The impact of COVID-19 and natural disasters on water consumption across sectors: case studies of Barbados and Trinidad and Tobago

Warm crystal blue waters, palm trees and white sandy beaches are popular pictures that come to mind whenever the Caribbean is mentioned. Yet these benign images obscure the fact that the Caribbean plays host to several natural hazards that impact it with seemingly increasing frequency (Robinson 2018). These include hydroclimatic events, such as tropical storms and hurricanes, accompanied by floods, debris flows and mass earth movements; extreme temperatures; droughts; geohazards, such as earthquakes, volcanic eruptions and tsunamis; and recurrent epidemics often associated with vector-borne diseases, particularly mosquitos (UN Office for the Coordination of Humanitarian Affairs (OCHA) 2020). To this list of hazards, we can now add public health as exemplified in the COVID-19 pandemic (Jansen & Bulbena 2021). Some of the hazards the Caribbean faces are cyclical in nature, while others are either fast or slow in their onset and impacts (Robinson 2018). Given the level of exposure and the vulnerability of Caribbean Small Island Developing States (SIDS) to natural hazards (Economic Commission for Latin America and the Caribbean (ECLAC) 2011; Ram et al. 2019), focus must be placed on understanding the cumulative impacts. This is of particular importance as these recurring hazards result in complex social, economic, and ecological consequences that can last long after an event, leaving the Caribbean States in a constant recovery mode.

Climate change is expected to magnify hydroclimatic hazards (World24 Meteorological Organization (WMO) 2021). The consequent adverse effects on weather-sensitive sectors, such as water and sanitation, impact a country's water security, biodiversity, agriculture and food security, and health (Rozenberg et al. 2021). However, it is not just hydroclimatic hazards and climate change that are affecting water security. The impact of COVID-19 has highlighted the many ways in which the provision of water services is vulnerable to the effects of health-related restrictions on supply chains, personnel and finances (Janson et al. 2021). COVID-19 was declared a global pandemic by the World Health Organization (WHO) on 11 March 2020. Furthermore, the nature of the pandemic appears to have brought on changes in modes of employment and working practices, and changes in consumption patterns which could have profound long-lasting impacts on how water services are managed (Poch et al. 2020).

The pandemic also demonstrated the strong national dependence across economic sectors on the reliable provision of water and sanitation services in efforts to curb the spread of the virus (Abu-Bakar et al. 2021; DeMaagd et al. 2021; Giné-Garriga et al. 2021). In addition to COVID-19, in 2021, Barbados was impacted by two natural hazards. The first event, on 9 April, was the eruption of the La Soufrière volcano, which is located on the nearby island of Saint Vincent. This produced an ash plume reaching approximately 10,000 m, which drifted eastwards towards the Atlantic Ocean (King 2021; United Nations Development Programme (UNDP) 2021). The second event occurred on 2 July 2021, when Hurricane Elsa, a category 1 hurricane, passed over Barbados. This was the first hurricane in 65 years to make landfall in the country (Morgan 2021). The situation of more than one natural hazard simultaneously affecting the island highlights the vulnerability of the Caribbean. It also emphasised the increasing need to consider and plan for consecutive and multiple hazard events (Roopnarine et al. 2021a, 2021b).

The purpose of this review paper is to examine the impacts of COVID-19 on water consumption in Barbados and Trinidad and Tobago, differentiating among consumer categories. For Barbados, the assessment compared water consumption from February 2017 to March 2020 – the start of the pandemic – with water consumption from March 2020 to August 2021. For Trinidad and Tobago, changes in water consumption between January 2017 and September 2021 were assessed. The rationale for the choice of these two countries as case studies is that firstly they offer an opportunity to explore the impact of the COVID-19 pandemic on water consumption in SIDS with contrasting socio-economic conditions, one being a mainly tourism-based economy (Barbados) and the other an industrial-based economy (Trinidad and Tobago). Secondly, the availability of water consumption data is often a challenge in the region. This data access allows the authors to contribute in a small way to addressing the paucity of studies in the Caribbean that focus on water consumption patterns. A further aim of this paper is to examine the implications of the findings for the future monitoring and analysis of water consumption as an integral part of water planning and management with multi-sectoral implications for national disaster management.

The paper is structured to first contextualise and capture the emerging patterns of water consumption related to the implementation of COVID-19 public health containment strategies and their impacts. This was done in the form of a literature review. The two case studies are then presented as independent but contrasting examples of the differences between water consumption across both countries. A combined discussion outlines the implications of the findings in the context of Caribbean SIDS. In the final section, some conclusions are drawn in the context of a changing environment. Namely that as a consequence of climate change, consecutive hazards are likely to become more common and will require more emphasis on forward planning (Mishra et al. 2021).

Among the economic effects of COVID-19 globally, activity in service industries, tourism and hospitality, and retail, all of which employ large numbers of people, dropped dramatically. People employed in these sectors lost a large part of their livelihood and were required to stay at, or return to, their homes (Deb et al. 2022). Countries sought to ensure that essential industries and services were maintained, but this often entailed increased costs and lower productivity, which was associated with the introduction of health measures to protect staff. However, these measures were not sufficient to avoid staff reductions and a contraction of the workforce (ECLAC-Pan-American Health Organization (PAHO) 2021). Particularly affected initially were global supply chains, although these adapted and diversified to address potential shortages (Giné-Garriga et al. 2021). In the Caribbean, water utilities reported challenges, for example, in sourcing chemicals for water treatment. According to Jansen & Bulbena (2021), across Latin America and the Caribbean (LAC), a discernible shift in consumption patterns was also noted, with residential consumption increasing and non-residential decreasing. This resulted in a decrease in revenue for water utilities. Interestingly, the study also revealed that some utilities showed increased profitability and liquidity, which stemmed from reduced operating costs and capital expenditure during the period (Janson et al. 2021).

The widespread disruption of economic activity due to the pandemic brought about behavioural changes in water consumption. The adoption of smart metering in some countries has provided interesting insights into these changes, which may not have otherwise come to light. The changes in patterns of water consumption and overall consumption were often linked to changes in economic activity (Poch et al. 2020; Abu-Bakar et al. 2021). Studies have also shown that generally as tourism, commercial and industrial activities were curtailed, water consumption declined, whereas residential consumption often increased as those who would otherwise have been at work were now at home.

A study from southern Brazil (Kalbusch et al. 2020) found that industrial consumption declined by 53%, while residential consumption increased by 11%, with a greater increase for apartments than for houses. The same study noted that the impact on commercial consumption was more complex because of the different subsectors. Consumption associated with the retail trade dropped, while that of support facilities, such as hospitals, increased. A similar pattern of changes in water consumption across consumer categories was seen in a study in Tabriz, Iran (Feizizadeh et al. 2021). The authors reported a significant increase in residential water consumption and a decrease in non-residential consumption. Overall, consumption in Tabriz dropped by approximately 10% between 2019 and 2020, suggesting that the increase in residential consumption due to people staying at home was not enough to offset decreases due to the reduction in economic activity. The use of geolocated water consumption data provided insights into the spatial distribution of various consumption patterns and when compared with water loss data, yields information that can be used to understand dynamic patterns of water use. The study noted that the pandemic intensified existing shortcomings in water supply and sanitation services.

Given Barbados’ heavy reliance on tourism, it is important to review studies that have looked specifically at the pandemic's impact on water consumption in the tourism sector. In Phuket, Thailand, a major tourism destination, changes in consumption patterns were reported by Changklom et al. (2022). The study found that after the Thailand government declared a State of Emergency in January 2020, tourist numbers decreased by approximately 99%, and both residential (−13%) and non-residential (−55%) water consumption decreased. While a significant decrease in non-residential consumption was unsurprising, mostly associated with the drop in tourism and the closing of hotels, the drop in residential consumption was curious. The research found two explanatory trends. First, there was a significant exodus from Thailand of workers employed in hotels and tourism. These workers would have contributed to residential water consumption. Second, residential properties that had been rented out to tourists through peer-to-peer accommodation services (such as Airbnb) became vacant, so when tourist numbers collapsed, there was a knock-on effect on apparent residential consumption (Changklom et al. 2022).

Changes in water consumption were noted by DeMaagd et al. (2021) in their study of the impact of COVID-19 on tourism in Oahu, Hawaii, where they looked at both hotel occupancy and Airbnb lettings. Their data indicated a considerable decrease in hotel water use during the first quarter of 2020, which coincided with a similar drop in tourist numbers and hotel occupancy rates; reservations with Airbnb also collapsed. Unlike in Phuket, however, residential consumption saw an overall increase, ascribed to the switch to work-from-home arrangements for residents of the island (DeMaagd et al. 2021). Analysis of consumption data indicated that for every 1% reduction in hotel occupancy, there was a corresponding 0.64% drop in water consumption. The fact that the impact of reduced hotel occupancy was not higher was attributed to most hotels not completely shutting down and there being a remaining ‘fixed’ amount of water consumption, for example, for pools and other amenities. Without this, the impact would have been greater. On the other hand, the authors hypothesised that the drop in Airbnb reservations during the COVID pandemic was offset by an increase in work-from-home arrangements for the island's residents. This then had a spill-over effect on water consumption: the drop in consumption associated with the loss of Airbnb reservations was offset by residents increasing their home water consumption due to spending more time at home. This suggests that local demographic and economic activity profiles have an important influence on water consumption patterns.

In 2020, the Inter-American Development Bank (IDB) surveyed 11 water utilities across the LAC region to understand how their operations had been affected by the pandemic. This section is based on a summary of the findings from the five Caribbean utilities that were included (Janson et al. 2021).

As discussed above, many jurisdictions experienced an increase in residential water consumption after the onset of COVID-19 and the introduction of government mitigation measures; this was also the expectation for the Caribbean. Three out of the five utilities, namely in Dominica, Jamaica and Suriname, experienced increased residential consumption for April–September 2020 as compared to similar periods pre-COVID-19, with Dominica experiencing the highest increase of 13% (Janson et al. 2021). The other two countries – Belize and Trinidad and Tobago – experienced a very slight decrease (Janson et al. 2021). In the case of Trinidad and Tobago, the apparent decrease has been ascribed to the extremely low level of water meter penetration as well as the user categorisation, making the estimation of residential water demand problematic. In the case of Belize, during the pandemic, some residents of the larger urban areas moved back to their families in the rural areas as employment opportunities were reduced. Overall, the Caribbean utilities did not experience significant changes in the volumes of water supplied, except for Suriname, which registered a 9% reduction. Interestingly, across the utilities, the total number of customers, residential and non-residential increased by up to 8% in the case of Dominica (Janson et al. 2021), though no reasons for this increase were advanced.

As a result of the increases in residential consumption and decreases in non-residential consumption, the share of residential consumption rose across all five countries, which was consistent with the findings of assessments conducted outside of the region. However, the redistribution of consumption did not translate into significant changes in supply volumes and, similarly, did not change the levels of Non-Revenue Water (NRW), which were reported as ranging from 75% in the case of Jamaica, 53% for Dominica, 50% for Trinidad and Tobago, 44% for Suriname and 22% for Belize (Janson et al. 2021). This is perhaps not surprising when the overall volume supplied did not change considerably.

The IDB only surveyed a limited number of Caribbean utilities at the start of the pandemic, so there are limitations in the findings based on the progression of the pandemic in the society and its effect of the level of economic activity on different sectors. However, the general pattern that emerged accords with that seen in other parts of the world and with the feedback provided during various on-line discussions organised by the Caribbean Water and Wastewater Association with water sector stakeholders to share experiences. It would not be prudent to associate all the observed changes in water consumption with COVID-19 as there may be other factors at play, suggesting a need for caution in drawing general conclusions and a need for further research. The IDB study also highlighted the difficulties in gathering utility-level data for detailed analysis.

Sowby (2020), in reviewing the preparedness of the U.S. water sector, suggested that the impact of the COVID-19 pandemic on water utilities has been both novel and profound. The pandemic has brought in new working arrangements, loss in revenues, and shifts in water consumption patterns for residential and non-residential customers, some of which may be permanent. Sowby (2020) also noted that many policy changes emerge as a response to crises through professional associations and research. One of the consequences noted by commentators (Poch et al. 2020) is that there is a growing realisation that we need to better understand consumption patterns and incorporate this knowledge into future infrastructure and operational planning. As Poch et al. (2020) observed, the fourth water revolution or the transition to digitilisation will play a significant role in enabling water utilities to respond better to future crises.

The existence of smart water meters and monitoring has allowed researchers to undertake fine-grained analyses of water consumption patterns across locations, sectors and time (Abu-Bakar et al. 2021). These analyses can be used by decision-makers for policy and strategy development in the water sector. In some respects, COVID-19 has been a catalyst in accelerating the trend towards the use of this data to develop new functionalities aimed at more rational and sustainable management of water systems (Poch et al. 2020). Changing consumption patterns are having an impact on operational aspects for which analysis of consumption patterns can be useful both in the short and long term. This is illustrated by a study from England (Abu-Bakar et al. 2021) which looked at changes in household consumption patterns using smart meter data and the clustering of economic and social variables. The study found increases in residential consumption of up to 35% as compared to pre-pandemic levels of consumption. More interestingly, however, the study was able to show shifts in daily peaks of water consumption to discern changes in patterns of use. Some of the changes included decreases in evening peak consumption and increases in late morning peak consumption and a sharp decrease in early morning peak consumption. These changes are related to people not having to get up and go to work by a certain time, so that consumption peaks are reduced and spread out over the morning period. Such changes would not have been discernable without data from smart meters and suggest that peak hourly pumping rates to balance supply and demand could be reduced. However, as Mishra et al. (2021) point out, such changes are unlikely to filter through low-income countries any time soon. We now turn our attention to the two chosen case studies, in which we review the available water consumption information to examine what the effects of COVID-19 have been and what insights these might provide.

On 11 March 2020, the WHO recognised COVID-19 as a global pandemic, and on 17 March, the first two cases were recorded in Barbados. From 28 March, the Government of Barbados imposed a series of restrictions to contain the spread of the disease. These measures included school closures, work-from-home directives, closure of businesses including hotels and lockdowns to restrict movements. At the same time, travel restrictions imposed by countries in Barbados’ source markets resulted in a collapse in tourism numbers, as shown in Figure 1.

In terms of water supply and distribution, 99% of the population in Barbados is connected to the water distribution system, which supplies water 24/7. The average daily water production has varied over the last decade by between 142,000 and 176,000 m³. Production is principally from groundwater resources, with desalination contributing approximately 20% of the total production. Water production is affected by periods of dry weather associated with El Niño conditions, limiting aquifer recharge, and has been experienced during 2009 and 2010, 2014 to 2016, and 2019 and 2020 (Trotman et al. 2020). During these periods, average production levels have been approximately 155,000 m³/day (2019 annual total 57.012 Mm³). Analysis of billed consumption volumes between 2002 and 2019 obtained from the Barbados Water Authority (BWA) indicates that 65–70% is classed as residential, 20% as commercial, 6% by government bodies, and 5% by hotels – actual water consumption by tourists will be higher as the 5% figure does not include rented holiday accommodation, which falls within the residential category.

There are approximately 115 tourist establishments with 10 rooms or more (Charara et al. 2011): 23 are on the West Coast and 73 are on the South Coast, of which 69 were active in 2019. Tourism establishments on the West Coast are generally of a higher star rating and with a smaller number of rooms, compared to those on the South Coast. Work by Charara et al. (2011) showed that hotels with higher ratings generally have higher water consumption per capita.

Bridgetown, in the southwest, is both the capital city of Barbados and the commercial and administrative centre of the country, while the South Coast has, as noted, most hotels and, by extension, tourism-related businesses. Both regions include high-density residential areas with up to 3.57 persons per household, compared to the national average of 2.86 persons per household (authors’ own calculations based on figures from the 2010 national population census obtained from the Barbados Statistical Service). All properties are metered, and meters are read manually on a monthly cycle. To assess the impact of the COVID-19 restrictions on water consumption patterns, we examined the billing records of customers in the Bridgetown and South Coast areas of Barbados.

To enable comparison between pre- and during COVID-19 consumption patterns, monthly water consumption data from 2017 to 2021 were made available by the BWA. The BWA provided the customer district, rate schedule, customer type, parish, service type and start date, and monthly consumption records for the two areas noted above. The data entries were checked to address meter readings and entry errors. Table 1 shows the relative change in water consumption by the sector between 12 months before (April 2019–March 2020) and 12 months after (April 2020–March 2021) the start of the pandemic, during which the imposition of health restrictions and the effective shutting down of the hotel and tourism sector took place.

Information in Table 1 reveals a dramatic decline in hotel water consumption during the pandemic when compared to the pre-pandemic period. The average monthly water consumption declined by approximately 56%. Interestingly, even during the period when hotels essentially had no guests, water consumption was approximately 44% of its pre-COVID-19 level. This is similar to the observation of decreases in hotel water consumption in Hawaii (DeMaagd et al. 2021). In other words, hotels were still using and being charged, even though they had reduced staff and no income. In contrast, average monthly residential water consumption increased by approximately 9% for the period assessed, which was consistent with the expectation, given that business closures, work-from-home directives and school closures meant that there were more people at home for more of the time. The average monthly consumption for both the government and commercial categories followed a similar trend to that shown by hotel consumption, with declines during the pandemic of approximately 22 and 6%, respectively.

The twin-island Republic of Trinidad and Tobago is the second case study country, with a population of around 1.5 million (World Population Review 2022). It is the largest oil and gas producer, and the most industrialised country in the Caribbean (GoRTT 2022). As a result, it has one of the highest GDP per capita in the LAC region of US$ 14,878 (World Bank 2021), although it had been falling steadily since 2015, prior to the pandemic, because of declining production rates. Extractive and petrochemical industries are substantial and place a high demand for more reliable and larger volumes of water. Other water-intensive sectors typically include agriculture; mining and quarrying, manufacturing; food, beverage, and tobacco production; textiles, clothing, leather, and wood; tourism – especially in Tobago; and aggregate industries.

Trinidad and Tobago is also vulnerable to annual hydroclimatic hazards, particularly in the rainy season, June to December (McShine et al. 2019), when flash flooding is a common occurrence (Roopnarine et al. 2018, 2021a). For the period 2019–2021, however, there were no hydroclimatic events that significantly affected the water sector.

Trinidad and Tobago has a mean annual rainfall of 2,200 mm with long-term annual renewable surface water estimated at 3,740: 3,600 million m³ in Trinidad and 140 million m³ in Tobago. The long-term average annual renewable groundwater resources are estimated at 614: 545 million m³ in Trinidad and 69 million m³ in Tobago. Although Trinidad and Tobago is significantly developed, with over 99% of the population having access to water – 89% with a piped supply and the other 11% via public standpipes (Ministry of Public Utilities (MPU) and Water and Sewerage Authority (WASA) 2022) – it is difficult to determine actual demand since residential water consumption is largely unmetered (only ∼3% of residential customers are metered). In contrast, WASA's industrial, commercial and cottage (businesses on residential properties) customers are relatively well metered, with 89, 59 and 25% of customers currently metered in these categories, respectively (MPU and WASA 2022).

A substantial proportion of WASA's customers receive an intermittent supply that varies because of both seasonal and operational issues that affect the ability to supply water 24/7. In any given reporting quarter, less than 40% of WASA's customers receive a consistent and reliable supply. The scheduled supply forces households to rely on household storage. Paradoxically, although the population and WASA struggle with an intermittent supply, the per capita consumption rate is 329 litres per capita per day (lpcpd) in Trinidad and 314 lpcpd for Tobago (Genivar 2009), which is one of the highest in the region. In comparison, the figure for Barbados is 220 lpcpd. This suggests that the intermittent water supply is a self-reinforcing system that adds pressure to the water supply network by increasing water demand so the network has difficulty in supplying its customers. Even though residential customers make up approximately 96% (MPU and WASA 2022) of WASA's customer base, industrial customers are the most significant contributors to revenue: making up only approximately 1% of WASA's customers, industrial customers’ payments contribute to approximately 33% of WASA's annual revenue.

Trinidad and Tobago's first confirmed case of COVID-19 was on 12 March 2020 (Office of the Prime Minister of Trinidad and Tobago 2020). To manage the inevitable outbreak, the government instituted border control and restrictions on economic activity, which severely affected all forms of economic activity. Initially, public health interventions comprised non-pharmaceutical preventive measures, including face masks, frequent washing of hands and physical distancing, as well as national lockdowns. As expected, the imposed restrictions strongly influenced the mobility of the population. Similar measures were employed globally as reported by Bakchan et al. (2022). Measures that revolved around strategies to promote social distancing have been noted to impact businesses (Nicola et al. 2020), the natural environment (Mostafa et al. 2021) as well as water demand (Cooley et al. 2020). During lockdowns and the enforced Work from Home (WFH), the Ministry of Health urged a general upgrade in personal hygiene and sanitation measures. Activities such as increased showering, washing of hands, cleaning, laundry and cooking potentially increased residential water consumption as was noted by Cooley et al. (2020). Decreases in non-residential consumption may have offset potential increases in residential consumption and demand. However, we were unable to support this assertion due to the limited metering of residential customers noted above.

With the phased opening of businesses for operation as safe zones and persons being required to return to work, hand washing facilities or stations (fixed or portable) were installed and equipped with soap and running water at most business establishments. The same applied to workplaces and job sites. Workers were required to wash their hands frequently when leaving their workstations for breaks; when using the bathroom; before and after eating, drinking, or using tobacco products; and after touching any surfaces suspected of being contaminated. Similarly, with the partial reopening of schools, strict sanitation measures were employed. These activities are likely to have increased the consumption of water in commercial and other non-residential sectors. At the same time, it is possible that residential consumption remained high as the population placed more emphasis on personal hygiene and sanitation, and a reluctance to return to normal activities thus increasing their water demand.

Industrial consumption activity stabilised during the period March to October 2020, coinciding with the earlier closure of certain activities in these sectors due to public health regulations. Like residential consumption, industrial consumption for the same period displayed a declining pattern with no discernable impact during the pandemic. However, unlike industrial consumption patterns, residential consumption patterns only provide a small snapshot of the entire customer grouping because, as noted above, only 3% of WASA's residential customers are metered. Industry consumption was less erratic during 2020 compared to 2019 and 2020 indicative of possible adjusted controls in demands and reductions in production and supply. The cottage category showed a smooth, slightly declining trend for water consumption, largely unaffected by the pandemic and other economic disruptions. This category accounted for the lowest consumption and seemed to display the highest buffering capacity.

Further analysis of data for the pre-COVID (April 2019–March 2020) and during COVID (April 2020–March 2021) periods revealed decreases in average monthly consumption across all consumer categories (Table 2). Average monthly residential consumption decreased by approximately 10%, which is a result of the utility classifying primary and secondary schools in this category. Since the closure of schools was part of the national public health restrictions, the decline in water consumption of schools is reflected in these residential consumption values. Noting that all customer categories showed unit reductions within the range of 3–13% (pre-COVID vs. during COVID), it is likely that decreased water consumption patterns were solely influenced by the pandemic and may reflect worsening socio-economic conditions.

In both Barbados and Trinidad and Tobago, there were contrasting impacts based on the consumer consumption data assessed. In the case of Barbados, discernible trends in consumption were evident. After the onset of the pandemic, in March 2020, there were decreases in both the tourism and commercial sectors, while there was an increase in residential consumption. These observations accord with findings reported in the growing body of literature reporting on case studies from across the globe on the impact of the pandemic on overall water consumption. However, deficiencies regarding water meter readings and entry errors complicate the analysis of consumption data and limit the potential management and operational benefit that could be derived from such information. By the way of illustration, in April 2021, Barbados experienced an ashfall from the eruption of the La Soufrière volcano on the nearby island of Saint Vincent which occurred on 9 April 2021. There is some evidence from the analysis of BWA's corresponding water consumption data that residential consumption may have increased by 25% due to the extensive cleaning by residents of properties, yards and vehicles by power washing. In fact, BWA issued appeals to the public to not use excessive volumes of water for clean-up (The Barbados Advocate 2021). However, metering challenges and the fact that there are more residential properties than other types of properties do not allow the impact to be fully appreciated across the different consumer categories. Similarly, it was not possible to detect if there was any impact on water consumption from the passage of Hurricane Elsa in July 2021. The inability to separate the impact of ‘out-of-the-ordinary’ or rapid onset events from background consumption behaviour, particularly those events that occur within the monthly water meter-reading cycle limits not only a water utility's ability to optimise the operation of water supply systems but also to prepare for the impact of hazard events.

In the case of Trinidad and Tobago, however, the impacts of the COVID-19 pandemic on the elements of the water sector examined were far less pronounced, indicating that either the pandemic had limited influence on the water consumption associated with the sectors examined – unlikely, or that the limited measurement of consumption prevented it from reflecting broader consumer consumption and patterns – more likely. However, there was a noticeable and sustained trend in 2020, relative to other years for the commercial sector, which coincided with the imposition of health restrictions and associated sanitation requirements. This trend reflects the increased need for water in sanitation among a particular category of commercial operations (groceries, pharmacies and hardwares) as these remained open for people to access their goods and services in the absence of Government-mandated closure of the other categories of commercial operations.

However, the key difference between the two countries is the prevalence of metering, notwithstanding the issues noted above, whereas BWA has universal water metering for all its customers, WASA has less than 5% of its total customer base metered, and in the residential customer grouping, only 3% of these customers are metered. Metering is a critical tool for data gathering to determine the demand and subsequent consumption patterns of the water utility's customers. Metering can support decision-making in servicing the water and sanitation needs of the population in fighting public health crises and overall socio-economic development. Furthermore, as is evident from the literature reviewed in section 2.5, smart metering and the ability to analyse data that it provides are becoming an increasingly important management tool for utilities. It is enabling the optimisation of the management of water supply systems and ensuring the adequate provision of water to all sections of society (Poch et al. 2020; Abu-Bakar et al. 2021; DeMaagd et al. 2021; Changklom et al. 2022). The relatively low penetration of metering in Trinidad and Tobago and the current absence of smart metering in Barbados hamper the potential to improve water management.

Adaptive and integrated water management will become increasingly important as the impacts of climate change become more apparent and intense, and future epidemics become increasingly likely (Daszak et al. 2020). The analysis of monthly readings only allows the identification of trends that become discernible over longer timescales. These monthly readings were suitable to analyse the impact of COVID-19 on the consumption of different customer categories in both countries, as the effects continued for several months. Monthly readings, however, do not allow a more fine-grained analysis over shorter time scales, which prevents the effects of rapid onset events such as hurricanes from being analysed. In this respect, the current COVID-19 pandemic should spur a more progressive approach to the use of technology and the opportunities it affords to record and analyse data and understand consumption behaviour.

An increased coverage of smart metering and sensor-based techniques to determine water consumption across sectors can significantly improve our understanding and provide greater insights into consumer trends and customer behaviour. Trinidad and Tobago's residential sector is significantly lacking meters; therefore, a proper understanding of demand is difficult. This will simultaneously make it difficult to meet the growing demands of the population without a fundamental understanding of how demand changes on a daily, seasonal or ad hoc basis and the complexities of the various sectors that require water and sanitation. In the case of Trinidad and Tobago, it was noted that residential consumption decreased slightly with the onset of the pandemic and resultant COVID-19 guidelines, which diverged from the expectation because schools are grouped in the residential category. The 5% metering coverage overall, the unstructured nature and misclassification of the existing customer base and the quarantine arrangements in public facilities for infected persons in the initial phases are other plausible contributory factors for this slight decrease. This highlights the need for metering and monitoring to better assess consumer behaviour.

More generally, the adoption of Information and Communication Technology (ICT) – of which smart metering is one aspect – combined with the analysis of socio-economic datasets has the potential to provide useful insights into consumption patterns across different customers and consumer categories. A holistic approach to understanding consumption patterns is favourable, as it will allow water utilities to understand all the drivers of water consumption. These drivers are likely to include behavioural patterns that are related to knowledge, attitudes and practices around water and sanitation, as well as the plumbing installations and configurations of water consumers. Analysis of these drivers will be important for the management and operation of water supply systems, enabling utilities to better plan for and respond to challenges such as climate change, disaster management and socio-economic growth that influence water demand. The Caribbean water utilities can benefit from a more holistic and proactive approach to water consumption.

The analyses of water consumption data from Barbados and Trinidad and Tobago indicate that the COVID-19 pandemic has had a discernible effect on patterns of water consumption across different consumer categories. Examining the full extent, however, has been limited by the extent of metering, particularly in the case of Trinidad and Tobago, and the quality and granularity of the available data, which is evident in both Barbados and Trinidad and Tobago. Looking ahead, to effectively address water demand, Caribbean water utilities need to recognise the value of data, seize the opportunity to make use of data and develop data analytics capacity – in-house or otherwise. The Caribbean water sector should capitalise on the age of widespread digital transformation and these techniques can be used to inform the operational management of natural resources and support financial sustainability. Reiterating what Poch et al. (2020) have observed, we need to understand consumption patterns better, to incorporate this knowledge into future infrastructure and operational planning, and to respond better to future crises. Such understandings are critical to the delivery of sustainable water services. Whether the BWA or the WASA could have managed their water systems better under COVID-19 if they had greater meter coverage and/or smart infrastructure metering is an open question. However, as has often been said, it is difficult to manage what is not properly measured. In this respect, better metering practices, preferably smart infrastructure metering and the ability to analyse it, would bring many benefits such as increased system resilience, better levels of service, optimisation of water system operation and the ability to address water losses and reduce the stress on water resources. This study illustrates the challenges in understanding water consumption and suggests that there are opportunities to improve. Although the study is limited to Barbados and Trinidad and Tobago, the work has relevance for other Caribbean Small Island Developing States.

The authors would like to express thanks to the Barbados Water Authority (BWA), Barbados, in particular the General Manager, Mr Keithroy Halliday, and Mr Charles Leslie, Manager Engineering. Our gratitude also extends to the staff of the Ministry of Public Utilities (MPU), the Water and Sewerage Authority (WASA) and the Water Resources Agency (WRA), Trinidad and Tobago, for provision of raw data utilised in the analysis. We also wish to acknowledge the contribution of Dr Sara Humphreys in refining the contents of this article.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.