by Jesse Jankowski

Crop evapotranspiration (ET) is the biggest managed loss of water in California, accounting for roughly 80% of human net water use, and includes crop water applications transpired from plants and evaporated from soil. Methods to estimate ET have been developed based on a robust scientific understanding of its physics and data collected in the field or remotely by aircraft or satellites. Irrigation decisions often incorporate approximations of ET to help meet crop water demands, aided by field data from a state-supported network of weather stations.

For water managers, accurate ET estimation is important in large-scale accounting to calculate water available for multiple uses. In the Sacramento-San Joaquin Delta, consumptive water use informs project operations and affects the availability of environmental flows for fish habitat and salinity management. Water rights transfers and impacts of land fallowing also can be quantified by comparing ET from specific crops and bare soil. States like Idaho, with simpler agricultural systems, rely on ET models with remotely-sensed data to administer water rights.

As local parties work to implement the Sustainable Groundwater Management Act (SGMA) in California, ET measurements and estimates will be particularly important in modeling surface and groundwater availability and interactions.

At the request of the State Water Resources Control Board Office of the Delta Watermaster, UC Davis’ Center for Watershed Sciences convened seven modeling teams and one field team from the UC Davis Department of Land, Air and Water Resources to measure and estimate ET from agricultural lands in the Delta during the 2015 and 2016 water years (October 2014 through September 2016). Participating modeling teams included the Department of Water Resources (CalSIMETAW and DETAW models), the USDA Agricultural Research Service (DisALEXI model), the Irrigation Training & Research Center at Cal Poly (ITRC-METRIC model), NASA’s Ames Research Center (SIMS model), and UC Davis (UCD-METRIC and UCD-PT models).

Field stations were deployed to five fallow fields in 2015 and 14 fields of the dominant land uses, alfalfa, corn, and pasture, in 2016. These sites provided ground-based data for comparison to the modeled ET estimates. Land use surveys for each year were used to analyze ET estimates and gauge trends for specific land uses; 26 crop categories were selected to compute total agricultural ET within the Delta.

The estimates vary but give a sense of the total annual ET in the Delta, show trends for major crops in the area, and help quantify uncertainties for different land covers and times of year. The comparative study also provided policy insights for the use of model and field data in water resource management and to help improve estimation of ET in California.

Delta Service Area, ET model estimates, and field stations deployed for this study.

A 2016 Interim Report included a “blind comparison” of the models for the 2015 water year only. The Final Report contains results which benefited from group learning, standardized input datasets, and access to the UC Davis field data for the 2015 and 2016 water years. The seven models estimated about 1.4 million acre-feet of annual consumptive water use in the Delta; each model was within 11% of this average. Most ET occurs during the summer growing season (March through September) from five major land uses: alfalfa, corn, fallow lands, pasture, tomatoes, and vineyards.

Estimated annual ET, in thousand acre-feet, from crops in the Delta in 2015 and 2016.      *Information about specific models is provided in technical appendices to the Final Report.

The most common crops in the Delta in both years were alfalfa, corn, and pasture, which made up about 40% of Delta agricultural land and nearly 60% of its annual crop ET. Fallow lands increased to about 17% of the Delta’s agricultural lands and crop ET in 2016, though the 20 days of field measurements over bare soil in 2015 suggested that evaporation could be lower than predicted by the models. The largest differences between models occurred late in the growing season for almonds, corn, and potatoes, and relative variations between estimates were larger in the non-growing season when lands are typically fallow and ET is low due to colder temperatures and cloudy skies.

Detailed comparisons between models suggest that model assumptions, alternate input datasets, and interpolation between satellite images contributed to most differences and uncertainties among ET estimates. Although estimates could be improved with further calibration, ET models will also differ due to their human components: even when automated with computers, the expertise of a modeler is required to develop, run, and use them.

Estimated total monthly ET, in thousand acre-feet, from major crops in the Delta in 2016.

Several policy insights and recommendations arise from this work:

• Land use surveys such as the ones done for 2015 and 2016 in this study are valuable for water planning. Years of data shows trends in land use, like the increased fallowing in the Delta in 2016 which might show preparation for permanent tree or vine crops. When combined with other input data from satellites, field stations, or models, land use surveys also can help estimate ET from specific locations and areas. Support for land use surveys and additional information on irrigation methods, winter crops, and native versus invasive vegetation will be important for a variety of water management and policy decisions.
• Remote sensing-based water use estimates are less labor-intensive, offer better coverage, and provide more standardized estimation than diversion reports from individual water users. Even if such estimates are sometimes less accurate, they are more consistent through space and time. The results of the seven different models in this study help quantify when and where these uncertainties might be larger, as even small errors represent water with potentially high economic value, particularly during droughts.
• More accurate and understandable evapotranspiration data can support better water planning at regional scales with full-coverage information. Farm-scale irrigation management can be improved with high-resolution estimates, and water trading is aided by quantifying the water saved from crop shifts or land fallowing. Better ET estimates also improve the accuracy of water balances, for Delta water operations and for groundwater balances in critical basins across California implementing the Sustainable Groundwater Management Act.
• Meteorology data collected in the field can be used to estimate ET on much finer scales. Microclimates like the “Delta Breeze” and other temperature and humidity variations may cause ET to be lower than expected by models, and specific irrigation and crop maintenance practices will have their own impacts. Additional field data is needed for ET from fallow lands, particularly on Delta islands below sea level. A new field study is underway for the 2018 growing season. More comparisons and cooperation across field measurements and model estimates will be especially useful for unique regions like the Delta.
• Although this study focused on crop ET, about 12% of the Delta is natural vegetation such as woodlands, riparian zones, and floating plants. Another 18% is open water or urban areas. Some non-agricultural lands may have higher ET than crops, so habitat restoration efforts could increase regional consumptive water use. Because most estimation methods are tailored towards agricultural ET, further model refinements and more field data from both upland and riparian vegetation are needed.
• Having the State of California establish a collaborative group of agencies, research centers, academic institutions, and consultants to continue the study of evapotranspiration in the Delta and elsewhere would help improve ET estimates and increase the effectiveness of state and local investments in ET estimation. The exchange of common datasets and standards enhances transparency, access to technical information, public knowledge, and reduces overall costs. The large amounts of data made publicly available through this project alone are a great opportunity for further research.

The Final Report for the project and full model and field datasets can be viewed at the project website.

Jesse Jankowski (jjankowski@ucdavis.edu) is a graduate student in Civil Engineering at UC Davis and a research assistant at the Center for Watershed Sciences.

Further Reading

Allen, R.G., Pereira, L.S., Raes, D. and Smith, M. Crop evapotranspiration- Guidelines for computing crop water requirements. Food and Agriculture Organization of the United Nations irrigation and drainage paper 56, 1998.

California Department of Water Resources. California Irrigation Management Information System- Resources. 2018.

California Department of Water Resources. Land Use Viewer. 2018.

Idaho Department of Water Resources. Mapping Evapotranspiration. 2018.

Medellín-Azuara, J., Paw U, K.T., Jin, Y. Jankowski, J., Bell, A.M., Kent, E., Clay, J., Wong, A., Alexander, N., Santos, N., Badillo, J., Hart, Q., Leinfelder-Miles, M., Merz, J., Lund, J.R., Anderson, A., Anderson, M., Chen, Y., Edgar, D., Eching, S., Freiberg, S., Gong, R., Guzmán, A., Howes, D., Johnson, L., Kadir, T., Lambert, J.J., Liang, L., Little, C., Melton, F., Metz, M., Morandé, J.A., Orang, M., Pyles, R.D., Post, K., Roosevelt, C., Sarreshteh, S., Snyder, R.L., Trezza, R., Temegsen, B., Viers, J.H. A Comparative Study for Estimating Crop Evapotranspiration in the Sacramento-San Joaquin Delta- Final Report, Appendix set, and project datasets. Center for Watershed Sciences, University of California Davis, 2018.

Mount, J., Chappelle, C., Gray, B., Hanak, E., Lund, J., Cloern, J., Fleenor, W., Kimmerer, W., and Moyle, P. California’s Water: The Sacramento-San Joaquin Delta. Public Policy Institute of California, 2016.