AgWaterAI

Groundwater program method

Method

Method v1.1, October 7, 2026

How the three groundwater numbers are made.

A farm in the program carries three numbers, each with its own label, its own gate and its own source. An efficiency figure against a published benchmark is a performance claim. A measured reduction is pumping below the farm’s own weather-adjusted history, measured and signed by us. A validated benefit is that reduction carried through an independent validator. The three are shown side by side and never mixed.

How it works

From the farm’s wells to a sponsor payment, in five steps.

  1. The farm’s wells

    Metered pumping on the farm’s own wells. Where there is no meter, logged pump run-time, with that basis stated on the record, until one goes in.

  2. Our sensors and readings

    Soil probes on the blocks, pressure chamber readings by our crew on the monitoring plan’s schedule, and CIMIS weather.

  3. Signed record

    Probe reading summaries are signed into a tamper-evident record. Each issued order is kept with the readings behind it. Each record states whether a value is measured, calculated, estimated or agency-reported.

  4. Independent check

    The three numbers are computed and labelled. A validator reviews before anyone claims a volumetric benefit.

  5. Sponsor payment

    The sponsor pays for the sensors and the evidence. The grower receives 50% of the net. Nothing is paid until a sponsor signs.

The three numbers

Definitions, labels and gates.

  1. Available now

    Benchmark efficiency

    The farm’s own water figure for a full season, compared with a named and dated published figure for the same crop and county on the same basis. Each record states its source: pump run-time, flow meters or the GSA’s satellite ETAW. Reported as acre-feet per acre below or above that benchmark, with the benchmark’s source and basis beside it.

    Withheld when

    Needs a full water year on a stated basis: applied water from meters or from logged pump run-time and the block application rate, or the GSA’s satellite ETAW, compared with a benchmark on the same basis. A partial year is never compared with an annual benchmark.

    Sample data

    Efficiency against a published benchmark, for the farm's own fields. A performance claim, not a reduction.

    A comparison with a published figure
  2. 2027 season

    Measured reduction

    As proposed: acre-feet pumped this season below an accepted site-specific baseline, for example the mean of three or more prior metered seasons, each scaled to this season’s weather. The final baseline rule is subject to technical and assurance review. Reported before and after the deduction stack.

    Withheld when

    Needs three or more prior metered years. With fewer, the fallbacks in order are the GSA account history, a pump energy backcast, or a measured baseline year, and the number waits.

    Sample data

    Measured reduction against the farm's own weather-adjusted history. Measured and signed by AgWaterAI, not independently validated.

    Measured, not independently validated
  3. 2027 vintage, validation after the season

    Validated benefit

    The measured reduction with the baseline capped at the lower of the farm’s history and its GSA allocation (the cap is our proposal, subject to the validator’s acceptance), carried through an independent validator who is named with the number.

    Withheld when

    Called validated when the validator has completed the review, and not before. Until then it is a candidate volume, shown as not validated.

    VWBA 2.0

    Validator named on the record. Baseline capped at the allocation.

    Validated volumetric benefit under VWBA 2.0, baseline capped at the allocation, independent validator named.

    Independent validator named on the record

Benchmark sources

The published figures the efficiency number is compared with.

Each record states its source, and names the row it was compared with. Applied water is the volume delivered through the irrigation system, including irrigation inefficiency. ETAW is crop ET less the rain the crop consumed, a consumptive-use figure, lower than applied water for the same orchard. Crop ET demand is what the trees would use when well watered.

A record’s water figure and its benchmark are on the same basis. Applied water is compared with an applied-water estimate. ETAW (evapotranspiration of applied water: crop ET less the rain the crop consumed, a consumptive-use figure) is compared with an ETAW estimate on the same definition, never with an applied-water estimate. Crop ET demand rows are context and are never the comparison.

Pump run-time (the 2026 citrus records)

Applied water calculated from logged pump run-time and the block application rate, not from a flow meter, compared with a named and dated DWR county applied-water estimate for the crop. Each record states the run-time basis.

The GSA’s satellite ETAW (the 2026 Madera County records)

For farm units confirmed in the applicable Madera County GSA account, the GSA’s own satellite ETAW (Hydrosat IrriWatch), compared with a DWR county ETAW estimate on the same consumptive-use basis. Calendar year, final in January. The farm’s meter records are kept alongside the satellite figures for comparison; the record stays on the GSA’s figure until a full season of both shows how they compare.

Flow meters (future farm units)

Applied water from flow meters, for farm units that have elected meters with their GSA, compared with a DWR county applied-water estimate.

Crop

Citrus

14 published rows

San Joaquin Valley South

applied water

2.50

30 inches

navel oranges, low-volume (micro-sprinkler)

Table B: 30 acre-inches a year from year 6 on, April through October, at 85 percent distribution uniformity, plus 2.2 acre-inches for frost protection. Kern and Tulare County advisors.

Sample Costs to Establish an Orchard and Produce Oranges, Navels, San Joaquin Valley South, Low Volume Irrigation

UC Davis Agricultural and Resource Economics, UC Cooperative Extension, 2021

San Joaquin Valley South

applied water

2.50

30 inches

Tango mandarins, low-volume

Table B: 30 acre-inches a year at maturity, April through October, at 85 percent distribution uniformity, plus 2.2 acre-inches for frost protection. Tulare and Kern County advisors. Archived study; no newer mandarin study exists.

San Joaquin Valley South

ETc demand

3.30

39.54 inches

citrus, mature, wide spaced, no cover crop, fanjet

Sheet "Weekly ET", mature total 39.54 inches on 57.9 inches normal-year grass ETo (CIMIS zone 15).

Citrus ET by Age (CIMIS Average ET Estimates by Crop Age)

UC ANR Cooperative Extension, Kern County (Blake Sanden), 2011

Kern

applied water

4.05

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching. Kern: the South Central Region Office block (the valley floor), not the desert side.

Tulare

applied water

3.71

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching.

Fresno

applied water

3.53

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching.

Kings

applied water

3.79

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching.

Madera

applied water

3.46

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching.

Kern

ETc demand

4.26

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency. Kern: the South Central Region Office block (the valley floor), not the desert side.

Tulare

ETc demand

4.11

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency.

Fresno

ETc demand

4.01

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency.

Kings

ETc demand

4.07

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency.

Madera

ETc demand

3.96

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency.

Madera

ETAW

3.11

Citrus & Subtropical (DWR crop category), includes olives and avocados

Seasonal ETaw (evapotranspiration of applied water) per acre, mean of water years 2016 to 2019 in the county sheet, South Central Region Office block. DWR defines ETaw as the net irrigation water a crop needs, not including application efficiency, from modelled water-balance irrigation schedules built on soil, crop, precipitation and ETc data; its applied water is ETaw divided by application efficiency. The same consumptive-use basis as a GSA satellite ETAW figure (total ET less the rain the crop consumed). Modelled for a typical orchard of the category, where the GSA figure is measured on the farm unit; water years here, where the Madera County GSA year is the calendar year. Water years 2016 to 2019: 3.24, 3.11, 3.36, 2.74. Water year 2020 (3.11) is also in the sheet and is left out, so every DWR row here shares one four-year window.

Pistachios

14 published rows

San Joaquin Valley South

applied water

4.13

49.5 inches

pistachios, low-volume (drip)

Table A: 49.5 acre-inches a year from year 8 on, April to mid-November, at 90 percent irrigation efficiency, non-saline soil, no leaching water included.

Sample Costs to Establish and Produce Pistachios, Low-Volume Irrigation, San Joaquin Valley South

UC Davis Agricultural and Resource Economics, UC Cooperative Extension, 2020

San Joaquin Valley South

ETc demand

3.70

44.4 inches

pistachios, mature, non-saline soil, micro-irrigation

Table C: measured actual ET of well-watered mature orchards on non-saline soil, April 1 to November 15, 44.4 inches (UC research 2016 to 2019, southwestern San Joaquin Valley). Moderately saline 38.5, highly saline 30.1.

Sample Costs to Establish and Produce Pistachios, Low-Volume Irrigation, San Joaquin Valley South, Table C

UC Davis Agricultural and Resource Economics, UC Cooperative Extension, 2020

San Joaquin Valley South

ETc demand

3.70

44.35 inches

pistachios, mature (year 9, over 65 percent cover), drip, bare middles

Sheet "Biweekly ET", mature total 44.35 inches on 57.9 inches normal-year grass ETo. The broader SSJV crop ET table on the same site lists pistachios at 46.53 inches with a different coefficient set.

Kern

applied water

3.96

Almonds & Pistachios (DWR crop category)

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching. Kern: the South Central Region Office block (the valley floor), not the desert side.

Tulare

applied water

3.94

Almonds & Pistachios (DWR crop category)

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching.

Fresno

applied water

4.04

Almonds & Pistachios (DWR crop category)

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching.

Kings

applied water

3.88

Almonds & Pistachios (DWR crop category)

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching.

Madera

applied water

4.00

Almonds & Pistachios (DWR crop category)

Seasonal applied water per acre, mean of water years 2016 to 2019 in the county sheet (the tool is titled 2016-2020; the county sheet holds four years). Modelled from CIMIS ETo, crop coefficients, soil and rainfall, adjusted for irrigation efficiency and leaching.

Kern

ETc demand

3.72

Almonds & Pistachios (DWR crop category)

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency. Kern: the South Central Region Office block (the valley floor), not the desert side.

Tulare

ETc demand

3.69

Almonds & Pistachios (DWR crop category)

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency.

Fresno

ETc demand

3.81

Almonds & Pistachios (DWR crop category)

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency.

Kings

ETc demand

3.65

Almonds & Pistachios (DWR crop category)

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency.

Madera

ETc demand

3.83

Almonds & Pistachios (DWR crop category)

Seasonal crop evapotranspiration per acre, mean of water years 2016 to 2019 in the county sheet. ETo x Kc, before any irrigation inefficiency.

Madera

ETAW

3.60

Almonds & Pistachios (DWR crop category)

Seasonal ETaw (evapotranspiration of applied water) per acre, mean of water years 2016 to 2019 in the county sheet, South Central Region Office block. DWR defines ETaw as the net irrigation water a crop needs, not including application efficiency, from modelled water-balance irrigation schedules built on soil, crop, precipitation and ETc data; its applied water is ETaw divided by application efficiency. The same consumptive-use basis as a GSA satellite ETAW figure (total ET less the rain the crop consumed). Modelled for a typical orchard of the category, where the GSA figure is measured on the farm unit; water years here, where the Madera County GSA year is the calendar year. Water years 2016 to 2019: 3.97, 3.44, 3.78, 3.22. Water year 2020 (3.38) is also in the sheet and is left out, so every DWR row here shares one four-year window.

Sources: UC Davis Agricultural and Resource Economics, UC Cooperative Extension, UC ANR and California Department of Water Resources. Every figure was read from the publisher’s own document on October 7, 2026. A record is compared with the row on its own basis for the farm’s county: applied water with the applied-water row, the GSA’s satellite ETAW with the ETAW row. Crop ET demand rows are context and are never the comparison. A crop and county with no published figure on that basis has no row and no number.

Weather adjustment

The farm’s own seasons, under this season’s weather.

Each prior season is scaled to the assessment season by the ratio of CIMIS ETo (reference evapotranspiration) for the two seasons, from the station we pair the farm with, so the baseline is what the farm’s own past seasons would have pumped under this season’s weather. As proposed, the farm’s own years form the baseline where admissible pumping evidence covers three or more seasons; otherwise a prospective measured baseline may be needed.

Deductions

  • A conservative deduction is taken from the gross volume before anything is counted; we propose 23 percent as the starting figure, and its final size follows the uncertainty assessment and the validator’s acceptance.
  • An optional consumptive-use adjustment applies if the validator prefers to count consumptive use rather than withdrawal; we propose about 0.80, and the factor is proposed by the technical team and assessed by the independent assurance provider, not preset.
  • The baseline is capped at the GSA allocation for the enrolled acres, as we propose it and subject to the validator’s acceptance: pumping above the allocation is never counted as history.

Counted once

  • A pumping cut the GSA already requires never counts.
  • Acres in a fallowing or land-repurposing program that already pays for the water are excluded for the years paid.
  • Proposed control: any year the GSA account shows a transfer from an enrolled parcel is a year with no claim, and water carried over and pumped later is subtracted. Final treatment follows the accepted project method and the applicable GSA account rules.
  • The same reduced withdrawal is claimed once, by the sponsor that funded it.

The two commitments

  • No sale or transfer of the water left in the ground, for the benefit period.
  • No added irrigated acres on the enrolled farm unit, for the benefit period.

What a sponsor can say

One sentence per number.

Example language. A sponsor’s own counsel reviews any public claim before use.

NumberExample languageNot for this number
Benchmark efficiency“We fund monitoring on California orchards whose 2026 water figure, with its source stated on each record, is below a named and dated DWR county estimate on the same basis for the crop.”Claims of water replenishment, compensation for other water use, or independent validation.
Measured reduction“We funded a measured reduction of N acre-feet in applied water on California orchards against the farm’s own weather-adjusted history, measured and signed by AgWaterAI, not independently validated.”Claims of water replenishment, compensation for other water use, or independent validation.
Validated benefit“N acre-feet of volumetric water benefit under VWBA 2.0, independently validated; the validator’s statement is attached.”Anything beyond the validator’s own statement.

With every claim

  • Pressure chamber readings document plant water status and agronomic response; they do not measure pumped volume.
  • An independent third party reviews AgWaterAI’s volume calculation before any validated volume is claimed.
  • The validated tier goes through an independent assurance provider using the VWBA 2.0 accounting guidance, reconciled against the grower’s GSA account, and protected so the same water is not counted twice.
  • No validated record has been issued under this program.

The record

What is signed, and what is not.

  • Reading summaries are signed into a tamper-evident record with ECDSA P-256 keys held in Cloud KMS, hash-linked block to block, tamper-evident from the point of signing forward.
  • Unit records are signed with Ed25519 and the signature is re-checked on every lookup against our published public key. None has been issued under this program yet.
  • On the farms in our order pilot, the draft order is held back and logged with a per-week fingerprint before the adviser makes the call. The issued order is kept with the readings behind it and is not signed.
  • Well-meter record signing is planned for project deployment and will be described as live after it is.
  • Each record identifies whether the underlying value is directly measured, calculated from measured inputs, estimated or agency-reported, together with its source and period.

Tamper-evident signing preserves the integrity of an attested record from signing forward. It does not by itself establish measurement accuracy, baseline eligibility, additionality, regulatory acceptance, a reduction in water use or a volumetric water benefit. Those questions depend on the underlying instruments, the project method, GSA records and any applicable independent assurance. AgWaterAI’s record-signing method is patent-pending.

What a grower gathers

The evidence checklist.

  1. A site boundary file for the enrolled acres: shapefile, KML, or the GSA parcel map.
  2. Well meter logs, or dated photos of each meter face.
  3. Pump energy bills for the assessment period and the prior seasons.
  4. Weather: the CIMIS station we pair the farm with.
  5. Sensor data from the probes on the blocks.
  6. Photos of the irrigation equipment: pumps, filters, emitters.
  7. The assessment period, with its start and end dates.
  8. The two commitments in writing: no sale or transfer of the water left in the ground, and no added acres.

You keep your raw data. The sponsor receives the right to make the stated sentence for the stated blocks, label and season, with the record behind it. The grower receives 50% of the net sponsor payment. Nothing is paid until a sponsor signs.

Growers: program page and water potential check. Sponsors: sponsor page. Privacy policy

Method v1.1, October 7, 2026. Superseded versions stay published with their dates.

  • Method v1.0, October 7, 2026: Superseded the same day by v1.1. One source for the benchmark number, the farm’s applied water, against the DWR county applied-water estimate. v1.1 rewords the benchmark label, has each record state its own source (pump run-time, flow meters or the GSA’s satellite ETAW) and compares a figure with a benchmark on the same basis.