Spatiotemporal variations of agricultural water footprint and its economic benefits in Xinjiang, northwestern China

Agriculture is the largest water user and is the main driving force behind water stress in Xinjiang, northwestern China. In this study, the water footprint (WF) (blue, green and gray WF) of main crop production and their temporal and spatial characteristics in Xinjiang were estimated in 2006, 2010, 2014 and 2018. The blue water footprint deficit (BWFd) was conducted and food productivity and economic benefits of WF were also analyzed via the water consumption per output value (food productivity and economic benefits). The results reveal that the WF increased from 22.75 to 44.16 billion m3 during 2006–2018 in Xinjiang, of which cotton, corn and wheat are main contributors of WF. In terms of different regions, corn has the largest WF in north Xinjiang and cotton has the largest WF in south and east Xinjiang. The BWFd broadened from − 11.51 to + 13.26 billion m3 in Xinjiang with the largest increased BWFd in Kashgar (from − 3.35 to 1.40 billion m3) and Aksu (from − 2.92 to 2.23 billion m3) of south Xinjiang and in Shihezi (from − 0.11 to 2.90 billion m3) of north Xinjiang. In addition, the water footprint food productivity does not well correspond with the water footprint economic benefits in prefectures of Xinjiang. It means we should consider the food yields priority and economic benefits priority to formulate a scientific and effective supervisor mode to realize the sustainable management of agricultural water in prefectures of Xinjiang.

www.nature.com/scientificreports/ in 2010 and 472 × 10 8 m 3 in 2015. Li et al. 5 found the total WF r of Xinjiang was 389.9 × 10 8 m 3 in 2018. Zhang et al. 18 proposed that the WF r increased from 73.91 × 10 8 m 3 in 1988 to 270.50 × 10 8 m 3 in 2015. We can find that the WF r calculated by different groups in the same year are inconsistent because of the selected different crops and parameters (e.g., irrigation coefficient). For example, Shen et al. 16 chosen wheat, corn, cotton, oil crops and soybeans, while wheat, corn, cotton, rice and fruit trees were selected in Wang et al. 17 . The averaged irrigation coefficient in Xinjiang was used by Shen et al. 16 and Wang et al. 17 , while Li et al. 5 used the irrigation coefficient of different prefectures. Several obvious deficiencies are also discovered: (1) WF gray is the important contributor to Xinjiang production due to the extensive use of chemical fertilizer 19 , but their data are relative scarce for Xinjiang crops; (2) few studies estimated the difference between the actual irrigation water and WF blue to measure the extent of blue water scarcity in Xinjiang; (3) the water footprint per unit of yield (WY) received more attention to reveal water productivity from the perspective of food production, while few studies describe water productivity from the perspective of economic benefits via the water footprint per output value (WV).
In this study, we are planning to address the above three issues: (1) the WF blue , WF green and WF gray for main 11 crops were calculated and their temporal and spatial characteristics were investigated in 2006, 2010, 2014 and 2018; (2) the new indicator-blue water footprint deficit (BWF d ) 3 were used to reveal the actual situation of water shortage in Xinjiang; (3) the WF per unit of yield (WY) and WF per unit of output value (WV) were calculated to estimate water footprint food production and economic benefits in Xinjiang. Our works could enrich indicators about water footprint of crop production and also offer an optional way for agriculture saving water in Xinjiang.
The climate is characterized by a temperate continental climate with mean annual precipitation of < 200 mm and mean annual temperature of 9.1 °C. The oases are mainly situated in the piedmont plains and their water resources primarily result from rivers originating from precipitation and from glacial and snow melt water in the mountainous regions (i.e., Tianshan, Altai and Kunlun Mountains). The cultivated lands are distributed in the oasis regions. Water use for agricultural irrigation has been accounted for > 90% of freshwater use in Xinjiang, which is much higher than the average level of China. The planting area and yield of 11 main crops in Xinjiang were shown in Supplementary Table S1. The irrigation area of Xinjiang increased from 3856.91 × 10 3 ha to    The contribution rates of WF were different in south Xinjiang, north Xinjiang and east Xinjiang (Fig. 4). In south and east Xinjiang, the largest contribution rate for WF among crops was cotton, accounting for 52.26-63.81%, while the smallest rate was potato, accounting for 0.10-0.72%. In north Xinjiang, the largest contribution rate was corn with an increase from 12.33 to 21%, while the contribution rates of wheat and vegetables respectively decreased from 16.33 and 6.81% to 12.79 and 3.61%. The smallest contribution rate was also potato, only accounting for 0.37-0.73%.

Trends of water footprint (WF
In terms of WF blue , corn is major contributor among crops and increases from 11.97 to 20.16% in north Xinjiang. Being converse with corn, the contribution rate of wheat has a decreasing trend from 14.44 to 11.08%. No obvious changes were showed in other crops. In south Xinjiang, the contribution rate of cotton increased from 55.83 to 63.02%, while that of corn and wheat decreased from 16.74 and 15.67% to 12.11 and 11.45%. In east Xinjiang, WF blue of cotton, corn and wheat decreased significantly from 60.40%, 3.63%, 12.30% to 53.44%, 1.16%, 10.10%.
In terms of WF green , the contribution rates of cotton and corn were increased from 32.29 and 15.18% to 39.16 and 24.16%, and it was decreasing in other crops with different rates, of which the contribution rate of wheat decreased obviously from 25.33 to 15.96% in north Xinjiang. Being consistent with changes of cotton in north Xinjiang, the contribution rate of cotton increased from 46.54 to 52.17% in south Xinjiang. Corn and wheat have decreasing trends from 19.23 and 22.71% to 16.26 and 16.80%, respectively. The observed changing trends of crops in east Xinjiang are similar with that in south Xinjiang.
In terms of WF gray , the contribution rate of cotton was decreasing from 47.30 to 37.90%, while that of corn increased significantly from 11.86 to 23.91% in north Xinjiang. The contribution rates of other crops were basically stable. The contribution rate of cotton was increasing from 50.35 to 51.65% in south Xinjiang. In east Xinjiang, the changeable trends of the contribution rate of crops were the same as those in south Xinjiang. www.nature.com/scientificreports/ Based on the above analysis, the main contributions of WF come from cotton, corn and wheat in Xinjiang during 2006-2018. The increasing contributions of WF in north Xinjiang were from WF blue and WF gray of corn and WF green of cotton and corn, while those in south Xinjiang and east Xinjiang were from WF blue , WF green and WF gray of cotton. The contribution rate of WF blue was much higher than that of WF green and WF gray , which suggests these planting crops were mainly depended on irrigation water in Xinjiang.

Blue water footprint deficit (BWF d ).
During 2006-2010, the BWF d increased from − 11.51 to − 6.28 billion m 3 due to the expansion of planting areas. Although the water-saving technologies are beginning to be applied (e.g., dropper technology) in Xinjiang, this was not enough to offset the rapid increase of water demand due to the expansion of irrigation farmland since 2010. The rapid expanded irrigation farmland results in the larger BWF d from − 6.28 billion m 3 in 2010 to 13.26 billion m 3 in 2018, leading to more severe shortage of blue water in Xinjiang.
Regarding the different prefectures in three regions (Fig. 5), the BWF d in south Xinjiang was lower than that in north Xinjiang. In south Xinjiang, the largest increases of BWF d were showed in Kashgar (from − 3.  www.nature.com/scientificreports/     (Fig. 7). In terms of the changing trends for crops, the largest decrease of 68.18% from 0.02 to 0.01 m 3 /kg was showed in soybean, while the smallest decrease of of 5.71% from 0.12 to 0.11 m 3 /kg was found in wheat. Corn has the largest increase from 0.11 to 0.16 m 3 /kg. Regarding the contribution rate in three regions (Fig. 8), the rate slightly decreased for WV green from 9.46 to 8.42% and increased for WV gray from 14.21 to 15.31% in north Xinjiang. The ratio of WV blue remained stable at around 76%. In south Xinjiang, the ratio of WV blue decreased from 80.74 to 78.73%. The ratio of WV green increased from 3.58% in 2006 to 9.86% in 2010 and then decreased to 6.37% in 2018. The ratio of WY gray decreased from 15.68 to 14.89%. In east Xinjiang, the ratios of WV green and WV gray increased from 1.53 and 6.12% to 11.68 and 17.64%, respectively. The ratio of WV blue decreased from 86.  (Fig. 9). The WV of cotton had a tendency of decreasing, increasing and then decreasing. The WV of other crops decreased continuously. Being consistent with the ratios of WY green , WY blue and WY gray , the ratio of WV blue was the highest (mean 40.78%) in cotton, while the lowest (mean 0.25%) was found in potato.  (Fig. 2). The rapid expansion of agricultural planting scale is the fundamental reason for the large increase of crops WF in Xinjiang 8,20,21 . The WF were showed from high to low in Xinjiang: WF blue > WF gray > WF green . It means that WF blue is the most important water consumption and the role of WF gray should be concerned in next researches because of WF gray > WF green . In terms of the changeable rates, their increasing trend is also WF blue (160.20 million m 3 / yr) > WF gray (26.  www.nature.com/scientificreports/ In terms of the increased rates of WF in three regions, they were significantly different: south Xinjiang (0.79 billion m 3 /yr) > north Xinjiang (0.82 billion m 3 /yr) > east Xinjiang (0.03 billion m 3 /yr) (Fig. 3). The WF in south Xinjiang (22.34 billion m 3 in 2018) is consistently higher than that in north Xinjiang (20.61 billion m 3 in 2018) in the studied interval, the gap between them was further shrinking from 2.11 billion m 3 in 2006 to 1.73 billion m 3 in 2018 due to its better agricultural development conditions, faster agricultural infrastructure construction and better large-scale operation foundation in north Xinjiang 16 . Correspondingly, the increased rate is WF blue > WF gray > WF green in three regions. For WF blue and WF gray , the rates are south Xinjiang > north Xinjiang > east Xinjiang. It means that the crop irrigation water consumption and fertilizer consumption in south Xinjiang were higher than those in north Xinjiang. For WF green , the rates are north Xinjiang > south Xinjiang > east Xinjiang. The increasing rates are consistent with significantly increasing precipitation in Xinjiang during 2006-2018 22 , which lead to higher effective precipitation (90-95%) in the conditions of the small amount  www.nature.com/scientificreports/ of deep seepage and the majority surface-soil interception 8 . The expanded planting area also plays an significant role in an increase of WF green , which makes the area of crops can withstand precipitation increase and then makes the utilization of effective precipitation increase correspondingly 18,19 . In addition, the replacement of natural oases by many expansion of planting areas results in the conversion of ecological environmental water into artificial consumption in terms of water consumption 8  Food productivity and economic benefits of water footprint. The issues of food safety have received increasing attention in the context of the explosive growth of population in Xinjiang and the availability of fresh water is the biggest challenge to food production. GDP is mainly affected by the fluctuating crop prices and the labor cost. From different angles (food productivity and economic productivity) of water footprint, it provides an alternative way for Xinjiang agriculture to save water.

Discussion
As shown in Figs . This means water footprint food productivity in Aksu was lower than that in Shihezi, but the water footprint economic productivity was reversed. Two aspects should be proposed based on the above analysis: (1) food productivity is well corresponded with economic benefits among crops and three regions, which means WV or WY plays a equivalent role in government and farmers' decisions about crop planting structures in Xinjiang; (2) different policies should be made from different perspectives (food productivity and economic benefits) in prefectures of Xinjiang, being consistent with the results of Zhangjiakou City 3 .
Next works. Xinjiang WF experienced a continuous increasing trend during 2006-2018. The rapid expansion of agricultural planting scale is the fundamental cause for a significant increase of WF for crops in Xinjiang. Under the condition of available water shortage in Xinjiang, we focus on a scientific view for the adjustment and transformation of crop structures to reasonably allocate water resources in Xinjiang. However, two following aspects should be done in future: (1) the spatial-temporal matching characteristics between water footprint and socioeconomic factors in each prefecture are needed to analyse using mathematical models (e.g., Gini coefficient and imbalance index 23 ); (2) the specific planting area of crops are available via the remote sensing and the field investigation. Emphasizing comprehensive consideration of a variety of social and economic factors and detailed planting distribution of crops, we can provide a detailed plan to put forward suitable measures and policies to adjust the crop structure for sustainable agricultural development in prefectures of Xinjiang.
Data sources and methods. Data sources. Total 66 meteorological stations for Xinjiang were collected from China Meteorological Administration (http:// data. cma. cn/). The selected parameters include the maximum temperature, the minimum temperature, mean monthly temperature, mean monthly precipitation, wind speed, air pressure, relative humidity and sunlight duration. The land use/cover data were downloaded from National Earth System Science Data Center, National Science & Technology Infrastructure of China (http:// www. geoda ta. cn). The NLCD maps (1 km) were produced by visual interpretation of Landsat Thematic Mapper (TM) images 24 . Socioeconomic data (crop types, planting areas, crops yields and regional GDP) were compiled from Xinjiang Statistics Yearbook (China). Irrigation water consumption for prefectures in Xinjiang were obtained from Xinjiang Water Resources Bulletin (2006, 2010, 2014 and 2018).

Methods
Experimental design. The water requirements of 11 crops in their growing period were firstly calculated using the CROPWAT 8.0 and then the temporal and spatial features of water footprint of crop production were estimated in 2006, 2010, 2014 and 2018. Combined with the irrigation water consumption, the blue water footprint deficits were calculated to reveal the situation of blue water. Finally, water footprint per unit of yield and water footprint per unit of GDP were estimated to depict water productivity from the perspective of food production and economic benefits in Xinjiang.
Data processing. Water footprint (WF) for crop production. It consists of blue water footprint (WF blue ), green water footprint (WF green ) and gray water footprint (WF gray ). The total WF of 11 major crops in Xinjiang is evaluated based on the calculation method of water footprint showed in 25  www.nature.com/scientificreports/ where WF blue is surface and ground water consumed (evaporated) by the production of a commodity; WF green is the consumption of green water during the growing period of crops, green water is actually the total amount of rain evaporation. WF gray is a product refers to the amount of fresh water required to assimilate contaminants according to existing environmental water quality standards.
Green water footprint (WF green ) and blue water footprint (WF blue ). To calculate WF green and WF blue , reference crop evapotranspiration (ET 0 ) was calculated through meteorological elements and crop evapotranspiration (ET c ) was calculated through crop regulation coefficient (K c ) 26 . Crop evapotranspiration includes evaporation of soil surface and transpiration of crop. The specific formula is as follows 26 .
In this equation, ET crop is crop evapotranspiration (mm/day); ET blue is the evapotranspiration of crop blue water; ET green represents the evapotranspiration of green water; K c is the crop regulation coefficient (dimensionless); ET 0 is the reference crop evapotranspiration (mm/day); factor 10 is the conversion of the depth unit mm to the volume unit m 3 of water; A is the crop planting area; lg P d=1 ET crop is the total evapotranspiration in the growing period of crops from the sowing date (the first day) to the harvest date, lgP represents the number of days in the growing period.
ET green was determined by comparing the potential evapotranspiration and effective precipitation (P e ) during the growing period of crops, when ET c > P e , ET green = P e , ET blue = ET c -P e ; When ET c < P e , ET green = ET c , ET blue = 0.
Gray water footprint (WF gray ). In this study, we adopts the groundwater quality standard (GB/T14848-93) and water quality standard for irrigation (GB5084-2005) in China, that is, nitrate (N) is less than 0.02 g/L, farmland irrigation water salinity in general should not be higher than 1.7 g/L. We selected C max = 1.7 g/L, the ambient background concentration of nitrogen is assumed to be 0, and the specific calculation formula of gray water is as follows 27 : where WF gray is the gray water footprint (m 3 ); AR is the pure amount of nitrogen fertilizer, kg; Ə is the nitrogen leaching rate, %; C max is the maximum environmental allowable concentration of nitrogen fertilizer (kg/m 3 ); C nat is the initial concentration of pollutants in water (kg/m 3 ).
Blue water footprint deficit (BWF d ). Due to water shortage and imperfect water supply infrastructures, crops can't always be fully irrigated in Xinjiang, WF blue may not record the extent of blue water scarcity. In order to distinguish the actual blue water footprint consumption (WF blue ′) from the crops requirement of blue water footprint, we used the blue water footprint deficit (BWF d ) introduced by Ma et al. 3 in this study, which refers to the difference between WF blue ′ and WF blue .
When BWF d < 0, it means a situation of blue water surplus. When BWF d > 0, it means a situation of blue water shortage.
where W i refers to the irrigation water, ƞ is the effective utilization coefficient in each prefecture. www.nature.com/scientificreports/ Water footprint per unit of yield (WY). The WF per unit of yield (WY) is the WF divided by the crop yield (Y) and includes WF blue per unit of yield (WY blue ), WF green per unit of yield (WY green ) and WF gray per unit of yield (WY gray ) 3 .
Water footprint per unit of GDP (WV). The WF per unit of GDP (WV) is the WF divided by GDP 3 . It also includes three parts (WV blue , WV green and WV gray ), which reflects the economic benefits of WF.