Skip to main content
NCBI home page
Springer Nature - PMC COVID-19 Collection logoLink to Springer Nature - PMC COVID-19 Collection
. 2023 Jan 24;20(3):3161–3176. doi: 10.1007/s13762-023-04761-8

Estimation of household water consumption pattern in a metropolitan area taking the impact of the COVID-19 pandemic

H Sabzchi-Dehkharghani 1, A Majnooni-Heris 1,✉, A Fakherifard 1, R Yegani 2
PMCID: PMC9870780  PMID: 36712838

Abstract

A new approach for estimating the household water consumption pattern was developed by taking the impact of the COVID-19 pandemic using geographical data. Water consumption data for two years before and a year after the outbreak of the pandemic were analyzed to recognize the consumption pattern on annual and bi-monthly time scales as well as in different spatial classes. Following the recognition of the pattern, the spatiotemporal distribution of household water consumption was estimated based on the discovered connections between consumption and geographical variables. Once a regression relationship between consumption and population density was observed, an idea was developed to investigate the linear equations and their coefficient of parameters in water consumption groups from very low to very high classes using the training data. The coefficients were then adjusted to account for the pandemic's impact on the consumption pattern. Results showed that the highest increases in consumption were 11% for May–July due to the impact of the pandemic while the impact was from decreasing type during lockdowns. A pandemic-induced decline in the mean of consumption was linked to temporary migration by high-income families, whereas the water consumption of others faced an increase. The impact has also increased the slope of the linear relationship between the annual water consumption and population density increased by 3.5%. The proposed model estimated the annual water consumption with the accuracy of %3.77, %1.82, and %1.85 for two years before, one year before and one year after the pandemic, respectively.

Keywords: Modeling, Pandemic-induced changes, Spatial variables, Urban water, Water consumption groups

Introduction

The management of urban water systems still follows traditional methods while urban water systems are being redefined for a digital age in light of the increasing use of cyber-attacks and the occurrence of pandemics (Moy de Vitry et al. 2019). Understanding water usage elements that can depict the urban water use pattern (UWUP) is essential for assessing the robustness of urban water supply infrastructure in the event of a pandemic (Kalbusch et al. 2020). The COVID-19 pandemic has proven lethal to human beings, let alone overwhelm the water supply and sanitation sector ever since its outbreak (Antwi et al. 2021). Households are among the largest water users (Murwirapachena 2021), accounting for the majority of urban water network subscriptions in the Tabriz metropolitan area (WWC-EAP 2021). As a result, analysis of their water use pattern can help develop appropriate measures to increase the resiliency of an urban water supply system in the face of traditional and newly emerging challenges.

Reducing water use through demand management strategies without compromising consumer service or severely influencing healthy, sustainable urban living poses a significant challenge to sustainable cities (Crouch et al. 2021). A variety of ranges have been reported for the required liters per capita per day (LPCD) from between 20 and 50 LPCD (Howard et al. 2003) to between 40 and 509 LPCD (Athuraliya et al. 2012; Biswas and Gangwar 2021; Hay et al. 2012). Such an extensive range shows its dependency on various factors such as the type of household water use, the efficiency of urban service, local climate, and building preferences (Crouch et al. 2021). Prior to the COVID-19 epidemic, studies revealed that investment in network maintenance is the most essential method to save water, followed by legislation, individual meters, and public campaigns advocating for the adoption of water-saving devices and water conservation behaviors. Domestic water use has been linked to a variety of other factors such as household size, demographics, and climate, as well as behavioral and psychological characteristics in these studies (Tortajada et al. 2019). Due to the necessity of observing personal hygiene and home-base activities during lockdowns, a portion of information about urban water consumption patterns (UWCPs) has been left unknown during the COVID-19 pandemic.

While the impact of the COVID-19 pandemic on the water ecosystem was under study, as was the case in 2020 (Ji et al. 2021), the study of UWCPs during pandemic conditions has also been considered by researchers. In a case study in Joinville, Southern Brazil, statistical tests on the measured residential water use data gathered by telemetry in two periods (before and after quarantine) were used to see if there was a difference between the investigated periods. Water use decreased in the commercial, industrial, and public categories while increasing in the residential category. The regression model findings confirmed the effect of restrictions on non-residential consumption (Kalbusch et al. 2020). The impacts of COVID-19 on multi-dimensional UWCPs with the aim of forecasting water demand in the Tabriz metropolitan area were studied using an integrated approach of spatial analysis and regression-based autocorrelation. Results showed that the annual water deficit in Tabriz increased up to 30% in 2020, which may still further increase to about 40–45% in 2021 (Feizizadeh et al. 2021a). In an investigation of how the COVID-19 pandemic influenced California's urban water use by water demand modeling, normal water consumption was subtracted from actual use. It was found that the pandemic response reduced California's urban water use by 7.9%. The drop was mostly due to an 11.2% drop in the commercial, industrial, and institutional sectors, which more than negated a 1.4% growth in the residential sector (Li et al. 2021). In Wroclaw, Poland, the share of household water consumption increased significantly from 75.8% in March 2020 to 81.6% in April of the same year. On the other hand, water consumption in commercial buildings has decreased from 9.1 to 6.8%, and in educational buildings has decreased from 3.4 to 1.9% (Kazak et al. 2021).

The effect of stay-at-home habits on urban water usage was revealed to be somewhat larger than the combined influence of other non-pandemic variables (Li et al. 2021). Analysis of water consumption data from 11,528 households in England over a 20-week period beginning in January 2020 revealed a respective rise in household water consumption across the network of 13%, 22%, and 29% in March, April, and May 2020 (Abu-Bakar et al. 2021). Water usage habits during the COVID-19 pandemic matched other prominent features of the new normal 'stay-at-home' lifestyle when families are starting their days later. The morning consumption peaks two hours later and the greatest rise in water usage occurs in the afternoon, when stay-at-home schoolers and employees take a break, get up to use the toilet, wash their hands, and prepare meals (Dzimińska et al. 2021). An increase in home water consumption and the decline in commercial demand during the COVID-19 pandemic were confirmed in the UK (Renukappa et al. 2021). Higher residential water consumption per day (approximately 14%) with higher peak morning and evening demands during the day has also been proven in Germany (Lüdtke et al. 2021). The impact of the coronavirus on consumption patterns varies even within the residential category, with peak demand being higher in low-income areas than in high-income areas, in the United Arab Emirates (Rizvi et al. 2021). Since the custom of traveling during holidays changes between cities, case studies are recommended to detect the potential effect of outbound tourism on water demand, especially after the outbreak of the COVID-19 pandemic (Bich-Ngoc and Teller 2020).

The COVID-19 pandemic has demonstrated the importance of water availability for the public health sector in the event of a pandemic (Renukappa et al. 2021). Understanding UWUPs can help with design, cost savings (substantiality), hydraulic relief, and an accurate representation of water usage by integrating it with a network's pump timings (Rizvi et al. 2021). Although the literature on the subject has so far confirmed an increase in household water consumption during the COVID-19 pandemic, on the other hand, the different ranges presented for it indicate that cities and their consumption patterns can be unique. Since the discovery of UWCPs in the pandemic condition is a new topic and also because of customary variances in cities, more case studies are still recommended. Previous attempts at the modeling stage were mostly on estimating the subscribers' water demand by linking weather data to individual water consumption using statistical and machine learning methods (Pesantez et al. 2020; Duerr et al. 2018). The role of geographical parameters and the change in their role due to the impact of the COVID-19 pandemic in an household water consumption pattern (HWCP) estimating model is still the missing part.

The present study tried to compare the HWCP in two years before and a year after the outbreak of the COVID-19 pandemic, to discover the HWCP on annual and bi-monthly time scales and also in different spatial classes. The study also investigated the contribution of spatial variables to the HWCP for the first time. Following the identification of the HWCP, a novel model was developed to estimate the spatiotemporal distribution of the annual household water consumption based on the discovered relations between water consumption and spatial variables. The parameters of the linear equations for each water consumption class (classes from very low to very high water consumption) were determined using the training data (data for two years before the pandemic). The model was then tested for the data belonging to one year before the pandemic and for the data belonging to one year following the COVID-19 pandemic. The results of this study can be of importance to urban water management sectors in similar metropolitans. The research was conducted using Tabriz city (a megacity located in the northwest of Iran) water usage data from March 21, 2018, to March 20, 2021.

Materials and methods

Study area

Tabriz is Iran's fourth biggest city, with a population of almost 2 million and an area of around 245 km2 with industrial and old cultural features, located at 1321 m above sea level. The city is home to a semi-arid climate regime that makes its water supply components vulnerable to climate-induced and human-induced droughts. The Water and Wastewater Company of East Azerbaijan Province (WWC-EAP) produces 136 MCM of water per year, mostly destined for drinking, from four sources. These resources consist of a river, a dam, wells, and qanats (aquifer-style structures), which account for 55, 33, 12, and 1% of the total water production, respectively. Drilling new wells, reclaiming abandoned wells, and increasing the release of water from the Nahant Dam are currently the only remedies to compensate for Tabriz's drinking water shortages (WWC-EAP 2021). Tabriz is located in one of the most susceptible locations of the Urmia Lake basin, where the water content of Lake Urmia has dried up 85% so far, and the Iranian government has embarked on efforts in collaboration with international bodies in recent years to revive it (Sima et al. 2021). Following the establishment of the Urmia Lake Restoration National Committee (ULRNC), efforts to raise public awareness about the water problem have lately increased in the Urmia lake basin including Tabriz city (ULRNC 2014). Out of a total of 650,000 household subscribers (Feizizadeh et al. 2021a), only subscribers with known geographical coordinates and who were members of the urban water network from the beginning of the study period (418,911 subscribers) participated in the study. Figure 1 depicts the Tabriz urban water network and the spatial distribution of its subscribers. Since early 2020, the city has endured numerous waves of the COVID-19 pandemic, and a home quarantine began in Tabriz at that time, creating changes in UWCPs (Feizizadeh et al. 2021a).

Fig. 1.

Fig. 1

Open in a new tab

Outline of Tabriz urban water network and spatial distribution of its household subscribers

Flowchart of the work

As indicated in Fig. 2, the whole process designed for conducting the presented study starts with preparing a data bank from the readily available geographical layers and periodic records of water consumption of the Tabriz household subscribers. These data were then used to extract the HWCP of the city before and after the outbreak of the COVID-19 pandemic. In order to be able to estimate the spatiotemporal distribution of the Tabriz household water consumption, it was necessary to identify the parameters affecting it and also expected changes in the coefficients of these parameters due to the pandemic. To achieve this goal a recognition phase was designed to detect the existing relationship between the geographical variables and the HWCP and then in the estimation phase based on the detected relationships a function was developed for estimating the HWCP.

Fig. 2.

Fig. 2

Open in a new tab

Flowchart of the research procedure including the three phases of preparation and analysis of the data, recognition of the HWCP and estimation of the HWCP

Data in use for recognizing the HWCP

In this study, the finest available scale of household water consumption data which is on a bi-monthly timescale was obtained from WWC-EAP for two years before and one year after the COVID-19 pandemic outbreak (from 21 March 2018 to 19 March 2021). Only subscribers who have been continuously present in the urban water network for all three years were preserved, while both new and sealed subscriptions were eliminated, in order to extract the impact of the COVID-19 pandemic on the amount and pattern of household water consumption more accurately. Since bi-monthly data recording by WWC-EAP is done according to the Iranian official calendar (solar calendar), Table 1 shows the equivalents of recording periods with intervals on the global calendar.

Table 1.

Conversion of local data recording periods to time intervals on global calendar

Time ID Cal Year Spring–Summer Fall–Winter
Bi-monthly time periods
P1 P2 P3 P4 P5 P6
Two years before the outbreak of pandemic 2-YBP Local 1397 Far. 1–Ord. 31 Kho. 1–Tir 31 Mor. 1–Sha. 31 Meh. 1–Aba. 31 Aza. 1–Dey 31 Bah. 1–Esf. 29
Global 2018–2019 Mar. 21–May 21 May. 22–Jul. 22 Jul. 23–Sep. 22 Sep. 23–Nov. 21 Nov. 22 –Jan. 20 Jan. 21–Mar. 20
The year before the outbreak of pandemic 1-YBP Local 1398 Far. 1–Ord. 31 Kho. 1–Tir 31 Mor. 1–Sha. 31 Meh. 1–Aba. 31 Aza. 1–Dey 31 Bah. 1–Esf. 31*
Global 2019–2020 Mar. 21–May 21 Mary 22–Jul. 22 Jul. 23–Sep. 22 Sep. 23–Nov. 21 Nov. 22–Jan. 20 Jan. 21–Mar. 19*
The year after the outbreak of pandemic 1-YAP Local 1399 Far. 1*–Ord. 31* Kho. 1*–Tir 31* Mor. 1*–Sha. 31* Meh. 1*–Aba. 31* Aza. 1*–Dey 31* Bah. 1*–Esf. 31*
Global 2020–2021 Mar. 20*–May 20* May. 21*–Jul. 21* Jul. 22*–Sep. 21* Sep. 22*–Nov. 20* Nov. 21*–Jan. 19* Jan. 20*–Mar. 20*
Open in a new tab

*Pandemic period

After obtaining bi-monthly and annual water consumption values for household subscribers, first, a comparison of consumption values in the recorded periods was performed between the years before and the year after the pandemic to identify the overall changes in the HWCP. The data analysis was then implemented by drawing scatter plots to test the probable dependency of the HWCP on some of the available spatial variables in the city. In this regard, four categories of geographical layers for Tabriz, including population density (GDRUD-EAP 2016), land value (GDTA-EAP 2013), elevation (Farr et al. 2007), and land surface temperature (LST) (Ermida et al. 2020), were prepared using available data resources in organizational reports and satellite images as shown in Fig. 3. The concept behind incorporating the variables into the analysis was to clarify the contribution of dense, high-income, heightened, hot, and their opposite areas in determining the HWCP before and after the pandemic. The concept behind incorporating the variables into the analysis was to clarify the contribution of high and low areas in terms of population density, income, height, and temperature in determining the HWCP before and after the pandemic. The incorporation made it possible to recognize the probable effect of different lifestyles on the HWCP while the probable climate and micro-climate-induced changes on HWCP can be identified by monitoring the HWCP variations against LST values. The involvement of the population data layer was due to the identification of the trend between the population and water consumption. To ease the depiction of the HWCP before and after the pandemic, a 500 m by 500 m grid was overlaid on the city, and the total and mean values for the annual household water consumption in each square were calculated.

Fig. 3.

Fig. 3

Open in a new tab

The HWCP analysis based on household water meter recorded data and prepared raster layers of parameters assessed for HWCP estimations

Estimation of the HWCP

After recognizing the HWCP in different spatial classes and discovering the relationship between water consumption and spatial variables, the training data on water consumption (annual household water consumption data in 2 year before the pandemic) were classified into six classes very low to very high water consumption classes as shown in Table 2. The thresholds of the classes were determined in consultation with an expert panel and based on factors such as network hydraulic capacity, consumption records, and distribution of the data. In each of the classes, the parameters of linear equations between the number of subscribers and population density (a spatial variable with which the linearity of the relationship has already been discovered) were determined. After the determination of the parameters of linear equations, variation of parameters was investigated by a change in water consumption class and an overall model was built using the detected linear and nonlinear equations to estimate the spatiotemporal distribution of the annual household water consumption using four main parameters including consumption class, a number of household water subscribers in each class, population density, and the representative amount of water consumption for the class. For those variables that did not show the linearity, an attempt was made to apply their effect by weights in the model. To do this, their impacts on the mean water consumption which was discovered in the first round of analysis were considered to determine the weights, and finally, the amount of consumption in grids was estimated using the discovered equations and weights.

Table 2.

Water consumption classes in Tabriz city

Class no 1 2 3 4 5 6
Class description Very low consumptions Low consumptions Less than usual consumptions Usual consumptions High consumptions Very high consumptions
Water consumption range (Cubic meter per year in a 500 × 500 m square) Less than 50 Between 50–100 Between 100–150 Between 150–250 Between 250–400 Greater than 400
Representative amount of water consumption for the class (Cubic meter per year in a 500 × 500 m square) 25 75 125 200 325 400
Open in a new tab

Results and discussion

Recognition of the HWCP

As shown in Fig. 4, in a city like Tabriz, which is equipped with a water network that continuously supplies water to household consumption for approximately two million people, the total bi-monthly household water consumption is at least 12.4 MCM. However, the difference in household water consumption in bi-month periods could vary up to 1 MCM when compared within two years before the pandemic (2-YBP) and one year before the pandemic (1-YBP) and up to 3 MCM when they were compared within one year after the pandemic (1-YAP). When comparing each bi-monthly period with the same period in another year, it was found that if the comparison is between 2-YBP and 1-YBP, the difference in total household water consumption is up to 1 MCM, and if it is between 1-YAP and a normal year, the difference in the total consumption can increase up to 1.93 MCM. The highest increases in water consumption due to the pandemic were 11% for May–July among bi-monthly periods, whereas the pandemic had a decreasing effect on water consumption during the peaks of the pandemic when severe lockdowns were implemented by the government (Fig. 4a). This may be explained as restricting activities that use a lot of water and shifting the water consumption pattern to more sanitary purposes during lockdowns. Although the percentage of the increase in the total yearly consumption was only 3% (Fig. 4b) for 1-YAP, the cumulative impact of such modifications can change the HWCP in the future. It must be considered that the percentage of the increase in water consumption is attributed only to old subscribers while new subscribers of 1-YBP and 1-YAP have their own increasing effect which was not the case in this study. Due to the fact that in this study, consumption by new network subscribers was not involved in the calculations, it can be concluded with high confidence that changes in the total household water consumption in 2-YBP and 1-YBP were due to public awareness programs for water-saving and in 1-YAP were due to the impact of the COVID-19 pandemic. Thus, according to Fig. 4a, although there was a tendency to save on the total household water consumption in 1-YBP the outbreak of the COVID-19 pandemic led to disrupting the trend by increasing consumption. The total bi-monthly household water consumption decreased in 1-YBP compared to bi-monthly periods in 2-YBP except in P6, which had a 2.3% increase in consumption. The increase that occurred in P6 was due to the fact that half of this period belongs to the period following the outbreak of the COVID-19 pandemic. As indicated in Fig. 4c, the trend of changes in the mean and 25 and 75 quantiles for 2-YBP and 1-YBP was more similar to each other rather than to 1-YAP. The effect of the pandemic on household water consumption was incremental, especially in the P2 and P5 periods. In general, it can be said that the COVID-19 pandemic has led household water consumption to have a wider range and follow a more different pattern in 1-YAP compared to ranges and patterns in 2-YBP and 1_YBP.

Fig. 4.

Fig. 4

Open in a new tab

Changes in totals and patterns in household water consumption on a bi-monthly and annual basis throughout two normal and COVID-19 pandemic years. a Total bi-monthly household water consumptions and the COVID-19 pandemic wave (The wave was obtained by standardizing the number of COVID 19 infections per month announced by the Iranian Ministry of Health and Medical Education for 1-YAP), b total annual household water consumptions and c changes in the average and range of household water consumptions on a bi-monthly and annual basis

As is clear from Fig. 5a–d, Tabriz incorporates a variety of population densities, land values, elevation, and LST in its area. Although a considerable part of the elevated lands in Tabriz have higher than average land values (Fig. 5b and c), the classified map of the population density and LST have more similarities to each other (Fig. 5a and d). The inverse relationship observed between building density and LST in the cities has already been discovered in China (Song et al. 2020). Figure 5e compares the total and mean values for the annual household water consumption in the classified areas based on the average amount of spatial variables. In the areas where population density, land value, elevation, and LST were higher than average, the share of total annual consumption values was approximately 78%, 34%, 40%, and 25% in all three years, respectively. Because there have been no substantial changes in the mentioned percentages before and after the outbreak of the COVID-19 pandemic, it can be stated that the pandemic has had no major impact on the existing pattern for the spatial distribution of the total annual consumption values.

Fig. 5.

Fig. 5

Open in a new tab

Spatial distribution of 500 m × 500 m squares in which a elevation, b land value, c population density, and d LST were above or below average to generate, e for comparing total annual household water consumption in each classification

As it is depicted in Fig. 5e, the mean of annual consumption was always higher in four classes where population density is below average and land value, elevation, and LST are above average compared to their opposite classes. This shows that the COVID-19 pandemic has not changed its ranking yet while it may not be kept as a constant pattern for the future. Figure 5e further shows that over 1-YAP, the mean of annual consumption in the regions with lower than average population densities declined by about 18% (about 62 cubic meters per year), whereas it grew by about 6% in the areas with higher than average population densities (12 cubic meters per year). This shows that the COVID-19 pandemic has amplified household water consumption in areas with concentrated populations. The mean annual consumptions during 1-YAP in classes with above-average land value, elevation, and LST have dropped compared to mean values during 1-YBP and 2-YBP, whereas mean values in classes with below-average elevation and LST have increased. Around the geographical center of the city, the mean annual water consumption has increased after the outbreak of the pandemic while by moving from the center to the margins especially southward, the mean value has decreased. Since the southern and southeastern parts of Tabriz have a higher land value and usually high-income groups live in these areas, the observed decrease in the mean of consumption after the pandemic in these areas can be interpreted as a change in the HWCP. This shift demonstrates that the pandemic has led the consumption pattern to shift from high consumption prior to the epidemic to limited usages such as hygiene and cleanliness after the pandemic in these areas.

To investigate any possible relationship between the variables for describing the HWCPs before and after the COVID-19 pandemic, Fig. 6 shows scatter plots of the total and mean values of annual household water consumptions in 500 × 500 m pixels versus spatial variables, as well as scatter plots of spatial variables versus each other. Figure 6a and b confirms the existence of a linear relationship between the sum of annual water consumptions in the pixels and population density (with R-square more than 0.9), the slope of which has been increased by 3.5% after the outbreak of the COVID-19 pandemic. In contrast with the mentioned positive linear correlation, the relationship was reversed between the sum of consumptions and the amounts of LST, which was due to the reverse correlation between LST and population density (Fig. 6d). No correlation was detected between the sum of annual water consumption in different locations and the two other spatial variables (land value and elevation). The mean of annual water consumption in different locations did not show any significant correlation with spatial variables before or after the outbreak of the COVID-19 pandemic either, while the plots revealed different amounts of scattering along Y axes (the amount of scattering in the left and right sides of the dotted lines in Fig. 6a and b). Figure 6c confirms that the pandemic has amplified the annual water consumption in different locations in Tabriz city since the slope of the line between consumption values in 1-YBP and 1-YAP is almost 3.3% higher than the 1:1 line. Figure 6e indicates that there was no significant correlation between land value, elevation, and population density while the dotted box in this figure shows only limited pixels with above-average land value located in dense areas. This indicates that residents of high-value lands in less populous regions are those who have contributed to the decrease in mean annual water consumption (the decrease which is shown in Fig. 5e), despite the fact that the contribution of other residents resulted in an increase in water consumption over 1-YAP.

Fig. 6.

Fig. 6

Open in a new tab

Scatter plots of total and mean values of annual household water consumptions in 500 m × 500 m squares a versus spatial variables in 1-YBP, b versus spatial variables in 1-YAP and c versus each other as well as the scatter plots for d land value, elevation and population density versus each other and e LST versus population density. The dotted lines show the average value for spatial variables and the dotted box shows number of high-value lands in dense areas

Figure 7 indicates a decrease and an increase in the amount of consumption before and after the outbreak of the COVID-19 pandemic, respectively, in garden houses. Since garden houses located on the border of Tabriz city are usually owned by high-income families to spend holidays and summertime, the huge increase in the water consumption of these residential places during the pandemic indicates temporary migration from house to garden houses. The existence of temporary migration confirms the amount of decrease in the mean household water consumption after the pandemic (detected in Figs. 5 and 6) in the south and southeastern areas of the city, inhabited by high-income families. According to the results obtained from Figs. 5, 6, and 7, a shift in the amount of water consumption between different user types can be expected in the city, which calls for further studies on the water consumption of different users.

Fig. 7.

Fig. 7

Open in a new tab

Percentage of decrease and increase in annual water consumption in garden houses. The blue bar indicates a decrease and the red bar indicates an increase in the amount of consumption before and after outbreak of COVID-19 pandemic, respectively, for this type of consumption

Estimation of the water consumption

As previously shown in Fig. 6, population density is the only spatial variable that is linearly related to the sum of water consumption in grids. This made it a concept for authors to investigate the linearity of the relation between the number of subscribers from different water consumption classes and population density and to use it for predicting annual water consumption. Figure 7 shows the number of subscribers with usual annual water consumption (class 4) in grids increases with the increment in population density by the highest amounts of slope and correlation while the slope of lines decreases for the classes that bias from the usual amount of water consumption. Based on depicted linear relations in Fig. 8, the amount of annual water consumption in a grid with a known amount of population density is predictable if the multiplied amounts of counts by the representative water consumption are accumulated. Figure 8 also indicates the detected polynomial relations between the parameters of the linear equation and household water consumption classes. Based on the detected linear and polynomial equation, the overall model for estimation of the annual household water consumption in grids can be written as Eqs. 1 and 2:

Yi=-94.117i2+665.95i-493.66X+8.3489i2-60.762i+63.435 1
HWC=∑(Ri×Yi) 2

where Yi is the number of household water subscribers in each class (number in 500 × 500 m square), X is population density (average in 500 × 500 m square), i is the consumption class, HWC is the annual household water consumption in grids (cubic meter in 500 × 500 m square) and Ri is the representative amount of water consumption for the class (cubic meter per year in a 500 × 500 m square). The integrated form of Eqs. 1 and 2 can be written as Eq. 3:

HWC=∑(Ri×-94.117X+8.3489i2+665.95X-60.762i-493.66X+63.435) 3

Fig. 8.

Fig. 8

Open in a new tab

a Liner relations between the number of subscribers in grids from different water consumption classes and the standardized values of population density in 2-YBP, b polynomial relations between the parameters of the linear equation and household water consumption classes

To improve estimations, the effects of spatial variables on each grid were also applied by attributing corresponding weights in the model. As previously shown in Fig. 5, the mean amounts of water consumption were considerably different where the values of spatial variables were above the average value compared to those where the variables were below the average value. Table 3 shows the attributed weights for the spatial variables according to the mean amounts of water consumption in the dual categories.

Table 3.

Attributed weights for averaging the annual household water consumption in grids based on the values of spatial variables

Spatial variables
Land values Elevation LST Population density
Weights for averaging in grids where the values of the spatial variables are above average 1.56 1.82 1.82 0.57
Open in a new tab

Figure 9 shows the interpolated map of the observed and estimated HWCP values in Tabriz city’s and the corresponding maps obtained from RMSE values in 2-YBP and 1-YBP as the trained and the tested years, respectively. Modified linear coefficients and weights were applied in the model to take the impact of the COVID-19 pandemic into consideration in estimating annual water consumptions for 1-YAP. As shown in Fig. 9, the model was able to produce almost a similar spatial distribution of annual household water consumptions to the observed maps both in train and test years. The produced RMSE maps imply that the amount of error is high only in some spotty areas while the model could estimate the target parameter with rather a uniform amount of RMSE over the city. Table 4 shows the validation statistics for the model by comparing the model outputs with the observed amounts of household water consumption. According to Table 4, the model was able to estimate the annual household water consumption with the accuracy of %3.77, %1.82, and %1.85 for 2-YBP, 1-YBP, and 1-YAP, respectively. The model could successfully estimate the annual household water consumption in grids for the studied three years with a total RMSE of less than a million cubic meters.

Fig. 9.

Fig. 9

Open in a new tab

Observed, estimated, and RMSE maps a for 2-YBP as the test year, b for 1-YBP as the training year and c for 1-YAP as the test year with the modified model that has considered the impact of COVID-19 on household water consumption pattern

Table 4.

Percentage estimation error, RMSE and MAE for the train and test years

Year Percentage estimation error (%) Sum of statistics (m3/year)
RMSE MAE
Train 2-YBP 3.77 872,470 41,312
Test 1-YBP 1.82 785,732 37,205
1-YAP 1.85 819,670 38,812
Open in a new tab

Conclusion

Research findings on the HWCP

This study opened a new insight into the interpretation of recorded data from household water meters, considering the attributed spatial characteristics of urban water network subscribers. The incorporation of spatial analysis helped scrutinize the impact of the COVID-19 pandemic on the HWCP more in detail while such a smooth footprint could not be captured only by recording water meter data. Local water consumption per capita (in applications such as bathing, washing, and toileting among others) tends to be a fixed number (Howard et al. 2003), and in cities with continuous water supply, the monthly and yearly increases in water consumption following the outbreak of the COVID-19 pandemic were not a multiplied number in the literature (Li et al. 2021; Kazak et al. 2021; Feizizadeh et al. 2021b). This investigation confirmed the findings of previous studies, indicating that the pandemic impact on household water requirements was a smooth rising with the largest recorded increases of 11% and 3% on bi-monthly and annual scales, respectively. The smoothness of the footprint does not mean that it can be ignored, because its cumulative effect in the coming years can disrupt the urban water system, particularly in a city like Tabriz, where water resources were already severely threatened by climate change, hydrological and managerial droughts prior to the pandemic (Feizizadeh et al. 2021a). It should also be noted that by integrating new subscribers into the analysis, for those who joined the urban water network between 1-YBP and 1-YAP, a bigger percentage rise in the total yearly water consumption (more than 3%) may be projected during 1-YAP.

Since the footprint of the pandemic on the HWCP was detectable even with yearly and bi-monthly scale data, it can be expected that water consumption data with smaller time scales contain more profound impacts of the pandemic. Hence, providing infrastructure for recording data at the finest time scales in Tabriz is critical for effective urban water management during the pandemic. Since bi-monthly data were the finest temporal resolution available for household water consumption in Tabriz the study has tried to provide a methodology to depict the HWCP by incorporating other spatial data. The effect of the pandemic on household water consumption was found to be increasing and decreasing in bi-monthly periods, with the consequence of the variations finally leading to a rise in annual consumption following the emergence of the pandemic. Residents of areas where population density was higher than average and areas where land value and elevation were lower than average had more proportions of the total annual water consumption and the outbreak of the COVID-19 pandemic did not cause any change. The southern and southeastern parts of Tabriz are inhabited by high-income groups and the observed decrease in the mean water consumption after the pandemic in these areas can be considered as the footprint of the pandemic on the HWCP. This change was linked to inner-city immigration by high-income families from southern areas to garden houses. During the lockdowns, restrictive measures causing a big increase in water consumption and shifting water consumption patterns from diverse to more hygienic purposes might be another factor for the observed change in the HWCP. The observed shift in the amount of water consumption between different categories can be a reason for conducting further studies on the water consumption of different users. In contrast to marginal parts, residents of central areas contributed to the increase in water consumption during the pandemic. It was found that due to the impact of the COVID-19 pandemic the slope of the linear relationship between the sum of annual water consumption and population density increased by 3.5%. Based on the findings of this study, it can be stated that the methodology proposed for the interpretation of water consumption data in combination with spatial variables can recognize the contribution of spatial classes to water consumption. This can be a base knowledge for implementing effective urban water management.

Research findings on the HWCP estimating model

Modeling and forecasting water demand at the user level using a sort of predictors including past demand, weather data, and household characteristics had already been conducted (Pesantez et al. 2020; Deoreo and Mayer 2012). In the present study, a novel model was built to capture the role of readily available geographical variables in estimating the spatiotemporal distribution of HWCP in both normal and COVID-19 pandemic scenarios. The annual household water consumption could successfully be estimated using the four main parameters: consumption class, the number of household water subscribers in each class, population density, and the representative of water consumption for the class. The role of geographical variables was taken into account by attributing the weights for averaging the annual household water consumption in each grid. Modification of the model for the COVID-19 pandemic situation was done by applying the pandemic-induced increase in the amount of the slope of the linear equation between consumption and population density (3.5%). Since the impact of the pandemic had also changed the mean of household water consumption in dual categories, the attributed weights for averaging the annual household water consumption were also modified to capture the effect of the pandemic. The average amount of the percentage estimation error during the studied period was 2.48% for the model indicating that the model is able to provide a reliable estimation of the annual water consumption of household subscribers both in normal and the pandemic situation. The impact of the pandemic on the HWCP indicates that the HWCP can be susceptible to change due to phenomena that shock the urban lifestyle. Hence, it can be recommended that the parameters of the model are updated each year before its use of it for estimating the spatiotemporal distribution of annual household water consumption for the following year.

Acknowledgements

We would also like to show our gratitude to the Water and Waste Water Company of East Azerbaijan Province for providing the data and to the Iranian Water Resources Management Company for arrangements.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declarations

Conflict of interest

The authors declare there are no competing interests relevant to the content of this article.

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.

References

  1. Abu-Bakar H, Williams L, Hallett SH. Quantifying the impact of the COVID-19 lockdown on household water consumption patterns in England. Npj Clean Water. 2021;4:1–9. doi: 10.1038/s41545-021-00103-8. [DOI] [Google Scholar]
  2. Antwi SH, Getty D, Linnane S, Rolston A. COVID-19 water sector responses in Europe: A scoping review of preliminary governmental interventions. Sci Total Environ. 2021;762:143068. doi: 10.1016/j.scitotenv.2020.143068. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Athuraliya A, Roberts P, Brown A. Residential water use study volume 2-Summer 2012. Melbourne: Yarra Valley Water; 2012. [Google Scholar]
  4. Bich-Ngoc N, Teller J. Potential effects of the COVID-19 pandemic through changes in outbound tourism on water demand: the Case of Liège (Belgium) Water. 2020;12:2820. doi: 10.3390/w12102820. [DOI] [Google Scholar]
  5. Biswas A, Gangwar D. Studying the water crisis in Delhi due to rapid urbanisation and land use transformation. Int J Urban Sustain Dev. 2021;13:199–213. doi: 10.1080/19463138.2020.1858423. [DOI] [Google Scholar]
  6. Crouch ML, Jacobs HE, Speight VL. Defining domestic water consumption based on personal water use activities. AQUA Water Infrastruct Ecosyst Soc. 2021;70:1002–1011. doi: 10.2166/aqua.2021.056. [DOI] [Google Scholar]
  7. Deoreo WB, Mayer PW. Insights into declining single-family residential water demands. J Am Water Works Assoc. 2012;104:E383–E394. doi: 10.5942/jawwa.2012.104.0080. [DOI] [Google Scholar]
  8. Duerr I, Merrill HR, Wang C, Bai R, Boyer M, Dukes MD, Bliznyuk N. Forecasting urban household water demand with statistical and machine learning methods using large space-time data: A Comparative study. Environ Model Softw. 2018;102:29–38. doi: 10.1016/j.envsoft.2018.01.002. [DOI] [Google Scholar]
  9. Dzimińska P, Drzewiecki S, Ruman M, Kosek K, Mikołajewski K, Licznar P. The use of cluster analysis to evaluate the impact of COVID-19 pandemic on daily water demand patterns. Sustainability. 2021;13:5772. doi: 10.3390/su13115772. [DOI] [Google Scholar]
  10. Ermida SL, Soares P, Mantas V, Göttsche F-M, Trigo IF. Google earth engine open-source code for land surface temperature estimation from the landsat series. Remote Sens. 2020;12:1471. doi: 10.3390/rs12091471. [DOI] [Google Scholar]
  11. Farr TG, Rosen PA, Caro E, Crippen R, Duren R, Hensley S, Kobrick M, Paller M, Rodriguez E, Roth L. The shuttle radar topography mission. Rev Geophys. 2007 doi: 10.1029/2005RG000183. [DOI] [Google Scholar]
  12. Feizizadeh B, Omarzadeh D, Ronagh Z, Sharifi A, Blaschke T, Lakes T. A scenario-based approach for urban water management in the context of the COVID-19 pandemic and a case study for the Tabriz metropolitan area. Iran. Sci Total Environ. 2021;790:148272. doi: 10.1016/j.scitotenv.2021.148272. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Feizizadeh B, Ronagh Z, Pourmoradian S, Gheshlaghi HA, Lakes T, Blaschke T. An efficient GIS-based approach for sustainability assessment of urban drinking water consumption patterns: A study in Tabriz city. Iran. Sustain Cities Soc. 2021;64:102584. doi: 10.1016/j.scs.2020.102584. [DOI] [Google Scholar]
  14. GDRUD-EAP: General Department of Roads and Urban Development of East Azerbaijan Province (2016) Development and civil (Comprehensive) plan of Tabriz (in Farsi). https://azarsharghi.mrud.ir
  15. GDTA-EAP: General Department of Tax Affairs of East Azerbaijan Province (2013) transaction value of Tabriz and its suburbs (in Farsi)
  16. Hay E, Riemann K, Van Zyl G, Thompson I. Ensuring water supply for all towns and villages in the Eastern Cape and Western Cape Provinces of South Africa. Water. 2012;38:437–444. doi: 10.4314/wsa.v38i3.9. [DOI] [Google Scholar]
  17. Howard G, Bartram J, Sanitation Water. World Health Organisation Domestic water quantity, service level and health. World Health Organization. 2003 doi: 10.1016/j.jece.2021.105357. [DOI] [Google Scholar]
  18. Ji B, Zhao Y, Wei T, Kang P. Water science under the global epidemic of COVID-19: Bibliometric tracking on COVID-19 publication and further research needs. J Environ Chem Eng. 2021;9:105357. doi: 10.1016/j.jece.2021.105357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kalbusch A, Henning E, Brikalski MP, de Luca FV, Konrath AC. Impact of coronavirus (COVID-19) spread-prevention actions on urban water consumption. Resour Conserv Recycl. 2020;163:105098. doi: 10.1016/j.resconrec.2020.105098. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kazak JK, Szewranski S, Pilawka T, Tokarczyk-Dorociak K, Janiak K, Swiader M. Changes in water demand patterns in a European city due to restrictions caused by the COVID-19 pandemic. Desalin Water Treat. 2021;222:1–15. doi: 10.5004/dwt.2021.27242. [DOI] [Google Scholar]
  21. Li D, Engel RA, Ma X, Porse E, Kaplan JD, Margulis SA, Lettenmaier DP. Stay-at-home orders during the COVID-19 pandemic reduced urban water use. Environ Sci Technol Lett. 2021;8:431–436. doi: 10.1021/acs.estlett.0c00979. [DOI] [PubMed] [Google Scholar]
  22. Lüdtke DU, Luetkemeier R, Schneemann M, Liehr S. Increase in daily household water demand during the first wave of the COVID-19 pandemic in Germany. Water. 2021;13:260. doi: 10.3390/w13030260. [DOI] [Google Scholar]
  23. Moy De Vitry M, Schneider MY, Wani OF, Manny L, Leitão JP, Eggimann S. Smart urban water systems: what could possibly go wrong? Environ Res Lett. 2019;14:081001. doi: 10.1088/1748-9326/ab3761. [DOI] [Google Scholar]
  24. Murwirapachena G. Understanding household water-use behaviour in the city of Johannesburg, South Africa. Water Policy. 2021;23:1266–1283. doi: 10.2166/wp.2021.157. [DOI] [Google Scholar]
  25. Pesantez JE, Berglund EZ, Kaza N. Smart meters data for modeling and forecasting water demand at the user-level. Environ Model Softw. 2020;125:104633. doi: 10.1016/j.envsoft.2020.104633. [DOI] [Google Scholar]
  26. Renukappa S, Kamunda A, Suresh S. Impact of COVID-19 on water sector projects and practices. Util Policy. 2021;70:101194. doi: 10.1016/j.jup.2021.101194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Rizvi S, Rustum R, Deepak M, Wright GB, Arthur S. Identifying and analyzing residential water demand profile; including the impact of COVID-19 and month of Ramadan, for selected developments in Dubai, United Arab Emirates. Water Supply. 2021;21:1144–1156. doi: 10.2166/ws.2020.319. [DOI] [Google Scholar]
  28. Sima S, Rosenberg DE, Wurtsbaugh WA, Null SE, Kettenring KM. Managing Lake Urmia, Iran for diverse restoration objectives: Moving beyond a uniform target lake level. J Hydrol Reg Stud. 2021;35:100812. doi: 10.1016/j.ejrh.2021.100812. [DOI] [Google Scholar]
  29. Song J, Chen W, Zhang J, Huang K, Hou B, Prishchepov AV. Effects of building density on land surface temperature in China: Spatial patterns and determinants. Landsc Urban Plan. 2020;198:103794. doi: 10.1016/j.landurbplan.2020.103794. [DOI] [Google Scholar]
  30. Tortajada C, González-Gómez F, Biswas AK, Buurman J. Water demand management strategies for water-scarce cities: The case of Spain. Sustain Cities Soc. 2019;45:649–656. doi: 10.1016/j.scs.2018.11.044. [DOI] [Google Scholar]
  31. ULRNC: Urmia Lake Restoration National Committee (ULRNC) (2014) Rehabilitation of Lake Urmia causes of drought and threats (in Farsi). https://www.ulrp.ir/en
  32. WWC-EAP: Water & Waste Water Company of East Azerbaijan Province (2021) Aannual water consumption reports of East Azerbaijan Province's cities (in Farsi). https://en.abfaazarbaijan.ir

Articles from International Journal of Environmental Science and Technology are provided here courtesy of Nature Publishing Group

RESOURCES