Water Resources Evaluation in Arid Areas Based on Agricultural Water Footprint—A Case Study on the Edge of the Taklimakan Desert

Abstract

Water scarcity is an important factor limiting agricultural development in arid areas. Clarifying and evaluating the current situation of water resources in arid regions is helpful for decision-makers in the rational use of water resources. This study takes a typical arid region located at the edge of Taklamakan Desert-Hotan region as the study area. The water footprint (WF) of the Hotan region was calculated based on 20 years of data information from 2000–2019. An evaluation system was established using four aspects of the WF: structural indicators, efficiency indicators, ecological safety indicators, and sustainability indicators. The results show that the WF of the study area is mainly dominated by blue water consumption, with a proportion of 65.74%. The WF of crop production is larger than that of livestock production. The produced WF of grain crops is the highest of all products with a share of 44.21%. The increase in the local agricultural WF reached 53.18% from 2000 to 2019, but it was still lower than the amount of water available for agriculture. The evaluation results indicated that the region’s WF import dependency is lower than the global level, with an annual average self-sufficiency rate of 91.13% and an increase of 878.95% in the WF economic efficiency index. The agricultural WF produced in Hotan is exported in the form of trade, but the quantitative contribution is small and does little to relieve water stress in other regions. The agricultural water consumption was still within the range of local water resources that could be carried but only 6 years of sustainable water use, and the future development was not optimistic. With the ratio of produced WF to available water resources maintained at about 58%, the local available water resources should be above 43.21 × 10⁸ m³ to initially ensure the sustainable use of water resources. There were 12 drought years in the study period, which are prone to droughts and high disaster levels. The drought-water scarcity systems behaved in three phases: 2000–2011 (uncoordinated level), 2012–2015 (transitional phase), and 2016–2019 (coordinated level). Water scarcity threatened by drought reduced. The occurrence of meteorological droughts was more related to natural factors while the changes in WF were mainly driven by socio-economic elements such as human activities.

1. Introduction

Water scarcity has become a worldwide problem that has seriously affected global food security and ecosystem health [1,2,3]. The scarcity of water resources will make the protection and security of water resources face serious challenges and increasingly affect food security [4]. Water footprint (WF) is an indicator of water use that characterizes the amount of water required for all goods and services consumed by any known population (a country, a region, or a person) over some time [5]. The WF includes both physical and virtual water, and no longer simply counts visible water such as surface water and groundwater.WF can reflect the actual demand for and use of water resources in a region or country, and the consequences of human and social dependence on water resources and the environment. As a result, WF is also becoming a more useful tool for relieving water stress in water-scarce areas [6]. WF consists of three components: consumption of surface and groundwater (blue WF), consumption of rainwater stored in the soil as soil moisture (green WF), and the amount of freshwater needed to assimilate pollutant loads (gray WF) [7,8,9].

Agriculture is the largest user of water and accounts for 70–75% of global water use [10]. Current research on agricultural water footprint (AWF) generally focuses on three areas. The first one is studying the AWF production of typical and major crops or comparing the AWF in different regions to provide new suggestions for solving water scarcity problems and optimizing water management [11,12,13,14,15,16]. The second is to assess the pressure of agricultural production activities on water resources, such as the structure and efficiency of water use based on AWF, to optimize the agricultural cropping structure [17,18]. The third is to initially explore the influence of factors such as meteorology and agricultural inputs on AWF, to regulate agricultural inputs and achieve efficient water use [19,20,21,22]. The WF theory has enriched and developed the traditional water consumption evaluation system, and the study of AWF provides new research ideas and methods to alleviate the pressure on regional water resources and analyze the new challenges facing food security in the study area to achieve regional sustainable development [23,24].

To date, existing studies have focused on water-scarce basins [25,26,27] and provinces [28,29] for a single study of WF changes in drylands or a simple water resource evaluation. Few studies have combined WF ecological safety evaluation with meteorological drought to further investigate the differences in WF under different degrees of drought conditions in extremely arid regions. Examples of existing WF evaluation systems such as Qi Rui et al. constructed a regional water resources utilization evaluation index system based on WF and carried out an empirical analysis using Dalian city as an example [30]. Liu et al. built a water resources security evaluation index system for the Yangtze River Economic Belt based on the conceptual framework of Driver-Pressure-State-Impact (DPSI) [4]. WF-based water resources ecological security evaluation is hot but with the increasing level of socialization water-poor areas urgently need to analyze and evaluate the regional water resources carrying level comprehensively under the premise of ensuring regional food security and resource sustainability. Drought is a non-negligible influencing factor in water-poor areas, and the relationship between the rational use of water resources and drought has become a topic of widespread interest. Numerous scholars have studied the correlation between drought index and agricultural activities [31,32,33] but few scholars have combined drought index with WF evaluation. In this study, a suitable standardized precipitation evapotranspiration index (SPEI) was selected to investigate the relationship between AWF and drought from a new perspective to further explore the ecological security of water resources in extremely arid regions, providing a new idea for a comprehensive study of agricultural water consumption and drought prevention in arid regions.

China’s agricultural water consumption accounts for 74.2% of the total water consumption in the next 20 years China’s water resources will be scarcer [34,35]. Limited water resources are also a major challenge to food security in China’s arid zones, water exploitation in agricultural production in arid lands such as Xinjiang affects the ecosystem, and high-water consumption and low-income food industry bring huge water and ecological pressure [36,37,38]. The Taklimakan Desert has an extremely arid climate and a vulnerable ecosystem [39]. Hotan oasis is a typical hyper-arid region located on the edge of the Taklimakan Desert. The total regional water resource in 2020 was 10.90 billion m³, and the total water supply was 4.21 billion m³. Hotan is highly dependent on agricultural economic development. Agricultural irrigation is an important part of water resource utilization, and agricultural water consumption occupied 92.14% [40]. Therefore, the water shortage has been an important factor limiting the development of the Hotan region.

To study the utilization of agricultural water resources and scientifically quantify the pressure on water resources brought by agricultural production activities, we analyzed the dynamic changes of AWF in Hotan from 2000 to 2019 and used WF as a medium to make a comprehensive evaluation of water resources and food security in Hotan region from the perspectives of agriculture, ecology, economy, and climate, and constructed water resources carrying capacity evaluation system based on WF. A coupled coordination model was also used to analyze the ecological security of water resources in Hotan area in combination with the SPEI. This study also provides a new perspective for further analysis of the relationship between agricultural growth and water resources development and utilization, to provide some scientific reference for the efficient utilization of regional water resources.

2. Materials and Methods

2.1. Study Area

Hotan is located in the hinterland of Eurasia, the southernmost part of Xinjiang Uygur Autonomous Region (77°20′~85° E, 34°~39°30′ N), as shown in Figure 1. The total area is 24.74 × 10⁴ km², of which 63% is the Gobi desert, and only 3.7% is an oasis. Hotan is in a warm-temperate, extreme arid desert climate zone. The average annual precipitation is less than 50 mm but annual evaporation may reach 3000 mm [41]. The annual sunshine hours in this area are 2470–3000 h. The vegetation cover is low, and the sandy soil is exposed to a large area, which makes sand and dust disasters frequent.

Most of the rivers in Hotan area are endorheic rivers, among which 30 rivers can supply water for irrigation of farmland and human life, and the surface runoff recharge mainly relies on glacial snow melt and some high mountain precipitation. The Hotan area is dominated by agricultural activities, and the main agricultural and animal products are wheat, cotton, melons, fruits, beef, mutton, etc. The multi-year average agricultural water use was over 95% of the total water used, with irrigation water for crop cultivation and animal husbandry water for animal breeding comprised of agricultural water use in a ratio of about 7:3 [42].

2.2. Date

The crops selected for this paper are according to the main crops grown in Hotan area in the statistical yearbook of Xinjiang Uygur Autonomous Region, including grain crops (rice, wheat, corn), cotton, vegetables, fruits (sweet melons, grapes, apples, pears, jujubes, peaches, apricots), and livestock products including six categories of pork, beef, mutton, poultry meat, milk, and eggs, to analyze the WF changes of different agricultural products in Hotan area from 2000 to 2019. The data on agricultural products and GDP required for the analysis and evaluation of WF was obtained from the Xinjiang Statistical Yearbook (2000–2020) and Xinjiang Production & Construction Corps Statistical Yearbook (2002–2020) [43,44]; hydrological data were from “Xinjiang Water Resources Bulletin” (2001–2019) [42]; meteorological data such as rainfall, average wind speed, relative humidity, and temperature were obtained from the National Meteorological Science Data Sharing Service Center (http://data.cma.cn (accessed on 1 August 2021)) [45], and missing data were supplemented by international meteorological observations in the CLIMWAT2.0 database. Water footprint for animal products was calculated using data related to water footprint for animal products in China such as Chapagain [9].

2.3. Agricultural Water Footprint Calculation Method

We calculate blue-green water for different product types in the cultivation and livestock sectors based on Hoekstra’s method and also distinguish between internal and external water footprints [5]. The AWF is equal to the sum of direct and indirect water use by residents consuming agricultural products and services, including both internal and external WF, and the agricultural production water footprint (PWF) within the region is used to add (subtract) the amount of agricultural WF input (output) within the region. This calculation method can better reflect the dependence of the study area on external water resources [46]. Because the proportion of gray WF is small, this study focuses on the blue WF and green WF, and the gray WF is not discussed.

$$AWF=IWF+EWF$$

(1)

where IWF is the total amount of water used within the region to produce the region’s agricultural products (internal WF), and EWF is the amount of water consumed by products produced in other regions and consumed by the local population (external WF).

$$IWF=PWF-EWF=PWF+VWI-VWE$$

(2)

$$PWF=W{F}_{cro}+W{F}_{ani}={{\displaystyle \sum }}_{i=1}^n{P}_i\times AW{R}_i$$

(3)

where PWF is the total regional production WF, which is composed of two parts, the crop production WF ($W{F}_{cro}$) and livestock products production WF ($W{F}_{ani}$) sum to get, VWI is the regional import WF, EWF and VWI are numerically equal, and VWE is the regional export water WF. P_i is the production of a certain agricultural product for the year, and AWR_i is the unit water demand of that product.

2.3.1. Crop Water Footprint

Crop Water Requirement (CWR, m³/ha) is the sum of crop and soil evapotranspiration for normal crop growth during the growing period, it is obtained by adding up the daily evapotranspiration (mm/d) during the growing period. We refer to the calculation method in other scholars’ studies [47,48,49] using CLIMAWAT2.0 and CROPWAT8.0 software to calculate AWR, the formulas are as follows:

$$AWR=\frac{CW{R}_{green}}{Y}+\frac{CW{R}_{blue}}{Y}=\frac{10\times {{\displaystyle \sum }}_{d=1}^{lgp}E{T}_{green}}{Y}+\frac{10\times {{\displaystyle \sum }}_{d=1}^{lgp}E{T}_{blue}}{Y}$$

(4)

$$E{T}_{green}=\min \left(E{T}_c, P_e\right)$$

(5)

$$E{T}_{blue}=\max \left(0, E{T}_c-P_e\right)$$

(6)

where CWR_green and CWR_blue are crop green and blue water use (m³/ha); Y is crop yield per unit area (t/ha); ET_green and ET_blue are crop green and blue water requirements (mm); 10 is a constant factor; lgp represents the length of the growing period and measured in days; P_e is the effective rainfall (mm/d); ET_C represents the cumulative evapotranspiration during crop growth and is usually calculated using the standard Penman formula recommended by FAO.

$$E{T}_c=K_c\times E{T}_0$$

(7)

K_c is the crop coefficient; ET₀ is the reference crop evapotranspiration water (mm/d).

P_e (effective rainfall available) usually refers to the portion of precipitation that can meet the evapotranspiration needs of crops and was estimated using the United States Department of Agriculture (USDA) Soil Conservation Service (SCS) method, the formula is as follows:

$$P_e=\left\{\begin{array}{ll}\frac{P_{month}\times \left(125-0.2\times P_{month}\right)}{125}& P_{month}\le 250\\ 125+0.1\times P_{month}& P_{month}>250\end{array}\right.$$

(8)

where P_month is the monthly rainfall (mm).

2.3.2. Water Footprint of Animal Products

The calculation process of $W{F}_{ani}$ is complicated, referring to the amount of water required by the animal from birth to slaughter consumption during the whole growth process, and its size is related to the type of animal, feeding system, growth conditions, etc. Because the WF data per unit mass of animal products remain largely unchanged, this study adopts the latest research results of Mekonnen and Hoekstra on virtual water per unit mass of animal products in China [9], including the WF of the main farm animal products raised in Hotan area (beef, sheep, pigs, chicken, milk, and eggs).

2.4. Regional Water Use Evaluation Index System Based on Water Footprint

In this paper, we constructed an evaluation system including regional WF structure, efficiency, ecological safety, and sustainability energy indicators based on the principle, composition, and relationship between WF and water resources utilization to evaluate the regional water resources development and utilization. Since the ecological security of water resources in arid zones is closely related to drought disasters, we further analyzed the relationship between water scarcity and meteorological drought using a coupled coordination model in the ecological security index.

Table 1. Regional water resources evaluation index system based on WF.

Water Footprint Indicators		Computing Method	Description
Structure indicators	Water import dependency (WD) Water self-sufficiency (WSS)	$EWF/AWF\times 100\%$ $IWF/AWF\times 100\%$	The extent of dependence on external water resources. The extent of regional use of local water resources.
Benefit indicators	Million tons of water footprint population density (MTWF) Water footprint economic benefit value (WFEB) Net trade volume of water footprint (NTWF) Water contribution ratio (WCR)	$TP/AWF$ $\mathit{GDP}/\mathit{AWF}$ $\mathit{VWE}-\mathit{EWF}$ $\left(\mathit{VWE}-\mathit{EWF}\right)/\mathit{WA}$	WF supports population capacity. WF is consumed by 1 RMB GDP. Status and role in water trade. Regional water availability contributes to other regions’ contribution levels.
Ecological safety indicators	Water scarcity (WS) Water pressure (WP)	$AWF/WA\times 100\%$ $\left(IWF+VWE\right)/WA$	Water scarcity status. The intensity of the region’s productive water demand on available water resources.
Water sustainability performance indicators	Water footprint rate of change (WFPR) Water resource availability (WAR) Water sustainability index (WSI)	$\left(AW{F}_2-AW{F}_1\right)/AW{F}_1\times 100\%$ $\left(W{A}_2-W{A}_1\right)/W{A}_1\times 100\%$ $\left|WFPR\right|/\left|WAR\right|$	Variation of regional water consumption. The magnitude of change in regional water availability. Status and capacity of sustainable use of regional water resources.

TP: The total regional population, WA: Total regional available water resources.

showed the detailed indicators under each index separately, and their calculation formulae and meanings [30], including: (1) Structural indicators mainly indicate which of the local water resources is dominated by imports or self-sufficiency. These include water import dependency (WD) and water self-sufficiency (WSS), both of which reflect the composition of water resources consumed by a region’s goods and services and indicate the strategic structure of the regional water resources. If a region is water-poor and its goods and services are heavily dependent on virtual water imports, WD is close to 100% and WSS is close to 0. The reverse is also true. (2) The regional water footprint benefit indicators include internal benefit indicators and external benefit indicators. Internal benefit indicators are used to measure the effects and benefits of water resources in the region, reflecting the role played by water resources in the region; external benefit indicators are used to measure the external effects and role of water resources, reflecting the region’s position in the virtual water trade and its impact on other regions. (3) Ecological security indicators are used to indicate whether local water resources can meet the consumption of local residents and the level of water stress. Water scarcity (WS) reflects the state of water scarcity. The larger the indicator, the more severe the water scarcity situation facing the region. Water pressure (WP) reflects the intensity of the water demand for goods and services produced in the region on the amount of water resources available. The larger the WP, the higher the water resources carrying pressure in the region. When WP > 1, the water resources system in the region is overloaded; when WP = 1, the water resources utilization reaches the maximum carrying capacity; when WP < 1, the water footprint for production in the region is within the available water resources carrying capacity; (4) Water sustainability indicators are used to determine the sustainable development of regional water resources, in which Water footprint rate of change (WFPR) reflects the variation of regional water resources consumption in a certain period, and its size indicates the speed of increase or decrease in regional water footprint. Water resource availability (WAR) reflects the variation of regional water resources availability over a certain period of time, and its magnitude indicates the speed of increase or decrease in regional available water resources. Water sustainability index (WSI) quantifies the intensity of sustainable water resources utilization. Using the three indicators of water resources sustainability, we can determine the sustainable status and capacity of water resources in the region, and the judicial process is shown in Figure 2.

The coupled coordination degree model is used to understand drought and water ecological security as two interacting systems, as drought occurs, and more water is needed for productive activities in water-scarce areas. Combining WS with SPEI in the analysis, the coupled coordination model was used to analyze the impact and relationship of drought on local water security. Need to calculate the coupling C value, the larger the value indicates the greater the interaction between the systems. Need to calculate the coupling degree C value, the larger the value indicates the greater the interaction between the systems, coupling coordination degree D value between 0–1, the larger the value indicates the higher the degree of coordination between the systems. Referring to other scholars to grade the coupling coordination degree, the criteria are shown in

Table 2. Coupling coordination level classification criteria.

Class	D	Coordination Degree	Coordination Grade
Uncoordinated Declination I	0.00–0.20	High Incoordination	I1
	0.21–0.40	Moderate Incoordination	I2
	0.41–0.50	Low Incoordination	I3
Transitional Development II	0.51–0.60	Near Incoordination	II4
	0.61–0.70	Low Coordination	II5
Coordinated Development III	0.71–0.80	Intermediate Coordination	III6
	0.81–0.90	High Coordination	III7
	0.90–1.00	Complete Coordination	III8

[50]. With the continuous interaction and coupling coordination between the two and the adoption of advanced water conservation technologies and other means to mitigate the impact of drought, the water security environment will gradually improve, and agricultural production will gradually break through the drought environmental constraints to further enhance and develop.

2.5. Standardized Precipitation Evapotranspiration Index (SPEI)

The SPEI is reasonable as a standard for drought assessment studies and the results are referable with good regional applicability, and the value of SPEI-12 is used in this study [51,52,53]. The potential evapotranspiration (PET) is calculated using the Thornthwaite method, taking into account the effect of temperature:

$$PET=16.0\times {\left(\frac{10T_i}{H}\right)}^a$$

(9)

where T_i is the monthly average temperature, H is the annual heat sum, and a is a constant.

Calculate the difference between precipitation and PET.

$$D_i=P{R}_i-PE{T}_i$$

(10)

PR_i is the monthly precipitation and PET_i is the monthly potential evapotranspiration.

To normalize the data series, the log-logistic probability distribution is used to fit the D_i series with negative values and find the cumulative function:

$$F\left(x\right)={\left[1+{\left(\frac{\alpha }{x-\gamma }\right)}^{\beta }\right]}^{-1}$$

(11)

where α, β, and γ are the scale parameter, shape parameter, and origin parameter, respectively, and are obtained by fitting with the linear moment method.

The SPEI is found by using the normal inverse transform as follows:

$$W=\sqrt{-2ln\left(P\right)}$$

(12)

$$SPEI=W-\frac{C_0-C_{1}W+C_2{W}^2}{1+d_{1}W+d_2{W}^2+d_3{W}^3}$$

(13)

where C₀ = 2.515 517, C₁ = 0.802 853, C₂ = 0.010 328, d₁ = 1.432 788, d₂ = 0.189 269, and d₃ = 0.001 308. when P > 0.5, P = 1 − P.

The current GB/T20481-2017 “Meteorological Drought Levels” was used to classify the drought levels. See

Table 3. Meteorological Drought Levels.

Grade	SPEI Value	Type
1	−0.5 < SPEI	Normal
2	−1.0 < SPEI ≤ −0.5	Slightly drought
3	−1.5 < SPEI ≤ −1.0	Moderate drought
4	−2.0 < SPEI ≤ −1.5	Severely drought
5	SPEI ≤ −2.0	Extremely drought

.

3. Results

3.1. Agricultural Water Footprint

3.1.1. Production Water Footprint

As can be seen from Figure 3, PWF fluctuated and increased over 20 years reaching 66% total growth from 17.54 × 10⁸ m³ in 2000 to 29.16 × 10⁸ m³ in 2019 reaching a maximum of 34.72 × 10⁸ m³ in 2015 which may be related to the increase in the total amount of agricultural products in that year. In total, the water footprint of grain production in Hotan reached 215.09 × 10⁸ m³ in 2000–2019, followed by meat with PWF of 106.92 × 10⁸ m³, followed by fruits with PWF of 85.42 × 10⁸ m³. Hotan produced a total of 486.55 × 10⁸ m³ of water footprint in 2000–2019, which is about 24.33× 10⁸ m³ per year. As seen in Figure 4, PWF showed a decreasing trend in 2017–2019 with the most obvious decrease in cotton which produced less in that year. The proportion of $W{F}_{cro}$ was larger than that of $W{F}_{ani}$, indicating that agriculture in Hotan region was still dominated by the cultivation and the scale of livestock was smaller. It can be seen in Figure 3 and Figure 4, the share of $W{F}_{cro}$ activities in which grain production is the main water consuming activity amounts to 44.21% of the water footprint of all products produced. After grain, fruits and cotton had an average PWF of 4.27 × 10⁸ m³ and 2.58 × 10⁸ m³ with proportions of 17.56% and 10.59%, respectively. While the remaining species account for a smaller proportion which is related to the differences in CWR between different crops. The overall PWF of grain and cotton showed a decreasing trend while the proportion of fruits was increasing year by year and meat showed a trend of increasing first and then decreasing. On the one hand, the PWF of Hotan area was increasing annually, on the other hand, it was also directly related to the change in planting area and production of various crops. From 2000 to 2019 $W{F}_{ani}$ showed an uneven growth trend with time, and in 2019 $W{F}_{ani}$ in Hotan area was 7.57 × 10⁸ m³, an increase of 171.36% compared with that in 2000. Changes in animal products production have a direct impact on $W{F}_{ani}$ change, with a relatively stable feed structure and unchanged service water and drinking water standards, the production of livestock of all sizes in Hotan area increased from 60,800 tons in 2000 to a peak of 251,819 tons in 2016 then declined to 146,424 tons in 2019. $W{F}_{ani}$ in meat consumption mainly accounted for 21.97% of the proportion second only to grain crops.

Figure 5a showed that the proportion of WF_blue in agricultural water in Hotan was about 65.74% and WF_green was only 34.26%. A total of 320.78 × 10⁸ m³ of WF_blue and 167.18 × 10⁸ m³ of WF_green were produced in Hotan from 2000–2019, reaching an average of 16.04 × 10⁸ m³ and 8.36 × 10⁸ m³ per year. WF_blue reached a maximum of 21.29 × 10⁸ m³ in 2015 and increased by 11.08% in 2019 compared with 2000. WF_green at least was only 3.75 × 10⁸ m³ in 2000 and it reached its highest in 2016 with 13.81 × 10⁸ m³. Due to the climate conditions of extreme drought and scarce rainfall in Hotan, agricultural water is mainly groundwater and alpine ice melt and snowmelt runoff, making the proportion of WF_blue in agricultural water generally at a high level. WF_green has a lower opportunity cost compared to WF_blue for irrigation water due to factors such as precipitation. On the whole WF-blue and WF-green were increasing year by year, reaching a peak in 2015 and 2016. This is due to the expansion of agricultural cultivation area and livestock farming in Hotan area, the increase in agricultural products production and the increase in agricultural water input resulting in the increase in blue and green water footprint. Figure 5b showed that most of the products PWF in Hotan could meet the consumption of the local population, which indicated that Hotan mainly exported agricultural and livestock products virtual water in WF trade. From 2000 to 2019 still need to import fewer types of products but in increasing quantities, which indicates that the mismatch degree and gap between local WF production and consumption is gradually increasing, and attention should be paid to timely adjustment in the subsequent development planting system and livestock production species, and rational use of water resources. For example, increase the planting area of sunflower, rape and other oil crops, expand the scale of sheep and poultry breeding, and increase milk production to meet the consumption of local residents.

3.1.2. Comparative Analysis of AWF and WA in Hotan Region

After calculating various types of water consumption in Hotan from 2000 to 2019, the AWF results were obtained after integrated consideration of the virtual water volume of import and export, as shown in

Table 4. Statistical table of various types of water footprints (10⁸ m³) in Hotan from 2000–2019.

Year	PWF	VWE	IWF	EWF	AWF	Per Capita AWF (m3)	WA	Per Capita WA (m3)
2000	17.54	9.81	7.73	1.13	8.86	526.89	-	-
2001	18.69	10.70	7.99	0.73	8.72	509.21	-	-
2002	17.65	9.02	8.63	1.04	9.67	554.60	-	-
2003	17.29	9.68	7.61	0.92	8.54	482.05	-	-
2004	18.96	11.17	7.79	0.63	8.41	467.07	39.04	2166.86
2005	18.97	11.05	7.92	0.59	8.51	459.10	37.60	2027.84
2006	19.17	11.17	8.00	0.62	8.62	456.28	38.70	2047.84
2007	21.27	15.13	8.149	0.72	8.86	462.34	39.13	2041.32
2008	23.01	14.78	8.24	0.74	8.98	462.04	38.90	2001.84
2009	24.67	15.70	8.97	0.65	9.62	482.95	42.87	2151.33
2010	23.41	14.43	8.97	0.44	9.41	452.90	44.83	2157.25
2011	26.77	16.52	10.26	0.58	10.83	511.82	47.01	2220.98
2012	26.68	15.20	11.48	0.53	12.01	554.59	44.09	2036.37
2013	29.18	19.59	9.59	1.02	10.62	483.08	47.67	2168.93
2014	31.28	20.74	10.54	1.18	11.72	508.63	46.73	2028.24
2015	34.72	22.76	11.96	0.99	12.95	545.51	47.05	1982.04
2016	34.31	21.97	12.33	1.20	13.54	541.24	47.64	1904.73
2017	29.86	18.15	11.71	1.82	13.53	524.41	45.60	1767.39
2018	28.58	16.04	12.54	1.89	14.43	556.12	42.83	1650.39
2019	29.16	17.34	11.82	1.75	13.57	521.22	42.02	1613.64

. The AWF in Hotan showed a fluctuating upward trend of 53.18% during 2000–2019, AWF has a total of 211.39 × 10⁸ m in 20 years. The changing trend of IWF (Import WF) and EWF (Export WF) was consistent with that of AWF. In total, the IWF of the Hotan region from 2000–2019 was 192.23 × 10⁸ m³ and the EWF was 19.16 × 10⁸ m³. The trend of AWF was overall negatively correlated with the change in the water footprint output, which generally remained around 13–14 × 10⁸ m³ after 2016 and reached a maximum of 14.43 × 10⁸ m³ in 2018, and the per capita WF also reached a maximum of 556.12 m³/person in that year, mainly due to the consumption of grain, cotton, vegetables, and fruits by residents in that year very large. In addition, the minimum per capita AWF was 452.90 m³/person in 2010, which was related to the growth of the population in the Hotan region. The resident population of Hotan is 2.60 million in 2019, an increase of 922,213 people in 20 years. The GDP increased from 2.71 billion RMB in 2000 to 40.69 billion RMB in 2019, and economic growth and development brought about the growth of WF.

The WA in Hotan area exhibited a trend of increasing and then decreasing and the starting time was 2004 because the local WA volume statistics started in 2004. The closest 2004 data is used for the relevant WA for 2000–2003 in subsequent calculations. The total number of WA in Hotan during the 16 years 2004–2019 was 691.71 × 10⁸ m³, with an average of 43.23 × 10⁸ m³ per year. After 2009, the WA remained above 40 × 10⁸ m³ and the per capita WA was much larger than the per capita AWF, which suggested that the water resources in Hotan area were not reasonably developed and utilized and there was room for further development and utilization. The per capita WA in Hotan was gradually declining which was related to the rapid population growth in Hotan in the past 20 years. From the perspective of WF, the closer the AWF is to WA, the better the water utilization is. The total water supply in the Hotan area increased yearly while the total water resources decreased slowly. Since the area is in an arid zone, the adoption of a traditional cropping system in irrigation areas, the serious waste of agricultural water resources, as well as the continuous increase in population, a comprehensive understanding of exploring how to effectively use water resources from the perspective of WF is an effective way to solve the water shortage in Hotan arid zone.

3.1.3. Comparative Analysis of AWF in Normal and Dry Years

Combined with Figure 6 and Table 2 meteorological drought level judgment table, there were 8 years (2002, 2003, 2005, 2010, 2012, 2016, 2017, 2019) without drought in Hotan area in 2000–2019, and the rest of 12 years had different degrees of meteorological drought occurring, the proportion of which reaches 60%. Among them, there were 8 years (2006, 2007, 2008, 2009, 2011, 2013, 2015, 2018) with more than severe drought degree, and the proportion of drought years reached 66.67%, among which 2011 and 2018 reached extreme drought degree, and there was only one year of light drought in 2014, except for the rest of years of moderate drought. By analyzing the SPEI, we found that the Hotan area is prone to drought hazards, and the degree of drought is more severe once it occurs.

The follow-up divided the Hotan region into three scenarios for 2000–2019 by SPEI level: Normal, Slightly and Moderate drought, Severely and Extremely drought. Further analysis of the change in WF is shown in Figure 7. In normal years, the WF_green in Hotan was 8.85 × 10⁸ m³ and the WF_blue was 15.42 × 10⁸ m³, while the relative light and medium drought years were 6.3 × 10⁸ m³ and 15.32 × 10⁸ m³, the WF_green decreased more and the WF_blue increased but not much. In heavy drought years compared with normal years, WF_green was close to but WF_blue increased significantly, which is because Hotan area is mainly irrigated agriculture in drought years then mainly relies on blue water irrigation to recharge. However, WA was higher at 43.63 × 10⁸ m³ in normal years but decreased in drought years. In general, the little change in WA suggested the limited amount of water resources in Hotan area, which increased the local water stress even more in drought years when water was urgently needed for irrigation. PWF was least on average in light to medium drought years, with little difference between perennial and severe drought years, but VWE were much higher in severe drought years than in the other two cases, indicating that local water scarcity crisis can be alleviated by local import and export water trade when drought disasters occur. The small difference in AWF between the three scenarios suggested that the water resources consumed by the population in Hotan were relatively stable.

The pie chart in Figure 7 revealed that the AWF in Hotan was dominated by grain. In light of medium drought years, the proportion of grain was 52.55%, and the proportion of cotton in that year was also larger than the other two, as cotton requires more water which may aggravate the local water stress. The proportion of cotton and meat and vegetable dairy in heavy drought years was also greater than in normal years but the overall proportion did not change much. The pie chart revealed the largest change in the proportion of agricultural products in light to medium drought years, and the proportion of vegetables, fruits, and meat, in this case, declined most significantly, which also reminded us to analyze the changes in the local planting structure while studying the changes in agricultural water resources in drought areas, and to reasonably adjust the agricultural production system to achieve the best production methods obtained in drought years.

3.2. Water Resources Evaluation

3.2.1. Analysis of Water Footprint Structure Indicators and Benefit Indicators

Figure 8 was obtained after calculation according to Table 1. Figure 8 presents the structure and benefits indicators of the WF evaluation. Water import dependency (WD) in Hotan showed a trend of decreasing and then increasing from 2000 to 2019, reaching a maximum of 13.44% in 2017 with an average of 8.87%. Compared with the global 13%, the WD of Hotan region was small, but compared with the average of China, the WD of Hotan was larger. This indicates that Hotan region is located in the extremely arid zone of China with a fragile ecological environment, insufficient internal water resources, backward economic development, and certain dependence on external agricultural water resources. The Water self-sufficiency (WSS) remained above 86%, reaching a maximum of 95.61% with an annual average of 91.13%, while the world and China’s average WSS were 73% and 80.5%, respectively [54]. In contrast, the self-sufficiency of AWF in Hotan was high, which indicates that the agricultural water resources in Hotan can still meet the agricultural production in Hotan, the agricultural production can meet the needs of local people and there is still room for exporting products, and the overall water footprint structure is reasonable, but the utilization rate of water resources is low.

The Million tons of water footprint population density (MTWF) changed between 17.98–22.08 persons/10⁴ t in 20 years with a small change, and the water resources carrying capacity of Hotan area was more stable with the change of AWF. The years 2011–2012 and 2013–2016 were declining phases, with the largest decline in 2010–2012, which was due to a large increase in AWF during these two years, specifically related to the increase in agricultural production, industrial restructuring, etc. The overall Water footprint economic benefit value (WFEB) was low but increased year by year, from 3.06 RMB/m³ in 2000 to 29.98 RMB/m³ in 2019 an increase of 878.95%. Water resources have not yet reached efficient utilization in agricultural production in Hotan area, and the economic benefits generated are still low, but with the development of the local economy, planting technology, and agricultural income, the economic benefits of WF will continue to improve in the future.

Net trade volume of water footprint (NTWF) showed a fluctuating upward trend from 2000 to 2015 which reached a maximum of 21.77 × 10⁸ m³ in 2015 with a difference of 13.79 × 10⁸ m³ from the lowest value in 2002 because the AWF was larger and the trade volume was larger in 2015, and after 2018 has been a decreasing trend cause that the AWF decreased after 2015 and also reduced the trade exports of agricultural products and the regional trade volume of agricultural products was closely related to AWF. However, the NTWF was always positive, which indicates that Hotan is an agricultural water exporting area, and the exported water footprint is larger than the imported water footprint, and the locally produced water footprint is used to satisfy the consumption of local residents while exporting virtual water to other regions by means of product trade. The Water contribution ratio (WCR) was less than 1 for 20 years with a slight but insignificant increase. The low WCR reflects the low contribution of water resources in the Hotan region indicating that the amount of virtual water trade is small compared to the available water resources, and the generated virtual water for trade and available water resources do not relieve the pressure on water resources in other arid regions.

3.2.2. Analysis of Water Resources Ecological Safety Indicators

After calculation according to Table 1, we get the overall trend of Water scarcity (WS) in Hotan from 2000–2019 was increasing, but most of the time it was below 30% reaching a maximum WS of 33.70% in 2018. WS and WA were inversely proportional, which suggested that the decline in WS required some increase in total water use. The Water pressure (WP) of Hotan area for 20 years was less than 1, which indicated that the agricultural water use in the area was still within the available water resources carrying range at present. However, the WP was showing a trend of increasing year by year and reached more than 0.7 in 2015 and 2016, and the PWF was also higher in these two years, which showed that the increase in agricultural production and area in Hotan raised the PWF, but at the same time, it would also bring more pressure on the local water resources. The natural environment of Hotan region determines that precipitation is low and unevenly distributed. Under the premise that the total amount of water resources is not controllable, if we do not pay attention to the rational use of water resources, it will lead to the overload exploitation of water resources in Hotan region and increase the conflict between supply and demand of water resources. The current water consumption in Hotan is within the carrying capacity of available water resources, but if no moderation measures are taken a state of overload may occur in the future.

After the calculation of the coupled coordination model, the coupling degree (C), coordination degree (D), and coupling coordination level of SPEI-WS were obtained, which were presented in

Table 5. Ecological security indicators of the WF in Hotan 2000–2019 and coordination degrees.

Year	WS	WP	C	D	Grade
2000	22.70%	0.45	0.87	0.50	I3
2001	22.33%	0.48	0.83	0.46	I3
2002	24.77%	0.45	0.94	0.66	II5
2003	21.87%	0.44	0.52	0.53	II4
2004	21.56%	0.49	0.61	0.40	I2
2005	22.64%	0.50	0.67	0.60	II4
2006	22.28%	0.50	0.88	0.43	I3
2007	22.65%	0.59	0.96	0.43	I3
2008	23.08%	0.59	0.99	0.44	I3
2009	22.45%	0.58	0.94	0.42	I3
2010	20.99%	0.52	0.20	0.31	I2
2011	23.04%	0.57	1.00	0.42	I3
2012	27.23%	0.61	0.98	0.77	III6
2013	22.27%	0.61	0.91	0.41	I3
2014	25.08%	0.67	0.98	0.63	II5
2015	27.52%	0.74	0.93	0.59	II4
2016	28.42%	0.72	1.00	0.78	III6
2017	29.67%	0.65	1.00	0.83	III7
2018	33.70%	0.67	0.20	0.32	I2
2019	32.30%	0.69	0.99	0.86	III7

. The reason for the smaller C of the two systems in 2003, 2010, and 2018 could be that the SPEI in that year showed a larger difference and presented more extreme meteorological phenomena that negatively affected the development of the system coupling. In the rest of the years, SPEI-WS showed a high level of coupling, especially in 2011, 2016, and2017 when the C-value approximated 1.00. This proves that the interaction between meteorological drought and water scarcity is more obvious and the degree of intrinsic correlation between the two systems is better. The overall SPEI-WS system in Hotan was at the uncoordinated level (I–II) until 2011 during the study period. Because the Hotan area is in an extremely arid zone water conservation and utilization technology is backward water resource utilization efficiency is low when drought occurs and can’t effectively mitigate the harm caused by water shortage. And then after it reached grade III6 in 2012, it entered the uncoordinated level again can be understood as a transitional stage. The Hotan region was in a situation of figuring out how to deal with the relationship between drought-water shortage and improve water use efficiency. It was after 2015 that the system entered grade III of which only 2018 was uncoordinated due to an exceptional drought phenomenon. The significant increase in the coupled coordination level suggested that water scarcity was improving in the drought state. At this stage, with the construction of various water conservation projects and the promotion of water-saving technologies, and the improvement of farmland irrigation facilities, the effective irrigation area increased, and the unit water consumption decreased to alleviate the local water shortage due to meteorological drought. Although the SPEI-WS system achieved a coordinated development but is still in a fragile development stage, the emergence of extreme weather will bring a greater impact on the production environment and so on in the Hotan area.

3.2.3. Analysis of Water Sustainability Performance Indicators

As shown in

Table 6. Water sustainability performance indicators for 2000–2019 in Hotan.

Year	WFPR	WAR	WSI
2000	-	-	-
2001	−1.60	-	-
2002	10.91	-	-
2003	−11.73	-	-
2004	−1.43	-	-
2005	1.16	−3.69	0.31
2006	1.29	2.93	0.44
2007	2.78	1.11	2.50
2008	1.31	−0.59	2.22
2009	7.19	10.21	0.70
2010	−2.20	4.57	0.48
2011	15.11	4.86	3.11
2012	10.84	−6.21	1.74
2013	−11.58	8.12	1.43
2014	10.37	−1.97	5.26
2015	10.50	0.68	15.34
2016	4.54	1.25	3.62
2017	−0.05	−4.28	0.01
2018	6.67	−6.08	1.10
2019	−5.96	−1.89	3.15

, the Water footprint rate of change (WFPR) of Hotan region showed a negative growth trend in 2001, 2003, 2004, 2010, 2013, 2017, and 2019, and the rest of the years showed positive growth. The largest increase was in 2011 indicating that the AWF in the Hotan region grew at a faster rate in 2010–2011 Water use efficiency was low. The slowest growth rate in 2017 was due to the small difference in production and PWF in 2017 compared to 2016, and the negative growth may be explained by the fact that there was still a decrease in production. Water resource availability (WAR) had the largest variation of 10.21 in 2009, with a WA of 42.87 × 10⁸ m³ in that year compared to only 38.90 × 10⁸ m³ in 2008. Agricultural water resources efficiency variation from WAR was stable but agricultural water resources conservation was not enough, which was related to the insufficient surface water storage capacity and limited water supply capacity in Hotan. In addition, drought disasters were frequent in Hotan area, water resources pollution was caused by the unreasonable use of water resources, and the situation of water shortage was greatly aggravated.

From

Table 7. Sustainability Evaluation of Water Resources Development in Hotan.

Year	WFPR	WAR	WSI	Sustainability Evaluation
2000	-	-	-	-
2001	<0	-	-	-
2002	>0	-	-	-
2003	<0	-	-	-
2004	<0	-	-	-
2005	>0	<0	<1	Unsustainable
2006	>0	>0	<1	Sustainable
2007	>0	>0	>1	Unsustainable
2008	>0	<0	>1	Unsustainable
2009	>0	>0	<1	Sustainable
2010	<0	>0	<1	Sustainable
2011	>0	>0	>1	Unsustainable
2012	>0	<0	>1	Unsustainable
2013	<0	>0	>1	Sustainable
2014	>0	<0	>1	Unsustainable
2015	>0	>0	>1	Unsustainable
2016	>0	>0	>1	Unsustainable
2017	<0	<0	<1	Unsustainable
2018	>0	<0	>1	Unsustainable
2019	<0	<0	>1	Sustainable

combined with Figure 1 water resources sustainable use judgment figure, we can conclude the sustainability of water resources use in Hotan area. From the WF perspective, the agricultural water resources in Hotan were only used sustainably in 2006, 2009, 2010, 2013, and 2019, and unsustainably in all other years. Especially from 2011, the percentage of unsustainable years was relatively large, and improved by 2019. This also demonstrated that the sustainable development and utilization of agricultural water resources in Hotan were not optimistic, and the water security problem was serious. The total amount of water resources and available water resources in Hotan area are greatly influenced by the natural environment, such as rainfall, temperature, and other uncontrollable factors, and Hotan area is in an extreme arid zone, so the total amount of water resources is unstable and water shortage limits the development of local irrigated agriculture. With the increase in agricultural activities, the local population and economic growth also pose greater challenges to water security and local ecological protection.

A comparison of the PWF, AWF and WA for the sustainable and unsustainable years of water resources yielded an average WA above 43.21 × 10⁸ m³ for the sustainable years and around 42.11 × 10⁸ m³ for the unsustainable years, with PWF and AWF of 25.12 × 10⁸ m³ and 10.37 × 10⁸ m³ for the former and 24.37 × 10⁸ m³ and 10.64 × 10⁸ m³ for the latter, respectively. The proportion of PWF to available water resources in sustainable years reaches 58%, while the proportion of AWF is 24% and the proportion of unsustainable years is 58% and 25%, respectively. Therefore, we can tentatively conclude that the available water resources to be provided in Hotan should be maintained above 43.21 × 10⁸ m³ under the condition that the variation of produced water footprint is less floating, which can tentatively ensure the sustainable use of water resources in the dry and water-scarce Hotan area. In the case of future production water footprint increase under the existing production conditions, the available water resources should be increased, otherwise water utilization needs to be improved to reduce water waste to ensure the sustainable use of water resources.

4. Discussion

This study evaluated water use and sustainability in arid areas based on the WF. It was obtained that PWF, and AWF were increasing in Hotan area during the study period. The share of grain crops in the water footprint of all products produced is 44.21% but is gradually decreasing, the proportion of cash crops PWF increased, the proportion of livestock products increased first and then decreased, and the proportion of WF_blue was larger than that of WF_green. Some scholars have concluded that the increase in PWF in southern Xinjiang is greater, in which the proportion of food crops gradually declined, the proportion of cash crops gradually increased, and that the increase in cash crops is greater than that of food crops [1]. Regarding the blue-green water consumption of WF in arid zones, WF_green contributed less to the WF of crops compared with WF_blue [55,56]. However, more agricultural water footprint studies did not involve livestock only crops were studied, which is more single, and this study combines crops and livestock more widely involved.

Regarding the water resources evaluation, this study revealed more significant benefits and effects within the Hotan area and less sustainable local water resources in the future. The per capita water availability index is one of the most widely used blue water scarcity indicators [57]. Some scholars proposed the water footprint-based principal component analysis (WFPCA) method to evaluate the water resources carrying capacity [58]. This study used WF indicators including physical water and virtual water to join the study of water resources sustainability, which is no longer just a single physical water usage and can be combined with more scientific evaluation body construction methods for water resources sustainability evaluation in arid areas in subsequent studies. In addition to studying the historical WF, it is also possible to understand the changes in the WF of production and consumption that may occur in the future in each region to show the dependence of each region on the water resources of other regions under different possible futures, and it is possible to further assess how humans will mitigate freshwater shortages in the future [59].

Some studies revealed that AWF depends mainly on agricultural management rather than regional climate [60,61]. To further explore the relationship between meteorological drought and WF, we conducted correlation tests on 13 factors, including SPEI, Pe, AWF, and GDP, and the results are shown in Figure 9. Only Pe, RH, and AWg(water requirement per unit of green water) correlated strongly with SPEI. The correlation between the remaining factors and SPEI did not perform well, probably due to the complexity of the calculations and the more complex effects generated by the factors. The AWF, IWF, EWF, and PWF in Hotan showed significant positive correlations among the factors except for a small correlation between EWF and PWF. The correlation between AWF and meteorological drought did not perform well indicating that the occurrence of the meteorological drought was more related to natural elements while AWF was more driven by human activities such as agricultural production input elements.

This also reflects the comprehensiveness of the water footprint calculation, which is no longer the visible natural water such as surface water and groundwater that were used to simply calculate water consumption but includes the invisible water generated by social and economic activities, which further reflects the innovation of the water footprint theory.

The water footprint provides a new approach to the evaluation of water use in agricultural production processes. However, this study also has some limitations. On the one hand, the data used to calculate the water footprint, such as crop yields, crop planting data, and agricultural input data, are mainly derived from statistical data. Therefore, there may be discrepancies between data and reality. On the other hand, in the calculation process, the natural environment and agricultural production activities are different in different regions, and we have not analyzed the impact of changing the cropping pattern on the water footprint in more detail, such as the development of facility agriculture in Hotan region in recent years, and we should further develop a water footprint assessment method suitable for different regions according to the actual situation.

5. Conclusions

In this paper, we measured the WF of different agricultural product types in Hotan region for 20 years from 2000–2019 using meteorological, hydrological, and economic data to construct a water use evaluation system based on WF to briefly explore the relationship between meteorological drought and WF. The results of the study are as follows:

PWF, WF_green, WF_blue, and AWF showed an overall fluctuating increasing trend over 20 years in Hotan area. $W{F}_{cro}$ was larger than $W{F}_{ani}$, the proportion of $W{F}_{cro}$ dominated by grain crops among all products was 44.21%; the proportion of $W{F}_{ani}$ dominated by meat was 21.97%. The proportion of grain crops decreased gradually, the proportion of cash crops increased progressively, and the proportion of livestock products increased first and then decreased. The blue-green water footprint with WF_blue as the main consumption accounted for 65.74%. The AWF reached its highest in 2018 with 14.43 × 10⁸ m³ and the highest per capita WF in that year with 556.12 m³/person. The Hotan region needs to further adjust its agricultural production structure, such as increasing the acreage of sunflower, rape and other oil crops to reduce oil crop imports, expanding the scale of sheep and poultry breeding, and increasing milk production to meet the growing consumption of meat, eggs and milk nutrition products by local residents. The area of grain, vegetable and fruit crops can be reduced on the basis of improving crop yield through scientific and technological means. The AWF in Hotan is lower than WA, indicating that the water resources in the area have not been reasonably exploited and there is more room for further exploitation in the future. Water conservation can be achieved by adopting facility agriculture and improving irrigation methods such as using drip irrigation and sprinkler irrigation instead of traditional large water irrigation.

The evaluation results of water resources in Hotan region revealed that the annual average value of WD was 8.87%, which was lower than the global level but higher than China and had a certain dependence on external agricultural water resources. The annual average value of WSS reached 91.13%, indicating that the agricultural water resources in the area can currently meet the local agricultural production activities. With the increase in populations and GDP, the overall economic efficiency index of water footprint in Hotan region was rising, in which WFEB increased by 878.95%. NTWF was always positive, indicating that the produced water footprint in Hotan region was exported outward in the form of agricultural trade.

However, the WCR was lower and the level of contribution to water resources consumption in other areas was smaller. The WP < 1 for 20 years in Hotan region showed that the agricultural water consumption was still within the bearable range of local water resources, but the overall trend of WP was increasing, which indicated that with the increase in PWF, water security pressure increased, and overload development might occur. The years of unsustainable development of water resources in Hotan region reached 14 years, and the last 10 years came to 2019 to present a sustainable situation. It demonstrated that the sustainable development of agricultural water resources in Hotan region is not optimistic, and the problem of water security is serious. The water resources evaluation yielded that in order to achieve sustainable water resources use in Hotan, it is best to ensure that the available water resources for agriculture are above 43.21 × 10⁸ m³ and that the proportion of produced water footprint to available water resources does not exceed 58% under existing production conditions.

By analyzing the SPEI, it was found that there were 12 years of drought in Hotan area, among which there were 8 years of severe drought, meaning that the ecological environment of Hotan area was fragile and prone to drought and the degree of the disaster was more serious. The coupled coordination model suggested that the SPEI-WS system behaved in three phases: 2000–2011 (uncoordinated level), 2012–2015 (transitional phase), and 2016–2019 (coordinated level). The improved coupling coordination level signified that the water scarcity situation was changing for the better under the drought condition alleviating the local water scarcity due to meteorological drought, but still in a vulnerable development state highly susceptible to extreme climatic phenomena. Correlation studies showed that the occurrence of meteorological drought-related more to natural elements while WF was more driven by human activities.

In general, WF theory is a more comprehensive and specific study of the development and utilization of water resources in arid areas, which have outstanding water security problems and fragile carrying capacity and are highly susceptible to unsustainable development. Water scarcity and drought restrict regional agricultural development, and water resources should be planned rationally in the future to achieve sustainable development.