Accounting for water: A global review and indicators of best practice for improved water governance

While water accounting can provide integrated hydrological and economic information to support water governance, its widespread use in decision-making remains, like environmental accounting, limited in scope and by country (Vardon et al., 2016a, Vardon et al., 2016b). This is despite the fact that multiple opportunities have been identified to use water accounting for water management (Meijer et al., 2020), assessing land management trade-offs (Keith et al., 2017), full cost recovery water pricing (Romero et al., 2017), and wildlife management (Vardon et al., 2017).

Our goal is to show how water accounting has developed over time and how it could be improved to better inform water governance. Section 2 outlines the main features of water accounting, Section 3 describes current practices (e.g., number and types and accounts produced, institutional arrangements, data sources and methods used for account production), and Section 4 includes our key results and findings. Key insights are highlighted in Section 5, while Section 6 details the learnings from current practices and water accounting best practices for improved water governance.

Physical water asset accounts show stocks of water contained in the inland water resources by asset class (groundwater, soil water, various forms of surface water) and how they change from one period to the next. The physical water asset account is closely aligned with a water balance (Vardon et al., 2012). For theoretical completeness, the water in rivers is included in the asset account, which is the volume at a point in time, for example, the volume calculated by the length, width, and depth of rivers on 1 January. Some water resources are irrelevant to some areas, for example, those without snow, ice, glaciers, or groundwater, or may be considered of little utility, hence not included. This is in line with the “...flexible and modular approach to implementation within national statistical systems which can be aligned with the particular policy context, data availability and statistical capacity of countries” (UN et al., 2014, pp. viii-ix).

Water quality or condition accounts show water by different quality classes for each type of water resource or show water quality indicators by type of water resource or individual water resources. For example, the Australian Bureau of Statistics (ABS) and Australian Bureau of Meteorology (BoM) (2019) water quality account uses a range of indicators, including pH, conductivity, dissolved oxygen, P and N levels, for three rivers, four lakes and three catchments. SEEA Ecosystem Accounting also describes ecosystem condition indicators, with water quality being an indicator of overall ecosystem condition (UN et al., 2021).

Water budgets are a tool for quantifying water flows into and out of a hydrological system. They record all water stored and exchanged on the land surface (rivers, lakes), subsurface (aquifer, groundwater), and atmosphere (precipitation, evaporation) (e.g., Healy et al., 2007). A water budget resembles a physical asset account (Vardon et al., 2012).

Ecosystem Accounting covers aquatic ecosystems, several water-related ecosystem services, and water quality indicators that contribute to the assessment of ecosystem condition (UN et al., 2021). The Ecosystem Type Reference Classification used in SEEA EA is based on the IUCN Global Ecosystem Topology(GET). Ecosystem types related to water include artificial reservoirs³, lakes, rivers and streams, snow, ice and glaciers and groundwater as well as other water-related ecosystems, such as wetlands. Water-related ecosystem services, include water supply; water purification (retention and breakdown of nutrients and other pollutants, e.g., as provided by the vegetation in water catchments); water flow regulation; river flood mitigation services (this is the riparian vegetation which provides structure and a physical barrier to high water levels). Nursery population habitat (e.g., for harvested fish); recreation-related services (e.g., canoeing on a river); visual amenity (e.g., views of a river or snow, ice, and glaciers); spiritual, artistic, and symbolic services (e.g., the Ganga or Ganges River). SEEA Ecosystem Accounting provides more details on the concepts and structure of the accounts (UN et al., 2021).

Water accounts use many data sources and methods. A detailed examination of the data sources and methods used, however, is beyond the scope of this global review. The International Recommendations for Water Statistics (UN, 2012b) provide further details, and since its publication there have been many advances. For example, in the use of remote sensing technologies, hydrological modeling (e.g., Karimi et al., 2013; Pedro-Monzonís et al., 2016; Esen and Hein, 2020; Kind et al., 2020), and an increase in computer power that has resulted in the refinement and development of a range of tools for collecting and analyzing various types of data needed for water accounting (e.g., WA+⁴).

The are many actual or potential uses of water accounts. To be used, most decision-makers need to know how water accounting is relevant to their needs. Indicators that can be derived from the water accounts play a key role here. This starts with a mapping of their information needs and can be done for the four general water management and governance objectives (Table 1). Example of accounts application to specific topics include water security (Mahdavi et al., 2019), groundwater management (Mazzanti et al., 2014), managing trade-offs in catchments used for water supply (Keith et al., 2017).

Accounts were produced by a range of government agencies, international organizations, and academic journals, using different accounting frameworks, multiple types of accounts, and at regional and national levels (Fig. 6). Thirty-two (23 %) of accounts were produced by two or more agencies and in two cases, the Dominican Republic (CIDECA, 2016) and Colombia (WAVES, 2016), by four types of agencies (Supplementary Information Table S.6). In the cases of joint production, the water account was included in the producer category of each agency type, hence the number of account producers (Fig. 6a) is more than the number of accounts (n = 139). The most common producers of water accounts were academic organizations (n = 79), followed by national statistical offices (n = 43) (Fig. 6a). Twenty-five water accounts were produced by international agencies, for example, the World Bank (e.g., the WAVES¹¹ and GPS¹² programs), the UN (e.g. NCAVES¹³), the FAO, and the Asian Development Bank (ADB). International organizations played multiple roles, supporting accounting production through technical assistance or resourcing. The SEEA was the most used water accounting framework (n = 73 or 53 %) and the most used by national statistical offices (n = 33) and academic organizations (n = 36) (Supplementary Information Table S3). The next most used framework was WA+ (n = 35) (Fig. 6b).

Most water accounts produced were at the regional level based on physical boundaries, river basins, or other hydrologically defined areas (n = 86) (Fig. 6c). The spatial output varied according to the account producer (Supplementary Information Table S.7). Academic institutions mainly produced accounts for physical regions (n = 61) and statistics offices for the national level (n = 38). Twenty-one accounts had more than one spatial output. Many different types of accounts were produced (Fig. 6d), and spatial outputs in the accounts varied (Fig. 6c). Of the 271 accounting tables identified, physical supply and use tables were the most common type of account produced (n = 74), closely followed by physical asset accounts (n = 50) (Fig. 6d). Eleven accounts reported water use but not supply. Thirty accounts were for the water resources base, which were mostly water balances that align closely with physical asset accounts (Vardon et al., 2012). The number of accounts with the ecosystem service of water provision (n = 24), water filtration services (n = 16), and other water-related ecosystem services (n = 11), is not an accurate reflection of the number of accounts produced, including these services as the salient word “water” is absent from the title, abstract, or keywords of many articles that produce accounts for multiple ecosystem services. Thirteen countries produced monetary supply and use tables.

The frequency and time series length varied (Fig. 7). The overwhelming number of accounts produced by countries were “one-offs” (n = 37), with just 15 countries having annual water accounts (Fig. 7a). National statistical offices produced 14 of 23 annual accounts, while academic organizations 74 of 138 one-off accounts (Supplementary Information Table S.8). The time series length ranged from 1 to >10 years, and 25 countries had timeseries>10 years (Fig. 7b).

Historically, most accounts were published as printed documents or made available via pdf files (n = 121). The use of PDF files is common for low- and middle-income countries that produce water accounts, like Botswana. In contrast, high-income countries like Australia and the Netherlands have dispensed with printed and PDF documents, and the text is now entirely on webpages with downloadable tables (n = 18) (Fig. 7d).

In most countries, water accounting is in an experimental phase and has yet to be embedded in the decision-making processes of water governance. This is because, one, water accounts take resources and time to produce and rely on existing water data which usually has gaps and deficiencies (Section 5.2), and, two, decision-makers have little awareness or understanding of water accounts and what they offer (Section 5.3). With resources limited and understanding lacking, there can be competition rather than the collaboration between the different agencies needed for the ongoing production of water accounts. This can also be exacerbated by misunderstandings on technical matters between disciplines (e.g., hydrologists, economists, and accountants) (Vardon et al., 2018; World Bank, 2021).

We highlight the processes to help overcome the impediments to production of water accounts (Ruijs et al., 2019) and present three modes of water account production, each of which has its strengths and weaknesses:

• 1.

  Production led by government statistical agencies. This generally ensures close linking with other SEEA-based accounts, the SNA and associated economic data, with a strong emphasis on integration and a focus on ongoing account production. However, accounts can take several years to produce and require substantial capacity building and technical inputs and data from a range of water and research institutions, and statistical agencies are deliberately separated from account users¹⁵.

• 2.

  Production led by government agencies concerned with water, natural resources or environmental management and policy. In this approach, the accounts produced suit the needs of a particular agency. This ensures close linkages to policy analysis and resource management but can also mean that water accounts do not easily integrate into other types of environmental or economic information (e.g., other SEEA or SNA accounts).

• 3.

  Production led by research agencies (e.g., universities). Here, the accounts are generally produced relatively quickly and use the latest knowledge and information. Interpretation and analysis are prominent, but there is no view to ongoing production unless there is on-going funding and linkages to government agencies.

The United States of America (USA) released a strategic plan for environmental-economic accounting (US Department of Commerce, 2023). The strategy had a phased implementation plan with water accounts as part of the first phase. The USA's plan reflects the five factors in Table 2.

Water accounts are highly-technical exercises that rely on collaboration between different agencies, the knowledge and expertise of several disciplines, and data (1 Introduction, 2 Water accounting). Many countries lack coordination processes, with multiple accounts being produced within countries by different agencies using different accounting systems (Section 4.1).

A common barrier to account production is data availability and data quality (Vardon et al., 2012). Most countries have some of the data needed for the production of water accounts but no country has access to all the data needed to produce the full suite of water accounts: physical and monetary supply-use tables, asset accounts, environmental protection expenditure related to water, water quality, and emissions to water. Consequently, those producing accounts rely on various estimation methods to populate different parts of the water accounts. In some cases, data may exist, but the agency or agencies producing the accounts may need help accessing the data for legal, administrative, or technical reasons (Vardon et al., 2018). This, in turn, results in few types of water accounts being produced (Fig. 6), irregular production (Fig. 7), under-coverage of water resources, and low levels of industry detail (Fig. 8).

Salminen et al. (2018) provide a detailed analysis of data quality issues in Finland. They identify two potential sources of error in Finland's water use estimates. The first is poor data coverage of some industries, resulting in unreliable estimates of the water supply and use in these industries. The second source of error is allocating the amounts of water used by industries to different water sources (e.g., self-abstracted vs. mains water or between industries). The total water supply by the water supply industry appears to be reliable but the water use by industries and households is less so, and in some cases is unknown.

Weckström et al. (2020) identified the lack of industry detail as a limitation of accounts in the analysis of the economic importance of water. Their review highlighted the lack of industry detail (Fig. 8) but this limitation could be overcome with additional data. For example, water suppliers typically have registers of customers, which could be coded to industry or sector, as has been done in Botswana by the Water Utilities Corporation (DWS, 2021). Weckström et al. (2020) also note that water accounts generally report water consumption based on incomplete data, so there is a possibility that account users may incorrectly interpret the accounts.

Terminology is another challenge. Terms like “water consumption” are defined slightly differently in SEEA than in other frameworks. For example, for water footprints, Hoekstra et al. (2011) have a pure conceptual definition for water consumption: “The volume of freshwater used and then evaporated or incorporated into a product. It also includes water abstracted from surface or groundwater in a catchment and returned to another catchment or the sea. It is important to distinguish the term ‘water consumption’ from the term ‘water withdrawal’ or ‘water abstraction.” By contrast, the SEEA defines water consumption as the difference between supply and use and adds in practical considerations as examples: “.. apparent losses due to illegal tapping as well as malfunctioning metering.” Confounding the problem of interpreting water consumption is that some accounts overestimate consumption due to an underestimate of water return flows (e.g. from agricultural land).

A process of ongoing development, production and use of water accounts is, arguably, the most essential part of best practice. Engagement between account producers and users has previously been identified as critical to the effective use of environmental accounts (Ruijs et al., 2019; World Bank, 2021). Further, water accounting needs to evolve to incorporate new knowledge (data, methodologies, understanding, etc.) and perspectives which requires continuous engagement across stakeholders.

One important gap, currently overlooked in SEEA's development, is that of First Nations. We endorse the recommendation of Normyle et al. (2024) to establish a working group under the auspices of the UN for this purpose. How First Nations' values can be included in environmental accounts and how First Nations might be able to use environmental accounting are two questions that have received little attention in SEEA-based accounting (Normyle et al., 2022). These questions are equally valid for water accounts as water, in its many forms, connects to places, practices and values of First Nations' (Manero et al., 2022).

Expanding the relevance of water accounts means increasing the frequency, timeliness and detail of the accounts. Increasing the number of industries in accounts should increase their use in economic modeling (e.g., Wittwer, 2012), while adding information on physical flows allows some indicators, like water scarcity and water consumption, to be more readily derived from water accounts. As water availability is likely to become more variable with climate change (e.g., Konapala et al., 2020), sub-annual accounts and faster production of accounts could make accounts more relevant to managing short-term water demand and balancing the needs of households and industry (e.g., Grafton et al., 2020).

Accuracy is related to all the physical and economic data and methods used to populate the accounts (see Section 2.4). Accuracy will vary and errors may arise from many sources including in measurement, sampling bias, model assumptions, and processing errors (human and software). All data sources and methods used in accounts should be described and be publicly available. Quantitative (e.g., standard errors) or qualitative (e.g., “traffic lights”) measures of accuracy should also be included as best practice. Over time, anomalies between different data source can be resolved, data gaps filled, and the accuracy of accounts improved (Vardon et al., 2018).