Response of Domestic Water in Beijing to Climate Change

Abstract

Beijing, a megacity in northern China, has been long facing the challenge of water scarcity, and the problem of domestic water scarcity has been becoming more serious in recent years due to climate change and global warming. To cope with the adverse effects of climate change, it is urgent to build a prediction model for water consumption in Beijing under the background of climate change. Here, a climate domestic water use model was established based on the historical meteorological data and domestic water use data, and the future domestic water demand in Beijing and the response of domestic water use to climate change were projected. The results showed that the climatic water consumption in Beijing will increase with climate warming by 177.23 million m³/°C, and the per capita annual water consumption will increase by 8.1 m³/°C. Combined with the CMIP6 multi-model climate change scenario data, the climate domestic water consumption in Beijing in 2035 under the four scenarios of SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5 will be 169 million m³, 189 million m³, 208 million m³, and 235 million m³ respectively; by 2050, the climate domestic water consumption in Beijing will reach 338 million m³, 382 million m³, 395 million m³, and 398 million m³, respectively. Under SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5 scenarios, if all the increased climate domestic water consumptions are supplemented by groundwater, compared with 2019, the groundwater depth will decrease by 0.18 m, 0.22 m, 0.27 m, and 0.32 m in 2035, respectively, and the area of funnel area will increase by 6.84 km², 8.48 km², 10.11 km², 12.34 km² respectively. Compared with 2035, the groundwater depth in 2050 will decrease by 0.37 m, 0.43 m, 0.41 m and 0.36 m, respectively, the area of funnel area will increase by 14.13 km², 16.21 km², 15.61 km², and 13.68 km², respectively. If the increased climatic water consumption in Beijing is supplemented by external water transfer, the cost of external water transfer in 2035 will increase by CNY 391 million, CNY 485 million, CNY 578 million, and CNY 706 million, respectively, compared with that in 2019 under the four scenarios. Compared with 2035, the cost of external water transfer in 2050 will increase by CNY 808 million, CNY 927 million, CNY 893 million, and CNY 783 million, respectively.

1. Introduction

The water resource is fundamental for human survival and social development [1] and it is an important and irreplaceable natural resource. The water resource scarcity is listed as the first risk factor for social development by the world economic forum’s 2015 annual risk report, However, this risk is often underestimated or even ignored due to weak management and policy failures [2,3]. Currently, approximately 2.1 billion people today lack access to adequate clean water worldwide, and 4.5 billion people have poor access to safe drinking water [4]. China has been threatened by water shortage over a long history [5]. According to the statistical data of 1995, the per capita water resource in China was less than 2400 m³, only 27% of the world’s per capita water resource, ranking 121st in the world [6]. As the political, economic, and cultural center of China, Beijing has long been challenged by the problem of water resource shortage and low utilization efficiency of freshwater resources. The water resource shortage has become the bottleneck restricting the sustainable economic and social development of Beijing [7]. With climate warming, social and economic development, population increase, and improvement of people’s living standards, the domestic water consumption in Beijing presents a trend of continuous growth, and the contradiction between water supply and demand becomes more and more severe. Therefore, it is essential to forecast the amount of water used for domestic purposes in Beijing in the context of global climate change. At present, the studies on the impact of climate change on hydrological and water resources have been mainly focused on the impact of climate change on the average change of runoff in watersheds [8,9,10,11,12], with more attention paid to the deployment of macro water resources, and there have been fewer studies on the impact of small-scale and medium-scale regional meteorological factors on water supply and use.

Domestic water consumption includes water used for all residential purposes, including in-house water uses for drinking, bathing, washing, preparing food, flushing toilets, etc., as well as outdoor water uses for gardening, lawn watering, etc. [13]. As urban water consumption increases due to a warming climate and an improved quality of life for residents, more and more attention is being paid to research on predicting water consumption, especially the impact of climate change on domestic water consumption. Recently, many studies have been carried out for forecasting domestic water consumption in the different regions of the world. Babel et al. [14] developed a model based on the multivariate econometric approach to forecast and manage domestic water use in Kathmandu Valley, Nepal. Wang et al. [15] developed a statistical model to forecast domestic water demand in the Huaihe River Basin of China by considering climate change, population growth, urbanization, lifestyle changes, and water-saving technologies. Zhang et al. [16] established a trend model for urban domestic water use and forecasted the urban per capita domestic water use in Xi’an. Many studies have confirmed that climate change has a significant impact on regional water resources [17,18,19]. Climate change has a direct impact on regional domestic water use by affecting water supply and demand [20,21]. Warziniack et al. [22] found that the impact of climate change on water consumption overwhelmed all other factors under hot and dry futures, and thus, for reliable modeling of domestic water use, it is essential that climatic elements are fully taken into account. Some risk analysis methods of water supply have been developed, and they were proven effective to achieve improved safety and reliability [23,24].

In this study, we established a climate domestic water use trend model based on historical meteorological elements and domestic water use data, and combined CMIP6 multi-model climate change scenario data to predict the future domestic water consumption in Beijing and the response of domestic water use to climate change.

The main objective of this work is to better predict domestic water consumption in the context of climate change, to provide a basis for urban water supply planning and water management, and to provide a theoretical basis for later policy development.

2. Materials and Methods

2.1. Data Sources

The daily meteorological data of Beijing from 1990 to 2009 were obtained from the China Meteorological Administration (National Meteorological Information Center. Available online: http://data.cma.cn (accessed on 20 June 2021)), and the data mainly includes the daily temperature, relative humidity, air pressure, wind speed, sunshine hours, precipitation, etc. After screening, 18 stations were employed in this study, and the station layout is shown in Figure 1. All water resource data was extracted from the Beijing Water Resources Bulletin 1990–2019 (Beijing Water Authority. http://swj.beijing.gov.cn/ (accessed on 20 June 2021)). Future climate information, including daily maximum temperature, minimum temperature, precipitation, and solar radiation, were obtained from the CMIP6 multi-model ensemble of 20 Global Climate Models (GCMs). Generally, raw GCMs are at coarse temporal (monthly) and spatial (100–300 km grid spacing) resolutions and therefore cannot be directly used to feed water use models. Liu et al. [25] have developed a statistical downscaling method—NWAI-WG. Here we used this statistical downscaling method to downscale the monthly gridded data simulated by raw GCMs to daily climate data. This approach has been frequently used in recent climate change research [26] (SSP Database. Available online: for details, see https://tntcat.iiasa.ac.at/SspDb/dsd?Action=htmlpage&page=50 (accessed on 21 June 2021)).

2.2. Research Methodology

2.2.1. Anomaly Analysis

The anomaly is a quantity often used in the statistical analysis of time series of meteorological elements, often referred to as fluctuations, and is an assessment of deviations from the normal of the mean.

$$x_{di}=x-\overline{x},$$

(1)

where the anomaly x_di is the sample series in turn minus the sample series mean. The sample data is transformed into an anomaly series with a mean of zero, making the data more comparable.

2.2.2. Variability Analysis

Variability analysis is an important indicator of temperature and precipitation analysis as it can better characterize the degree of variability of a region’s meteorological factors over the years, reflecting the multi-year variability of temperature, precipitation, and relative humidity. Variability analysis can be divided into absolute variability analysis and relative variability analysis. The absolute variability is the average of the absolute values of the anomaly.

$$\overline{\overline{V}}=\frac{1}{\mathrm{n}}{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}\left|{\mathrm{v}}_{\mathrm{i}}\right|=\frac{1}{\mathrm{n}}{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}\left|{\mathrm{t}}_{\mathrm{i}}-\overline{\mathrm{t}}\right|,$$

(2)

where $\overline{V}$ is the anomaly, t is the actual value of the series, $\overline{t}$ is the mean of the series, and n is the number of samples in the series. The relative variability is the percentage of absolute variability from the series mean.

2.2.3. Coefficient of Variation

The coefficient of variation (C_v) is the percentage of the standard deviation of a set of data compared to its mean, and is a relative measure of the degree of dispersion of the data. As it does not carry units of measurement, it is suitable for comparison of data variation where the units of measurement are different or where the units of measurement are the same, but the number of concentrations differs significantly.

$${\mathrm{C}}_{\mathrm{v}}=\frac{\mathrm{S}}{\mathrm{M}} \times 100\%$$

(3)

where S is the standard deviation of the data series and M is the mean.

2.2.4. Grey Correlation Analysis

Grey systems theory considers the prediction of systems that contain both known and unknown or non-deterministic information as the prediction of grey processes that change within a certain orientation and are time-dependent. Although the phenomena displayed in the process are random and haphazard, after all, ordered and bounded, this collection of data therefore has a potential pattern. Grey prediction is the use of such laws to build grey models to make predictions about grey systems [27,28]. Grey forecasting begins with correlation analysis, which identifies the degree of dissimilarity between the trends of the system factors, generates the original data to determine the pattern of the system changes, generates data series with strong regularity, and establishes the corresponding differential equation model to predict the future development of things. Compared to statistical models, grey forecasting models have two main advantages: firstly, the accuracy of the model can be very high even with a small amount of data; secondly, the closer the model mechanism is to the current point in time, the higher the accuracy will be. Therefore, the predictive function of grey models is better than that of statistical models.

First, the reference series is the set X_0(k) = {X₀₁, X₀₂, X₀₃, …, X_0m in the time domain of multiple indicators}. Let Xi(k) be the set of indicators to be evaluated, Xi(k) = {X_i1, X_i2, X_i3, …, X_in}. The next step is to dimensionless size the data. The series is turned into a sequence X_0(k) = {X₀₁, X₀₂, X₀₃, …, X_0n} and the set of evaluation indicators is normalized as follows: X_ij = {X_i1, X_i2, …, X_ij, …, X_in}.

The correlation coefficient ξ (X_i) is calculated as follows:

$${\mathsf{\xi }}_{\mathrm{i}}\left(\mathrm{x}\right)=\frac{\underset{\mathrm{i}}{\min }\underset{\mathrm{j}}{\min }{|\mathrm{X}}_{0\mathrm{j}}{-\mathrm{X}}_{\mathrm{ij}}| + \mathsf{\zeta }\underset{\mathrm{i}}{\max }\underset{\mathrm{j}}{\max }{|\mathrm{X}}_{0\mathrm{j}}{-\mathrm{X}}_{\mathrm{ij}}|}{{|\mathrm{X}}_{0\mathrm{j}}{-\mathrm{X}}_{\mathrm{ij}}| + \mathsf{\zeta }\underset{\mathrm{i}}{\max }\underset{\mathrm{k}}{\max }{|\mathrm{X}}_{0\mathrm{j}}{-\mathrm{X}}_{\mathrm{ij}}|}.$$

(4)

A large correlation indicates that the comparative series has a strong influence on the reference series. It is generally considered that a weak association is 0 < r ≤ 0.35; a medium association is 0.35 < r ≤ 0.65; a strong association is 0.65 < r ≤ 0.85; and a very strong association is 0.85 < r ≤ 1.

2.2.5. Polynomial Simulation Method

The change in urban living water consumption is mainly affected by the social economy, population, quality of life, climate and other factors, and its composition equation is as follows:

y = y_a + y_b + y_c,

(5)

where y is the actual urban living water consumption; y_a is the trend amount of domestic water; y_b is the climate domestic water consumption; y_c is a random quantity.

The trend of domestic water is the part of water consumption affected by social and economic development, population growth, and the improvement of people’s living standards under the assumption that climate and other factors are normal. According to the trend of domestic water use, this paper uses the logarithmic function trend method to fit the trend quantity of domestic water use in Beijing city, and the trend model is:

y_a = 3.4483ln(t) + 4.8678 (t = 1, 2, 3 …),

(6)

where R² = 0.8385, p < 0.01

The climate water consumption refers to the part of water consumption affected by the fluctuation of climatic factors (temperature, precipitation, etc.). Random quantity refers to the part affected by random factors from other aspects. The proportion of water consumption in this part is generally small and not easy to separate, so it is usually ignored. Therefore, the actual amount of urban domestic water use is the sum of the domestic water trend and the climate domestic water use, and Equation (7) can be simplified as follows:

y = y_a + y_b.

(7)

3. Results

3.1. Changes in Climate Change and Water Consumption from 1990 to 2019 in Beijing

The annual changes in temperature and precipitation in Beijing were analyzed by using anomaly and anomaly percentages, respectively, and the results were shown in Figure 2. In terms of temperature, the annual average temperature in Beijing was 13.31 °C, and the trend rate of temperature was 0.414 °C/10a. After 2007, the positive anomaly in Beijing was significantly more than the negative anomaly. After 2015, the anomaly was always positive, indicating that the temperature showed a significant increasing trend, and the warming range was also becoming larger. In terms of precipitation, the percentage of precipitation anomaly in Beijing fluctuated greatly from 1990 to 2019, and the annual average precipitation was 538.07 mm. Precipitation had obvious fluctuation, and the change between wet and dry periods was obvious. The maximum precipitation was 708 mm in 2012 and the minimum value was only 373 mm in 1999. The maximum difference in precipitation was 335 mm. The absolute and relative variability of precipitation from 1990 to 2019 was 87.538 mm and 16.27%, respectively.

The relative humidity is also an important factor that can reflect climate change. A range-level analysis of relative humidity was shown in Figure 3. Relative humidity showed a clear downward trend from 1990 to 2019, with more negative than positive ranges after 2003, and after 2007 when the temperature was positive, the relative humidity was mostly negative, indicating that climate change presented obvious warming and drying trend.

Figure 4 showed the annual variation of urban water consumption in Beijing from 1990 to 2019. It can be seen that Beijing’s industrial water consumption presented a downward trend and the city’s industrial structure has changed considerably, due to some industries with high water consumption and low efficiency (e.g., paper making, textiles, printing, and dyeing, etc.) having dropped out of Beijing’s major industrial sectors. In addition to the industrial restructuring, Beijing also carried out the technological transformation of some units of high water-consuming industries to improve the reuse rate of industrial water and the reuse rate of production wastewater [29]. For agricultural water use, irrigation accounted for more than 90% of agricultural water use, and the reduction in the area sown for crops in Beijing in recent years has been the main reason for the reduction in agricultural water use [30]. Domestic water use has shown a significant increase, with domestic water use becoming the largest expenditure of total water use in Beijing in 2006, making Beijing a dominant city in terms of domestic water use. Exploring the main factors affecting changes in domestic water use can provide a theoretical basis for taking effective measures to alleviate the contradiction between urban water supply and demand.

3.2. Factors Influencing Domestic Water Consumption

Using grey correlation analysis, the correlation coefficients between temperature, precipitation, relative humidity, and domestic water consumption were shown in

Table 1. Grey correlation degree analysis of domestic water consumption in Beijing and different indicators in Beijing.

Indicators.	Average Temperature	Maximum Temperature	Minimum Temperature	Precipitation	Relative Humidity
Grey relation degree	0.797	0.652	0.514	0.537	0.679

. Temperature and relative humidity were strongly correlated with domestic water use, while precipitation was moderately correlated with domestic water use, and the maximum and minimum temperatures were strongly and moderately correlated with domestic water use, respectively (Table 1). The impact of temperature change on domestic water use was mainly in the form of changes in the amount of urban water used for comfort purposes such as humidification and cooling due to higher temperatures. Higher average annual temperatures increased the amount of dry and hot weather, with a corresponding increase in domestic water use; higher maximum temperatures increased the amount of water used for cooling, leading to an increase in domestic water use. Changes in humidity had an impact on human comfort and water requirements. A decrease in average humidity led to an increase in water consumption by human physiology, increasing in domestic water use [31].

3.3. Extraction of Climate Water Use

There was a strong correlation between changes in temperature and domestic water use, so the effect of temperature change on domestic water use was separated.

Using Equations (6) and (7), the climate domestic water use (y_b) was calculated with the inter-annual variation shown in Figure 5. The positive and negative values of climate domestic water use represented the change in the degree of impact of climate change on water use, with positive values indicating an increase in water use compared to the original trend due to climate change, and negative values indicating a decrease in water use compared to the original trend due to climate change. Using the coefficient of variation, the coefficient of variation for domestic climate water use was 16.76. The coefficient of variation showed that climate water use fluctuated widely, which reflected the high vulnerability of climate water use.

A regression analysis of climate domestic water use and annual mean temperature in Beijing from 2000 to 2019 revealed a significant correlation at the 0.01 level (Figure 6), and the relationship equation was presented in Equation (8).

y = 1.7723x − 23.756,

(8)

where y is climate domestic water consumption (m³), x is mean temperature (°C), R² = 0.6352, and p < 0.01.

This equation indicated that for every 1.0 °C increase in the average annual temperature, Beijing’s climate domestic water consumption will increase by 177.23 million m³, and Beijing’s per capita annual domestic water consumption will increase by 8.1 m³.

3.4. Domestic Water Use Response to Climate Change under Future Climate Scenarios

Using the climate domestic water use trend projection model combined with CMIP6 multi-model climate change scenario data, the climate domestic water use in Beijing in 2035 will be 169 million m³, 189 million m³, 208 million m³, and 235 million m³ under the four scenarios of SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5, respectively. The climate domestic water consumption in Beijing in 2050 will be 338 million m³, 382 million m³, 395 million m³, and 398 million m³, respectively.

The main sources of water for domestic use in Beijing are groundwater and external water transfers. In recent years, groundwater in Beijing has always been over-exploited, and the increase in external water transfers has effectively alleviated the problem of groundwater over-exploitation. The linear correlation between groundwater storage and burial depth was found to be significant at the 0.01 level (Figure 7), and the relationship is presented in Equation (9).

y = 4.5442x − 23.691,

(9)

where y is groundwater storage (m³), x is groundwater depth (m), R² = 0.9883, and p < 0.01.

This equation indicated that a 1 m drop in groundwater was associated with a 454.42 million m³ reduction in groundwater storage.

A significant linear correlation between the depth of groundwater burial and the area of the leakage zone at the 0.01 level was found (Figure 8). The relationship is given by Equation (10).

y = 38.125x + 147.09,

(10)

where y is funnel area (km²), x is groundwater depth (m), R² = 0.8582, p < 0.01.

The equation showed that for a 1 m drop in groundwater, the area of the funnel zone would increase by 38.125 km².

This paper assumes two water supply scenarios for climate domestic water use in response to climate change under different future scenarios. Scenario 1: Climate domestic water supply is provided by groundwater. Under the four scenarios of SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5, in 2035 the buried depth of groundwater will decrease by 0.18 m, 0.22 m, 0.27 m, and 0.32 m compared to 2019, and the area of the leakage zone will increase by 6.84 km², 8.48 km², 10.11 km², and 12.34 km², respectively. By 2050, the buried depth of groundwater will decrease by 0.37 m, 0.43 m, 0.41 m, and 0.36 m, respectively, compared to 2035, and the area of the leakage zone will increase by 14.13 km², 16.21 km², 15.61 km² and 13.68 km², respectively. Scenario 2: The climate and domestic water supply are provided by external water transfer. According to previous studies, the cost of transferring one cubic meter of water through the South-North Water Transfer Project is about RMB 4.8 [32,33] when only increasing the volume of water transferred without expanding the size of the existing water transmission trunk line. Under the scenarios of SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5, in 2035, the cost of external water transfer compared to 2019 will be CNY 391 million, CNY 485 million, CNY 578 million, and CNY 706 million, respectively; and in 2050 the cost of external water transfer compared to 2035 will be CNY 808 million, CNY 927 million, CNY 893 million, and CNY 783 million.

4. Discussion

In the past 30 years, the average temperature in Beijing has been increasing significantly with a tendency of 0.414 °C/10a. After 2015, the anomaly has always been positive, indicating a significant upward trend in temperature, and the warming rate is also becoming larger, which is consistent with the findings of Zheng et al. [34], Cao et al. [35], and Ji et al. [36]. Precipitation in Beijing has obvious fluctuations, with an absolute variation of 87.538 and a relative variation of 16.27% from 1990 to 2019, which is consistent with the findings of Wang et al. [37] and Zheng et al. [38]. In addition, the anomaly of relative humidity was analyzed, and a trend of decrease was found. There was a trend toward a warming and drying climate.

Though the increase in domestic water use in Beijing is mainly due to the rapid growth of the population and the improvement of the quality of life of the residents, the impact of climate warming on domestic water use cannot be ignored. Here, the climate domestic water consumption was calculated, and it was found that for every 1.0 °C increase in annual average temperature, Beijing’s climate domestic water consumption will increase by 177.23 million m³ and the annual domestic water consumption per capita will increase by 8.1 m³, reflecting the response of urban domestic water consumption to climate change. Dou et al. [39] pointed out that the impact of temperature change on urban domestic water use is mainly due to the changes in the amount of urban water used for increasing comfort such as humidification and cooling caused by temperature increase. Bai et al. [29] predicted that domestic water use in Beijing would be the greatest pressure on the city’s water security supply. Therefore, to better cope with the adverse effects of climate change on domestic water use in Beijing, this study built a water domestic water use prediction model applicable to the Beijing area in the context of climate change.

Zhang et al. [16] used a correlation analysis between domestic water use and average temperature to find that urban domestic water use increases as the climate warms. Here, a grey correlation analysis was used, and a grey correlation of 0.797 between temperature and domestic water was obtained. The climate domestic water use pre-diction model developed in this paper, combined with CMIP6 multi-model climate change scenario data, predicted the climate domestic water use in Beijing in 2035 and 2050 under different scenarios, which can provide a basis for urban water supply planning and water management, and provide a climatological basis for the government and different sectors when developing water conservation measures.

The depth of groundwater burial has a good linear relationship with groundwater storage volume and the area of the leakage zone. Here, it was found that a 1 m drop in groundwater will decrease groundwater storage volume by 454.42 million m³ and increase the area of the leakage zone by 38.125 km², in line with the findings of Wei et al. [40], Qin et al. [41], and Zhai et al. [42]. However, when considering groundwater and external water transfers to supplement domestic climatic water use, this paper makes two assumptions and considers only one way of supplementing domestic climatic water use. Further research is needed to find the most suitable adaptation paths to climate change for domestic water use in Beijing.

5. Conclusions

Here, the meteorological factors affecting domestic water use were analyzed, and a domestic water use prediction model was established, the impact of climate change on domestic water use in Beijing was disclosed under different climate scenarios. It was found that, during the period 1990–2019, the average temperature in Beijing increased by 0.414 °C/10a, the precipitation fluctuated, and the relative humidity decreased. The domestic water use in Beijing has been increasing and has become the largest expenditure of total water use after 2006. It was also found that the climate domestic water use presented an increasing trend with climate warming, showing that as the climate warms, Beijing will need more water, and will have to pay greater ecological and economic costs. These findings are very important for urban water supply planning and water management.

The methods and results presented here are intended to contribute to the adaptation process in human settlements located in areas of water scarcity, as well as to be a valuable and useful tool for stakeholders. The study is based on a simple conceptual framework and could serve as a perfect example to support decision-making on water use and climate change, and could also support the decision-making on water use and climate change adaptation-related issues. For example, in future research in the direction of water saving, we could establish a stepped water pricing system linked to temperature to adapt to the adverse effects of a warming climate. The methodology presented in this paper can help water resource managers in water-scarce cities in China estimate future domestic water demand in the context of environmental change. Reliable estimates of future water demand can assist in planning for the sustainable management of water resources.