AquaCrop assets and regional applications

AquaCrop is a relatively simple crop model with a wide range of applications at the point, field, and regional to continental scales. Its four main assets, distributed by FAO (https://www.fao.org/aquacrop/en/), are: (i) the standard program with a graphical user interface (GUI), (ii) the open-source version-controlled Fortran90 code available on GitHub, (iii) the stand-alone programs for Windows, macOS, and Linux, and (iv) its integration into systems for efficient regional-scale modeling, satellite-based data assimilation, and climate impact simulations. Specifically, a parallelized Python wrapper is available to run the stand-alone program, and the Fortran90 code is integrated into NASA’s Land Information System Framework (LISF).

This poster introduces AquaCrop's four assets and focuses on two regional-scale applications in Europe. First, we demonstrate the use of the parallelized Python wrapper in the context of a climate impact study, where we evaluated current and future maize yields. AquaCrop simulations were performed at a coarse spatial resolution (0.5°) to assess future changes in yields, and yield gaps (difference between actual and potential yield, without stresses). Second, the use of AquaCrop within NASA’s LISF is presented through a data assimilation experiment, in which AquaCrop simulations were performed at a 0.1° resolution. A generic type of C3 crop was used over the entire domain, and the crop stage lengths were parametrized using the VIIRS global land surface phenology. The uncertainty in simulations was assessed by perturbing meteorological inputs and soil moisture in the upper soil layers. To correct plant water stress, SMAP-enhanced surface soil moisture observations (9-km resolution) were assimilated using an ensemble Kalman filter. Results highlight (i) the need for careful mapping between AquaCrop-simulated and satellite-retrieved soil moisture and (ii) how small updates in soil moisture can propagate to significant changes in biomass development.