Main content area

Evaluation of AquaCrop model simulations of cotton growth under deficit irrigation with an emphasis on root growth and water extraction patterns

Tsakmakis, I.D., Kokkos, N.P., Gikas, G.D., Pisinaras, V., Hatzigiannakis, E., Arampatzis, G., Sylaios, G.K.
Agricultural water management 2019 v.213 pp. 419-432
Food and Agriculture Organization, Gossypium hirsutum, biomass, canopy, cotton, crop models, crop yield, crops, deficit irrigation, heat sums, root growth, simulation models, soil water, soil water content, topsoil, Greece
One of the most vital parameters for the robust crop growth models’ performance is the crops’ root growth pattern. However, its reference measurement methods are laborious, destructive and costly. In this paper we determined the root growth pattern of cotton (Gossypium hirsutum) in the 50 to 100 cm top-soil layer using soil water content measurements from the cotton cultivating seasons of 2015 and 2016 in Northern Greece. The estimated root growth pattern along with canopy cover, biomass, soil water content and final seed cotton yield measurements were then used to evaluate the capability of the FAO AquaCrop model to simulate a deficit irrigated cotton, cultivated under real farming conditions. To do so, a number of existing cotton crop files from the literature were tested. The results showed that the estimated root growth patterns were almost the same in 2015 and 2016 exhibiting root growing rates equal to 1.7 and 2 cm/d, respectively. When the model was run in growing degree days mode, it simulated root growth pattern, canopy cover, biomass and soil water content with fair accuracy for all the proposed crop files (R2 ≥ 0.93, modeling efficiency ≥ 0.91), but the seed cotton yield was simulated adequately only when the AquaCrop’s library file was used. In calendar days mode the model failed to simulate root growth pattern satisfactorily, but the simulation of canopy cover, biomass and soil water content was fair (R2 ≥ 0.75, model efficiency ≥ 0.72). Lastly, the seed cotton yield in the calendar days mode was once again simulated accurately only when the model’s default crop file was used.