Simulation of evapotranspiration and yield of maize: An Inter-comparison among 41 maize models

Bruce A. Kimball, Kelly R. Thorp, Kenneth J. Boote, Claudio Stockle, Andrew E. Suyker, Steven R. Evett, David K. Brauer, Gwen G. Coyle, Karen S. Copeland, Gary W. Marek, Paul D. Colaizzi, Marco Acutis, Seyyedmajid Alimagham, Sotirios Archontoulis, Faye Babacar, Zoltán Barcza, Bruno Basso, Patrick Bertuzzi, Julie Constantin, Massimiliano De Antoni MiglioratiBenjamin Dumont, Jean Louis Durand, Nándor Fodor, Thomas Gaiser, Pasquale Garofalo, Sebastian Gayler, Luisa Giglio, Robert Grant, Kaiyu Guan, Gerrit Hoogenboom, Qianjing Jiang, Soo Hyung Kim, Isaya Kisekka, Jon Lizaso, Sara Masia, Huimin Meng, Valentina Mereu, Ahmed Mukhtar, Alessia Perego, Bin Peng, Eckart Priesack, Zhiming Qi, Vakhtang Shelia, Richard Snyder, Afshin Soltani, Donatella Spano, Amit Srivastava, Aimee Thomson, Dennis Timlin, Antonio Trabucco, Heidi Webber, Tobias Weber, Magali Willaume, Karina Williams, Michael van der Laan, Domenico Ventrella, Michelle Viswanathan, Xu Xu, Wang Zhou

Research output: Contribution to journalArticlepeer-review

7 Scopus citations


Accurate simulation of crop water use (evapotranspiration, ET) can help crop growth models to assess the likely effects of climate change on future crop productivity, as well as being an aid for irrigation scheduling for today's growers. To determine how well maize (Zea mays L.) growth models can simulate ET, an initial inter-comparison study was conducted in 2019 under the umbrella of AgMIP (Agricultural Model Inter-Comparison and Improvement Project). Herein, we present results of a second inter-comparison study of 41 maize models that was conducted using more comprehensive datasets from two additional sites - Mead, Nebraska, USA and Bushland, Texas, USA. There were 20 treatment-years with varying irrigation levels over multiple seasons at both sites. ET was measured using eddy covariance at Mead and using large weighing lysimeters at Bushland. A wide range in ET rates was simulated among the models, yet several generally were able to simulate ET rates adequately. The ensemble median values were generally close to the observations, but a few of the models sometimes performed better than the median. Many of the models that did well at simulating ET for the Mead site did poorly for drier, windy days at the Bushland site, suggesting they need to improve how they handle humidity and wind. Additional variability came from the approaches used to simulate soil water evaporation. Fortunately, several models were identified that did well at simulating soil water evaporation, canopy transpiration, biomass accumulation, and grain yield. These models were older and have been widely used, which suggests that a larger number of users have tested these models over a wider range of conditions leading to their improvement. These revelations of the better approaches are leading to model improvements and more accurate simulations of ET.

Original languageEnglish
Article number109396
JournalAgricultural and Forest Meteorology
StatePublished - 15 Apr 2023
Externally publishedYes


  • Crop models
  • Evapotranspiration
  • Maize
  • Simulation
  • Water use
  • Yield


Dive into the research topics of 'Simulation of evapotranspiration and yield of maize: An Inter-comparison among 41 maize models'. Together they form a unique fingerprint.

Cite this