Solutions

Water accounting in irrigation canals

How to follow every cubic metre from the head regulator to the field, what the standard frameworks ask for, and why the measurements decide the answer.

At the end of the season a canal manager is asked a simple question: how much water went into the system, where did it go, and who got it? The records usually show gate openings, a few staff-gauge readings and the area irrigated. They rarely add up to an answer. Water accounting, the practice of tracking water supply, use and loss in a defined area over a defined period, is how that question gets answered with numbers instead of opinions. This article explains the water accounting framework used internationally, how it applies to a canal command, and what has to be measured for the accounts to be worth trusting.

What is water accounting?

Water accounting is, in the Food and Agriculture Organization (FAO) definition, "the systematic study of the current status and trends in water supply, demand, accessibility and use in domains that have been specified" [1]. The "domain" can be a field, a canal system or a whole river basin. FAO's own page puts it plainly: it is not possible to plan or manage water resources without first measuring and understanding them [2].

Two points follow. First, accounting is about quantities over a period: volumes, not only flow rates at one instant. Second, it needs a boundary. You cannot account for "the canal" until you say where the system starts (usually the head regulator), where it ends (the tail escape and the last outlets) and which season or year you are covering.

What is the difference between water accounting and water auditing?

Water accounting produces the numbers; water auditing places them in a wider context. In the FAO sourcebook, auditing "places the findings, outputs and recommendations of water accounting into a broader framework comprising governance, institutions, public and private expenditure, legislation, services delivery and the wider political economy" of the domain [1].

In a canal system this distinction is practical. Accounting might tell you, for example, that 40 % of the water released at the head never reaches an outlet. Auditing asks why: is it seepage through unlined reaches, unauthorised offtakes, an outlet that is larger than designed, or a rotation schedule nobody follows? The Central Water Commission (CWC) and Central Ground Water Board have issued General Guidelines for Water Audit and Water Conservation, and CWC's 2014 guidelines on improving water use efficiency recommend that state governments introduce water audit as a regular activity in irrigation projects [3]. In the same document CWC recommends charging for irrigation water on a volumetric basis [3], which is not possible without measured volumes.

How is water use classified in a canal system?

The framework that most water accountants follow comes from the International Water Management Institute (IWMI). Molden's 1997 paper proposed standard terms and indicators at three scales: a single use such as a field, a service such as an irrigation system, and a river basin [4]. FAO's sourcebook carries the key idea forward as water use fractions [1]:

FractionMeaningCanal example
Beneficial consumptionWater evaporated through crop transpiration or captured in biomassTranspiration by the wheat crop
Non-beneficial consumptionWater that evaporates without contributing to biomassEvaporation from bare soil, weeds, canal water surface
Recoverable non-consumed fractionWater that returns to a source that can be used againDeep percolation to a usable aquifer; drainage reaching a downstream user
Non-recoverable non-consumed fractionWater lost to further useFlow to a saline sink or a heavily polluted aquifer, or to the sea

The categories are from FAO's Box 3.11 [1]. The important lesson is that "lost" is not one thing. Seepage from a canal that recharges a good aquifer used by farmers downstream is not lost to the basin, even though it is lost to the canal. The classical canal efficiency figures do not make this distinction, which is why Perry and colleagues argue that debates about irrigation "waste" often misidentify how the water is split between evaporation, transpiration and recoverable return flows, and that the consumed water saved by advanced technology is limited [5].

How does this relate to the efficiency figures engineers use?

Indian planners use component efficiencies. CWC's 2014 guidelines list typical conveyance efficiencies of 55–60 % for unlined canals and 70–75 % for lined canals, field application efficiencies of 65 % for flood irrigation and 80 % for furrow, and an overall efficiency of 30–65 % for surface-water systems [3]. CWC's terms of reference for studies of completed major and medium projects ask for five efficiencies: reservoir or diversion, conveyance, on-farm application, drainage and overall project efficiency, each compared with the value assumed when the project report was prepared [6].

These are ratios of volumes. Conveyance efficiency is the water delivered at the outlets, divided by the water let into the canal. To compute it, someone has to have measured both. Water accounting is the discipline of doing that measurement consistently, and then splitting the loss into the fractions above.

What does a canal water balance look like?

The core of canal accounting is a water balance: inflow equals outflow plus losses plus change in storage. For a canal command over a season:

Inflow at head  =  Deliveries at outlets
                 + Tail escape (water leaving the end of the canal)
                 + Losses (seepage, evaporation, leakage, unauthorised offtake)
                 + Change in water stored in the canal

Losses = Inflow − Deliveries − Tail escape − Change in storage

Notice that losses are normally found by difference. Nobody measures seepage along 40 km of canal directly. This matters, as the example shows.

Worked example: a distributary over one season (illustrative)

Take a distributary that runs for 120 days. The mean discharge at its head, from a continuous record, is 4.0 m³/s. One cubic metre per second flowing for one day is 86,400 m³, or 0.0864 million cubic metres (Mm³).

Inflow        = 4.0 m³/s × 120 days × 0.0864 Mm³ per (m³/s·day) = 41.5 Mm³
Deliveries    = sum of metered outlet volumes                    = 24.9 Mm³
Tail escape   = measured at the tail                             =  5.0 Mm³
Storage change over the season (same level at start and end)     =  0

Losses        = 41.5 − 24.9 − 5.0 − 0                            = 11.6 Mm³  (28 %)
Conveyance efficiency = 24.9 / 41.5                              = 60 %
Where the inflow went (illustrative distributary, season total)
012.52537.550Inflow at headOutlet delive…Tail escapeLosses by dif…Mm³
Illustrative water balance for a distributary over a 120-day season. The loss term is not measured; it is what remains after deliveries and tail escape are subtracted from the inflow.

Now include measurement error. Suppose the head discharge is known to within 5 %, outlet deliveries to within 5 % and the tail escape to within 5 %. If the errors are independent, they combine in quadrature:

Head:      5 % of 41.5 = 2.07 Mm³
Deliveries: 5 % of 24.9 = 1.25 Mm³
Tail:       5 % of 5.0  = 0.25 Mm³

Uncertainty of losses = √(2.07² + 1.25² + 0.25²) = √5.90 = 2.4 Mm³

The loss of 11.6 Mm³ carries an uncertainty of about 2.4 Mm³, roughly ±21 % of the loss itself, even though every flow was measured to ±5 %. If the head measurement were 10 % uncertain, the loss would be uncertain by more than ±4 Mm³, over a third of its value. Small errors in the large terms become large errors in the small term. That is why the inflow at the head, which is the biggest number in the account, deserves the best instrument in the system. The general idea of propagating uncertainty is covered in measurement uncertainty for water data in plain language.

What measurements does canal water accounting need?

Water accounting stands or falls on a few measurement points, placed so that every flow in the balance is either measured or can be reasoned out.

  • Inflow at the head regulator. Continuous discharge, not occasional readings. Gate settings alone are weak evidence, because the flow through a gate depends on the upstream and downstream levels as well as the opening. Weirs, flumes and radar-based methods are compared in how to measure canal discharge.
  • Branch and distributary heads. These split the account into reaches, so a loss can be located to a stretch of canal rather than "somewhere on the system". Metering zone by zone is also the usual way to find where water is being wasted.
  • Outlets and escapes. Volumes delivered to farmers, and what leaves at the tail.
  • Rainfall and evaporation. Needed to separate canal evaporation and rainfall contributions from deliveries.
  • Calibration gaugings. Every rating relation needs periodic checks by independent gauging, as explained in the velocity-area method and rating curves and their uncertainty.
  • Canal cross-sections. The flow area at a radar or level station comes from a surveyed section, which must be re-surveyed because silt and weeds change it; see river cross-section surveys.

Water levels and discharge should then be checked for spikes, gaps and drift before they go into a seasonal total. A single stuck sensor over ten days can shift a season's volume; see data quality control for hydromet time-series.

How is water accounting done at larger scales?

Canal accounting looks inside the command area. Basin-scale accounting asks how much water the whole system depletes. The Water Accounting Plus (WA+) framework is one such approach. It reports resource base, evapotranspiration, productivity and withdrawal, and it can use satellite measurements, although the authors state that satellite data is not a precondition for applying it [7]. The same paper identifies the non-availability of flow data as one of the underlying reasons that operational basin water accounting systems are not in place [7]. For a canal manager, this is a direct argument for installing flow measurement: without ground measurements at the heads and outlets, neither the canal account nor the basin account can be closed.

From one season to a routine

A one-off audit is useful. A routine is better. The steps that make accounting routine, rather than an annual scramble, are modest:

  1. Define the domain and the period. A distributary, a season, the water year.
  2. Write the balance for it, as above, with every inflow and outflow named.
  3. Put a measurement on every large term, starting with the head.
  4. Compute the totals on a fixed schedule, such as weekly during the season, and compare reaches.
  5. Investigate the reach with the largest unexplained loss, which is where accounting turns into audit.
  6. Report the uncertainty with each figure, so a loss of 11.6 Mm³ is not mistaken for one of 11.6 plus or minus 4.

CWC's own studies compare estimated efficiencies against those assumed in the project report [6]. A system that keeps its own accounts every season can make the same comparison without waiting for a study.

Key points

  • Water accounting tracks supply, use and loss in a defined domain and period; auditing then explains the results in their governance and management context [1].
  • Losses in a canal balance are found by difference, so the uncertainty of the big flows controls the uncertainty of the loss, as the worked example shows.
  • "Loss" must be split into consumed and non-consumed, recoverable and non-recoverable fractions before it is called waste [1][5].
  • Measurement at the head, at branch heads and at outlets, with periodic calibration gaugings, is the foundation of every canal account.
  • CWC recommends water audit and volumetric charging for irrigation projects [3]; both depend on measured volumes.

Where ClimMaTech fits

At ClimMaTech we work on the measurement side of canal accounting: continuous level and flow records at the points where a water balance needs them. PRAVAAH combines an AWLR (Automatic Water Level Recorder) with an SVR (Surface Velocity Radar) in one non-contact unit, and it records on board so that data are kept when the network is down. Our canal irrigation monitoring page describes how readings at regulators, branches and tail reaches fit together into a canal's account.

References

  1. C. Batchelor, J. Hoogeveen, J.-M. Faurès, L. Peiser, Water accounting and auditing: a sourcebook, FAO Water Reports 43, Food and Agriculture Organization of the United Nations, Rome, 2016 (definitions in Boxes 1.1 and 3.11 and section 1.4). openknowledge.fao.org
  2. FAO, Water accounting (Land and Water Division web page). fao.org
  3. Central Water Commission, Ministry of Water Resources, Guidelines for Improving Water Use Efficiency in Irrigation, Domestic & Industrial Sectors, November 2014. cwc.gov.in
  4. D. Molden, Accounting for water use and productivity, SWIM Paper 1, International Irrigation Management Institute, Colombo, 1997. doi:10.3910/2009.366
  5. C. Perry, P. Steduto, R. G. Allen, C. M. Burt, Increasing productivity in irrigated agriculture: agronomic constraints and hydrological realities, Agricultural Water Management 96(11), 1517–1524, 2009. ideas.repec.org
  6. Central Water Commission, Terms of reference for conducting water use efficiency studies of completed major/medium irrigation projects, 2014. cwc.gov.in
  7. P. Karimi, W. G. M. Bastiaanssen, D. Molden, Water Accounting Plus (WA+): a water accounting procedure for complex river basins based on satellite measurements, Hydrology and Earth System Sciences 17, 2459–2472, 2013. doi:10.5194/hess-17-2459-2013
water accountingirrigationcanalwater auditconveyance efficiencywater balancedischarge measurementcommand area
Dharkan Anand
Written by
Co-Founder & CEO