BRENNEN

Industries / Food, farming and forestry / Agriculture and farming / Irrigation

Saving water

Cutting the water farms take, and checking whether water saved on one field is really saved for the river, the aquifer or other users.

7

Milestones achieved, from 1983 to 2024.

2

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1983

    Drain water poisons birds

    Irrigation leaves salts in the soil, so farmers drain water away through pipes and ditches. Kesterson Reservoir in California, 12 evaporation ponds for irrigation drainage that the US Bureau of Reclamation completed in 1971 inside the San Luis National Wildlife Refuge, collected that water. In 1983 scientists found dead and deformed birds there: naturally occurring selenium had built up in the drain water, at 85 to 440 parts per billion, and passed up the food chain. California's State Water Resources Control Board ordered the reservoir closed in 1985, and work on the San Luis Drain, planned to run 290 miles to the ocean, stopped in 1986 with 85 miles built.

    LessonWater that drains off irrigated fields has to go somewhere, and it carries what it picks up. 90104-x).

    watereducation.orgdoi.org

    Credit

    US Bureau of Reclamation (responsible); Harry Ohlendorf and colleagues documented the harm to bird embryos

  2. 2003

    Farm water for San Diego

    Under California's Quantification Settlement Agreement, the Imperial Irrigation District agreed to transfer as much as 200,000 acre-feet (about 247 million m³) of Colorado River water a year to the San Diego County Water Authority for up to 75 years, which the Water Education Foundation calls the largest farm-to-city water transfer in the United States. The district had to create the water by conserving it; an amendment allowed fallowing through the fifteenth year, after which the two agencies were to negotiate a shift to efficiency-based conservation.

    LessonLess runoff from Imperial Valley fields sped the shrinking and salting of the Salton Sea; the water the two agencies sent to the sea to make up for it stopped when that part of the deal expired in 2018.

    watereducation.orgsdcwa.org

    Credit

    Imperial Irrigation District and San Diego County Water Authority, with the Coachella Valley Water District, the Metropolitan Water District of Southern California and the State of California

  3. 2009

    Punjab delays rice planting

    India's Punjab state passed the Punjab Preservation of Sub Soil Water Act, 2009, which bans raising rice nurseries before 10 May and transplanting before 10 June, to slow the fall of groundwater pumped for rice. Studies disagree on what it saved: Tripathi, Mishra and Verma (2016), using data from 1985 to 2011, found the water table rising 0.2 cm a year after the Act, while Sharma and colleagues (2021) found groundwater kept falling, faster in the main rice districts.

    LessonLater planting also moved the rice harvest, and the burning of its stubble, later into autumn; Balwinder-Singh and colleagues (Nature Sustainability, 2019) found a 39% higher peak fire intensity at a time of year when weather favours poor air quality.

    ageconsearch.umn.edunature.com

    Credit

    Government of Punjab

  4. 2015

    Rice paddies dry between floods

    In alternate wetting and drying (AWD), farmers let a flooded rice field drain down and re-flood it only when the water in a perforated "field water tube" sunk in the soil falls below a mark, instead of keeping it flooded. The International Rice Research Institute (IRRI) and national partners developed and promoted it; their 2015 review reported large-scale adoption in the Philippines, Vietnam and Bangladesh, and irrigation cut by up to 38% with no yield loss when done correctly.

    LessonTwo randomized trials in Bangladesh found that AWD saved water only where farmers paid for each unit of water, not where they paid a fixed charge for the season (Chakravorty, Dar and Emerick, 2023).

    doi.orgdoi.org

    Credit

    IRRI (Rubenito Lampayan, Bas Bouman and colleagues) and national research partners

  5. 2018

    Efficiency rarely saves basin water

    In Science, R. Quentin Grafton and ten co-authors set out the "paradox of irrigation efficiency". Drip and sprinkler upgrades cut the water a farm applies, but much of the water once "lost" to runoff and seepage was being reused downstream or from aquifers, and farms often use the saving to irrigate more land or thirstier crops, so "water saved at a farm scale typically does not reduce water consumption at a watershed or basin scale". They noted that Spain, Morocco and Australia's Murray–Darling Basin had each spent several billion US dollars subsidizing irrigation efficiency.

    cgspace.cgiar.org

    Credit

    R. Q. Grafton, J. Williams, C. J. Perry, F. Molle, C. Ringler, P. Steduto, B. Udall, S. A. Wheeler, Y. Wang, D. Garrick and R. G. Allen

  6. 2020

    Upgraded farms used more water

    Australian government programs pay Murray–Darling irrigators to upgrade farm infrastructure in return for part of their water rights. Using program records and farm surveys from 2009–10 to 2016–17, mainly on the On-Farm Irrigation Efficiency Program, Neal Hughes and colleagues found higher productivity and profit for participants, but farm water demand "significantly higher post-upgrade", which they call a "Jevon's paradox outcome". A 2026 model by Alec Zuo, Sarah Wheeler and Quentin Grafton puts government spending on on- and off-farm upgrades at about AUD 5 billion by 2023 and estimates that the subsidies raised southern basin irrigation extractions by up to 24% from 2015–16 to 2020–21.

    LessonBuying back water rights with efficiency grants can still leave farms taking more water.

    doi.orgdoi.org

    Credit

    Neal Hughes, Manannan Donoghoe and Linden Whittle (2020); Alec Zuo, Sarah Ann Wheeler and R. Quentin Grafton (2026)

  7. 2024

    Rice water credits voided

    Carbon credits for rice are earned by cutting the time paddies stay flooded, which reduces methane and usually water. In August 2024 Verra, which runs a large carbon-credit standard, began invalidating 37 rice projects that held 4.5 million credits under a methodology for adjusted water management, 99.9% of all rice credits issued. Verra cited weak proof that the change would not have happened anyway, projects classed as small-scale that were likely large, overstated areas, and too little evidence that the practice changed; it sanctioned the four bodies that had audited them.

    LessonPaying farmers to change how they water needs proof, field by field, that the change happened.

    rmi.org

    Credit

    Verra (review); summarized by RMI

Watch for

  1. Paid water savings in Chile

    Kilimo, a company that sells irrigation advice, says a three-year project with 11 farms in Chile's Maipo basin, run with Microsoft, saved 60 million cubic feet (about 1.7 million m³) of water, and a separate drip project 14 million cubic feet (about 0.4 million m³). It measures savings against each farm's average water use over the previous three years, adjusted for rainfall, and pays farmers once savings are verified.

    Achieved whenan outside party shows the saving in water consumed by the crops (evapotranspiration from satellites or flux towers) or in river flow, not only in water applied.

    news.microsoft.com

    Claimed by

    Kilimo, in a Microsoft feature, December 2025

  2. Rice drying credits in Pakistan

    NetZeroAg's alternate-wetting-and-drying project with 3,000 smallholders on 32,556 ha in Pakistan's Punjab has had 46,714 carbon credits issued under the Gold Standard, and reports irrigation water use "down by up to 30%" and yields up by up to 12%.

    Achieved whenmeasured water use on project fields, compared with similar fields outside the project, is published by someone other than the developer.

    goldstandard.org

    Claimed by

    NetZeroAg, in a Gold Standard project spotlight, October 2026

Open

  1. 01

    Audit water savings from orbit

    An independent check of irrigation projects that companies count as water "replenishment". For each project, compare the volume claimed with the change in water the crops actually consumed (evapotranspiration), estimated from satellite data on the enrolled fields before and after, against similar fields nearby.

    The brief
    Credit · unclaimed
  2. 02

    Log water in credited paddies

    A one-season record of the water actually applied to rice fields in an alternate-wetting-and-drying carbon project and to similar fields outside it. Each field gets a water-level logger in its field water tube, so every irrigation and every dry spell is recorded, plus a one-time flow test of the pump or channel that feeds it.

    The brief
    Credit · unclaimed

Open 01 · the brief

Audit water savings from orbit

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An independent check of irrigation projects that companies count as water "replenishment". For each project, compare the volume claimed with the change in water the crops actually consumed (evapotranspiration), estimated from satellite data on the enrolled fields before and after, against similar fields nearby.

Why it is still open

The World Resources Institute's Volumetric Water Benefit Accounting 2.0 guidance (September 2025) lets irrigation efficiency projects claim either reduced consumption or reduced withdrawal, depending on local context, and Kilimo measures savings against water applied. No published comparison of claimed volumes with satellite-measured consumption was found.

files.wri.orgnews.microsoft.com

Built with what exists

The OpenET API for US fields and FAO's WaPOR data elsewhere; Google Earth Engine; Python; an AI language model to pull project names, places and claimed volumes from company water reports, with every extracted figure checked by hand.

  1. First step

    List every irrigation project in Microsoft's, Google's and Amazon's latest water reports that names its farms or district, and pull five seasons of OpenET data for the first three US projects.

  2. Done looks like

    At least 10 named projects, each with its claimed volume, the satellite estimate of the change in consumed water with its uncertainty, and the share of the claim that estimate supports, published with code and with field boundaries where they are public.

Open 02 · the brief

Log water in credited paddies

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A one-season record of the water actually applied to rice fields in an alternate-wetting-and-drying carbon project and to similar fields outside it. Each field gets a water-level logger in its field water tube, so every irrigation and every dry spell is recorded, plus a one-time flow test of the pump or channel that feeds it.

Why it is still open

The Pakistan project gives its saving as "up to 30%"; the Bangladesh trials show savings depend on how farmers pay for water; and Verra voided almost all earlier rice credits partly for lack of evidence that practices changed. No published logger record from a credit project's fields was found.

goldstandard.orgdoi.orgrmi.org

Built with what exists

Pressure-transducer water-level loggers such as Onset's HOBO U20L, set in the perforated tubes; a portable flow meter for the one-time flow test; Sentinel-1 radar images, free from the EU's Copernicus programme, to check flooding between visits; Python for analysis.

  1. First step

    Ask one project developer for access to its fields, and test three loggers in field tubes for two weeks against daily manual readings.

  2. Done looks like

    40 project fields and 40 comparison fields logged through one season, with irrigation events, days flooded and estimated cubic metres of water per hectare for each field, weighed yields, and all raw data released.