Comparing Delta Consumptive Use: Preliminary Results from a Blind Model Comparison

By Josué Medellín-Azuara, Kyaw Tha Paw U, Yufang Jin, Quinn Hart, Eric Kent, Jenae’ Clay, Andy Wong, Andrew Bell, Martha Anderson, Daniel Howes, Forrest Melton, Tariq Kadir, Morteza Orang, Michelle M. Leinfelder-Miles, J. Andres Morande, William Li, and Jay R. Lund

As California works to improve its official accounting of water for a range of purposes, one major area lacking widely accepted quantification is the consumptive use of water for agriculture, particularly evapotranspiration (ET) from crops.  In the Sacramento-San Joaquin Delta, such estimates are important, along with other hydrologic flows, for a variety of water rights, operational, and regulatory purposes.

Consumptive use is the proportion of water removed that cannot be reused elsewhere in a basin. For crops in the Sacramento-San Joaquin Delta, this is mostly evapotranspiration. In a region’s water balance, consumptive use can become a keystone for estimating groundwater recharge, outflows from a basin, and the availability of water for water exchanges or market transactions. In places like the Sacramento-San Joaquin Delta (the Delta), crop consumptive use estimation may have the additional challenges of adjusting for a collection of localized factors such as fog, canal seepage, evaporation from canals, and widely varying wind conditions.


Figure 1. 2015 Land use in the Sacramento-San Joaquin Delta, by LandIQ.

Models for calculating consumptive use have been researched for many years and have the potential for improved accuracy in estimating Delta outflows and informing water management and diversion in this complex and fragile estuarine system.  Yet, given the many agencies, stakeholders, and institutions relying on this water system hub, self-reporting of water diversions and estimates of consumptive use are often a challenge.

In an effort to reduce the information gap and converge efforts on consumptive use research, estimation and measurement, the State Water Resources Control Board’s Office of the Delta Watermaster convened funding from several state, regional, and local Delta water agencies to do a comparative study on consumptive use for the Delta. Researchers from the Center for Watershed Sciences and the Department of Land, Air and Water Resources at UC Davis partnered with researchers from UC Cooperative Extension, DWR, NASA-Ames Research Center, USDA-Agricultural Research Service, and CalPoly’s Irrigation Training and Research Center to conduct this study. The two-season study covers the 2014-2015 and 2015-2016 water years and includes completion of land use surveys for each year, deployment of field measurement equipment (thanks to the cooperation of farmers and organizations within the Delta), and estimation of consumptive use using seven methods and models such as CalSIMETAW, DETAW, METRIC, Priestley-Taylor and SIMS.

The study includes an initial report on measurement of evapotranspiration in bare soil, and a blind comparison of evapotranspiration estimates using seven different methods, applied independently. The interim report is posted on the project website at:

Preliminary findings show that bare soil evapotranspiration at the end of the irrigation season was close to zero in four locations selected for field measurements during September-October of 2015 (Figure 2).


Figure 2. Daily actual ET (ETa) measured from bare soil stations (surface renewal and eddy covariance) along with daily reference ET (ETo) from the nearby CIMIS stations. Lines are mean values across stations and gray shading represents one standard deviation from the mean.

A more extensive field campaign during 2016 includes more than a dozen measurement stations over corn, pasture and alfalfa fields, and covers more of the growing season. This will greatly improve measurement and enhance future comparisons of ET estimation methods.

Preliminary findings are that the average annual crop evapotranspiration estimated by the ensemble of the models tested in the Delta Service Area is about 1,550 thousand acre feet (TAF) (Figure 3). That ensemble estimate is roughly consistent with the annual estimates used in DWR’s California Water Plan Update 2013. Crops with the highest variation in ET estimates across estimation methods under comparison are are tomatoes, vineyards and potatoes.  All estimates are within about 20% of the median estimate with the higher discrepancies during December and January, months with lower crop evapotranspiration, and September, in which some crops are being harvested.

The fact that all estimation methods do not closely agree is not surprising, and would be highly unusual in comparisons of ET estimation. However, further work should be able to increase the agreement among these estimates. In addition, it should improve understanding of why and how much different estimation methods are likely to differ, and how modeling and remote sensing estimates are likely to differ from field estimates of crop ET.


Figure 3. Total evapotranspiration estimates for selected crops in the Legal Delta and the Delta Service Area (October 2014-September 2015). Estimates are derived from monthly average estimates of daily evapotranspiration for each method. CalSIMETAW and DETAW coverage is limited to the Delta Service Area.

Discrepancies in model estimates are driven by several factors including variation in input datasets and processes, inherent methodological differences, and other varying meteorological conditions. Refined estimates produced by all of the methods during the second year of study will use common input datasets and protocols to enhance the value of comparisons.  In addition, improved algorithm calibration based on analysis of 2015 data is likely to reduce the range of variation in ET estimates across methods and improve the accuracy and credibility of all methods.

A final report on this two-year study will include final estimates of consumptive use from all methods for the 2014-2015 and 2015-2016 water years. The final report will benefit from a longer field measurement campaign and improved common protocols and input datasets. It is expected to be available during late spring 2017.

Improving estimation of consumptive use is a step toward improving the reliability and transparency of data and analysis for water management, with potential for reducing reporting and accounting expenses for farmers and the state.

Josué Medellín-Azuara, Jay R. Lund, and Andrew Bell are affiliated with the Center for Watershed Sciences, UC Davis. Kyaw Tha Paw U, Yufang Jin,  Quinn Hart, Eric Kent, Jenae’ Clay, and Andy Wong are afiilliated with Land Air and Water Resources, UC Davis. Martha Anderson is affiliated with USDA-ARS. Daniel Howes is affiliated with CalPoly Irrigation Training and Research Center. Forrest Melton is affiliated with NASA-Ames, Monterrey. Tariq Kadir and Morteza Orang are afilliated with the Department of Water Resources. Michelle M Leinfelder-Miles is affiliated with UC Cooperative Extension.

Further Reading

Medellín-Azuara, J., Paw U, K.T., Jin, Y., Hart, Q., Kent, E., Clay, J.,Wong, A., Bell, A., Anderson, M., Howes, D., Melton, F., Kadir, T., Orang, M., Leinfelder-Miles, M., and J.R. Lund. (2016). Estimation of Crop Evapotranspiration in the Sacramento San Joaquin Delta for the 2014-2015 Water Year. An Interim Report for the Office of the Delta Watermaster, State Water Resources Control Board. Center for Watershed Sciences, University of California, Davis. Last Access September 28, 2016

Medellin-Azuara, J. and Howitt, R. (2013) Comparing Consumptive Agricultural Water Use in the Sacramento-San Joaquin Delta: A Proof of Concept Using Remote Sensing, Center for Watershed Sciences University of California, Davis. Last Access September 28, 2016.

Siegfried, Lucas J.; Fleenor, William E.; & Lund, Jay R.(2014). Physically Based Modeling of Delta Island Consumptive Use: Fabian Tract and Staten Island, California. San Francisco Estuary and Watershed Science, 12(4). Jmie_sfews_20875. Last access September 28, 2016


The authors are thankful for funding and research support provided by the State Water Resources Control Board, California Department of Water Resources, the Delta Protection Commission, the Delta Stewardship Council, the North Delta Water Agency, the Central Delta Water Agency, and the South Delta Water Agency.  The authors also appreciate the assistance provided in various capacities by Nadya Alexander, J. Andrés Morandé, William Li, Cathryn Lawrence and Barbara Bellieu, who made this report possible.

This entry was posted in California Water, Delta, Uncategorized, Water System Modeling and tagged . Bookmark the permalink.

9 Responses to Comparing Delta Consumptive Use: Preliminary Results from a Blind Model Comparison

  1. Pingback: California Water and Drought News for October 10, 2016

  2. Pingback: Blog round-up: California’s fish emergency, Comparing Delta consumptive use; Dams, starving Delta smelt, and more … | MAVEN'S NOTEBOOK | Water news

  3. Does anyone have a contact that can run a scenario to verify the salinity affects on the Delta if a shipping lock was put in to protect the 3 bridges at Benicia and to stop the salt water intrusion from the shipping channel? This would only effect 1 of the 12 sections between the bridge supports, but since salt water is heavier than fresh water this is the major salinity contributor. Thanks,


    • jaylund says:

      Such solutions were considered in depth in the late 1920s. Regulating salinity with outflow supplied by dams was found to be more economical at that time.


      • Joseph Rizzi says:

        They studied dams and barriers but they did not study just putting in a lock system in the shipping channel only and leaving 11/12th’s of the bay alone for free travel for fish and small watercraft.

        The original study was to make the Delta a freshwater reservoir.


      • jaylund says:

        In the 1920s, they did examine locks for locations roughly between Benicia and Pittsburg. The most recent serious proposal I saw was a 1978 modeling study of putting a sill under Carquinez Strait to keep salt water out.


      • Great study, which states that 1.2 MAF of water is required to naturally keep the salt water to the west if DAM and Lock was installed. BUT, I am only looking at installing a lock blocking only 1/12 of the flow in both directions. This is clearly not a Dam or Barrier that was previously studied. The shipping channel is dredged yearly to keep the water way deep for ships, but this also keeps this shipping channel deep which allows the heavier salty sea water a prime channel to move it’s salty water into the Delta. PPIC studied and reported that 71% of fresh water released from northern dams like Shasta to hold back the salt water intrusion.

        If a Lock system to protect the 3 Benicia bridges (1/12 of the 1 mile span) also prevents 20% of the Salty Sea water intrusion that would all less fresh water needed to keep salty water to the west of Benicia and makes the Delta a fresher water area that supports more life(fish, plants, birds, etc…) The 11/12 of the Benicia span can be left just as it is today.

        I am looking to understand if the 1/12 section for the shipping lock will prevent 10%, 20% or 30% of the salt water intrusion into the Delta? If you know of someone that can run this scenario to give me a more scientific answer, that would be great.

        If you can find a link to that 1978 study, I would love to read it.



  4. Pingback: SCIENCE NEWS: Comparing Delta consumptive use, climate change and oysters, hydrogen-powered ferries, Southwest megadrought, gauging a warmer world, and more … | MAVEN'S NOTEBOOK | Water news

  5. Pingback: Watershed Health – Oak Services, LLC

Leave a Comment (moderated)

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s