Global investigation of impacts of PET methods on simulating crop-water relations for maize

• Published in: Agricultural and forest meteorology Vol. 221; pp. 164 - 175
• Main Authors: Liu, Wenfeng, Yang, Hong, Folberth, Christian, Wang, Xiuying, Luo, Qunying, Schulin, Rainer
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.05.2016
• Subjects: AET; agricultural statistics; APT; ASE; climate change; corn; crop models; crop yield; Crop-water relations; CWP; CWU; environmental policy; EPIC; evapotranspiration; Global scale; growing season; GSET; humid zones; irrigated farming; irrigation water; Maize; Modelling uncertainties; PEPIC; PET; PHU; PPT; prediction; PSE; Python; rain; uncertainty; water management; water requirement; Zea mays; CWU; PSE; APT; EPIC; Maize; ET; Crop-water relations; AET; PEPIC; ASE; Modelling uncertainties; PPT; Global scale; GSET; PHU; CWP; PET
• DOI: 10.1016/j.agrformet.2016.02.017
• ISSN: 0168-1923

•We simulate crop-water relations for maize on a global scale.•We compare uncertainties derived from different PET methods on the relations.•Uncertainties are significant, especially for crop water use and productivity.•Water availability plays an important role in the uncertainties.•A proper choice of PET methods is quite important for crop growth simulation. Crop models are commonly used to investigate crop-water relations over different spatial scales. Estimating potential evapotranspiration (PET) is a basis for this investigation. Most crop models have built-in PET estimation methods. Using different methods can lead to very different PET estimates; but little is known about the sensitivity of large-scale crop model predictions on the choice of the PET estimation methods. In the work reported here, we used PEPIC, a grid-based EPIC (Environmental Policy Integrated Climate) model with a Python environment, to investigate the impacts of five different PET methods on estimated crop-water relations for maize on a global scale at a resolution of 30arcmin. Results show that the estimated PET varied largely among different PET methods for the same climate zones, leading to uncertainties in estimating crop-water relations. Uncertainties in water-related variables such as growing season evapotranspiration (GSET) and irrigation water requirement were more relevant than uncertainties in crop yields. Water availability played an important role in the uncertainties. All PET methods showed similar performance with respect to simulations of GSET for rainfed maize cultivation in low-rainfall regions, while there were large differences for regions with high rainfall. For irrigated agriculture, the estimated irrigation water requirement varied widely among the five PET methods, with a factor of 2 between the smallest and the largest estimates. Overall, using the Priestley-Taylor method led to lowest yield but highest GSET estimates. The Baier-Robertson and Hargreaves methods produced rather high GSET estimates for tropical and humid regions. The Penman-Monteith method gave the best yield estimates, compared to agricultural statistics. The results highlight the importance of considering the uncertainties resulting from the selection of PET estimation methods in investigating crop-water relations, particularly in predicting impacts of future climate change and in formulating appropriate water management strategies.