Author:

Keywords:

AquaCrop model, soil fertility, semi-quantitative simulation approach, relative biomass production , Science & Technology, Life Sciences & Biomedicine, Agriculture, Multidisciplinary, Agriculture, SIMULATE YIELD RESPONSE, RESOURCE USE EFFICIENCY, FIELD-GROWN SUNFLOWER, DEFICIT IRRIGATION, WATER, WHEAT, MAIZE, MANAGEMENT, QUINOA, PARAMETERIZATION, 07 Agricultural and Veterinary Sciences, Agronomy & Agriculture, 30 Agricultural, veterinary and food sciences

Abstract:

Most crop models make use of a nutrient balance approach for modelling crop response to soil fertility. To counter the vast input data requirements that are typical of these models, the crop water productivity model AquaCrop adopts a semi-quantitative approach. Instead of providing nutrient levels, users of the model provide the soil fertility level as a model input. This level is expressed in terms of the expected impact on crop biomass production, which can be observed in the field or obtained from statistics of agricultural production. This study is the first to describe extensively, and to calibrate and evaluate, the semi-quantitative approach of the AquaCrop model, which simulates the effect of soil fertility stress on crop production as a combination of slower canopy expansion, reduced maximum canopy cover, early decline in canopy cover and lower biomass water productivity. AquaCrop’s fertility response algorithms are evaluated here against field experiments with tef (Eragrostis tef (Zucc.) Trotter) in Ethiopia, with maize (Zea mays L.) and wheat (Triticum aestivum L.) in Nepal, and with quinoa (Chenopodium quinoa Willd.) in Bolivia. It is demonstrated that AquaCrop is able to simulate the soil water content in the root zone (relative root-mean-square error (RRMSE) 6-13 %), and the crop’s canopy development (RRMSE 12-34 %), dry aboveground biomass development (RRMSE 13-22 %), final biomass (RRMSE 4-24 %) and grain yield (RRMSE 7-19 %), under different soil fertility levels, for all four crops. Under combined soil water stress and soil fertility stress, the model predicts final grain yield with a RRMSE of only 11-13 % for maize, wheat and quinoa, although of 34 % for tef. This study shows that the semi-quantitative soil fertility approach of the AquaCrop model performs well and that the model can be applied, after case-specific calibration, to the simulation of crop production under different levels of soil fertility stress for various environmental conditions, without requiring detailed field observations on soil nutrient content.