contributed the most to total food waste (Liu et al. 2013). The average waste rate in Chinese households is 11.28% (Li et al. 2021). It is estimated that if Chinese permanent residents try to reduce food water by 5 g per day, it would save 2.6 million tons of food and 1.79 billion m³ of water per year (Song et al. 2015). At the same time, some food service providers, such as restaurants and hotels, should be deeply aware of the seriousness of food waste. There is a serious food waste in the provision of hotel buffet, which increases the ineffective consumption of food resources, such as the leftovers that can only be poured out. Residents should also establish a sense of saving when dining out, reduce food waste and encourage packing.

Finally, we also need to pay attention to the virtual water consumption caused by energy consumption in the process of food cooking. Cooking generates energy consumption, such as electric energy, natural gas and coal gas consumption, which will lead to virtual water consumption hidden in the energy source. Both residents and catering enterprises should pay attention to energy conservation such as adopt more efficient ways to cook food, and reduce the virtual water of energy consumption, such as changing the way you cook and choosing energy-efficient pans (Carlsson-Kanyama et al. 2003).

Food and water security are global problems relating to the survival and development of mankind. In order to meet future food demand, food production must be increased, which is constrained by the rigid constraints of water supply. As a populous and agricultural country, China is facing more severe pressure on water resources and food security. Food security involves different links such as food production and consumption. Therefore, examining China's food production and consumption from water perspective is of great significance for formulating scientific agricultural production layout, optimizing dietary structure and international trade policies, as well as coordinating the relationship between ecological civilization construction and food security.

The study analyzed the food water footprint of production and consumption in China from the points of spatial, temporal and structure. From 2001 to 2019, the total and in China showed an upward trend. The production water footprint was about 1826.42 billion m³ per year while that of residents' consumption was 615.78 billion m³ per year. The huge difference is due to the variety of uses of grain in China. In addition to household rations, it can also be used for industrial grain, forage grain, seed grain, as well as a certain proportion of ineffective loss. Due to the production conditions, there were regional differences and a mismatch between water footprint and water stress among provinces. Shandong, Henan and Hebei contributed the most to the production water footprint and also were high water stress areas. However, there does not appear to be a significant correlation between and water resources stress. In terms of water types, green water contributed the most. The gray water was relatively high which indicates high pollution in food production. Specifically the water footprint of plant-based food contributed the most, both in production and consumption generally.

The study clarified the water used in different provinces and food structure and the types of water resources used, and put forward suggestions on sustainable production and consumption, which can provide reference for water resources management, agricultural production and diet optimization. The water footprint and its structure of unit products in different provinces are different. The measurement of production water footprint can provide reference for water resources management in various provinces and optimize the structure of green, blue and grey water. For agricultural production, it is necessary to improve production technology, improve the utilization efficiency of chemical fertilizers and pesticides, and adopt advanced irrigation technology to reduce the waste of water resources. The measurement of water footprint of residents' food consumption reveals the impact of diet on water resources and environment. The diet structure should be optimized on the basis of nutrition balance to realize the harmony between nutrition balance and environmental effect.

This work was supported by Shandong Provincial Natural Science Foundation (ZR2020QD014, ZR2021MD076), and Youth Science and Technology Innovation Program of Shandong Provincial Education Department(2020RWG010).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.