Generation of AquaCrop input files
Climate files
In AquaCrop, climate data are stored in four different files: a Temperature file, an ETo file, a Rain file, and a CO file. Additionally, there is a Climate file (with a '.CLI' extension) that serves as a cover file, indicating the filenames for the Temperature, ETo, Rain, and CO files. A Python script was developed for the automatic generation for each simulation cell of Temperature files, ETo files, Rain files, and Climate files from the climate database (see the 'Climate data' subsection within the 'Input data' section). For the CO file, the default CO files for each emission scenario were used in the simulations. In the historical scenario, the atmospheric CO concentration for the year 2020 was used as a reference for all years. This approach ensures consistency and avoids considering the influence of increasing CO levels over time, allowing for the evaluation of the specific responses of the other climatic variables.
2
2
2
2
2
2
Soil files
In AquaCrop, soil hydraulic properties are stored in a Soil file (with a '.SOL' extension). A specific Python script was developed for the automatic generation of soil files for each simulation cell from the data stored in the Geopackage database (see the 'Soil Data' subsection within the 'Input Data' section). No limitations for root development were considered. Additionally, the presence of gravel was neglected since it was accounted for during the estimation of soil hydraulic properties through the pedotransfer functions.
Crop files
To estimate the impacts of drought, AquaCrop simulations were performed in thermal time to account for the temperature effects on crop development, which caused interannual variations in the crop cycles. For this purpose, the crop files must express phenology and development in growing degree days (GDD). A Python script was developed for the automatic generation of crop files for each simulation cell, using the default crop files in GDD from the AquaCrop database as templates. For cassava, quinoa, teff, and sugarcane, no GDD crop files existed. Therefore, these files were generated based on data from various scientific articles on these crops worldwide (Geerts et al., 2009; Tsegay et al., 2012; Wellens et al., 2022; and Jones et al., 2021). The climate data needed for estimating the GDD for each development phase were obtained from ERA5 (see the 'Climate Data' subsection within the 'Input Data' section). For cassava, due to significant differences between the cultivars grown in Africa and America, two different cassava crop file templates were generated for these regions. In creating crop files for each simulation cell, adjustments were made to the crop file templates for the number of plants per hectare, maximum canopy cover (CCx) as a fraction of soil cover, and reference harvest index (HIo). Additionally, the maximum effective rooting depth was adjusted based on the rootable soil depth value from the Geopackage database (see the 'Soil Data' subsection within the 'Input Data' section). The rootable soil depth categories are: (1) Deep soil (>100 cm); (2) Moderately deep soil (<100 cm); (3) Shallow soil (<50 cm); and (4) Very shallow soil (<10 cm). A rootable soil depth of 1 m, 0.6 m, and 0.4 m was considered for the last three categories, while the first category was considered non-limiting, using the default maximum rooting depth for each crop in the model.
Irrigation files
In d-iap, the net irrigation requirements were estimated by using the net irrigation requirement mode in AquaCrop, specifying an allowable depletion of soil moisture, which was set at 50% of the readily available water (%RAW). The total amount of irrigation water required to keep the water content in the soil profile above this specified threshold is the net irrigation water requirement. The net irrigation requirement does not consider the extra water that has to be applied to the field to account for conveyance losses or the uneven distribution of irrigation water on the field (Raes et al., 2023) or to maintain salt balance.
Management files
In d-iap assessments, optimal crop management is assumed, with no nutritional stress or negative effects from weeds. Additionally, it is important to note that AquaCrop does not simulate the effects of pests and diseases. Lastly, no field management practices that would alter soil evaporation or runoff were considered.
Initial conditions files
Project files
To perform AquaCrop simulations using the standalone version, which allows for conducting extensive simulations as required by d-iap, project files need to be generated. These files are created for each simulation cell, taking into account the combination of crop, season, and management practices, facilitated by a Python script developed for this purpose. The “First day of cropping period”, a key variable in the project file, is determined based on specific criteria depending on whether the management is rainfed or irrigated. For rainfed management, it is set as the date when cumulative precipitation reaches 30 mm within the sowing window (refer to “Crop data” subsection within the “Input data” section) for each respective year. In contrast, for irrigated management, the sowing date is fixed at the midpoint of the sowing window. In cases where management is classified as “Irrigation/rainfed” in the crop database, if less than 80% of the years reach the 30 mm threshold, it is considered as irrigated management. The simulation period is synchronized across runs to minimize uncertainties in the soil water balance. A Python script was developed to estimate key dates such as the “First day of simulation period”, “Last day of simulation period”, “First day of cropping period”, and “Last day of cropping period” based on temperature data and the thermal time required to achieve physiological maturity for each crop.
For the first run, or year 1, of AquaCrop simulations (corresponding to 1960 in the historical scenario), it is necessary to estimate the initial soil water content. Subsequent runs use the final soil water content from the previous run as the starting point for the soil water balance, thereby reducing associated uncertainties. To establish initial conditions, it was assumed that soil moisture levels were at field capacity (FC) or permanent wilting point (PWP) on a specific date prior to sowing, adjusted based on the location. This date and soil moisture level were determined individually for each simulation cell according to the Koppen-Geiger climatic classification and hemisphere, taking into account the corresponding precipitation pattern. In regions such as India and Pakistan, which exhibit precipitation patterns different from other regions within the same Koppen class, they were assigned a distinct classification. A Python script was developed to automate this process, integrating an API that links cell coordinates to the Koppen class, thereby generating initial condition files for each cell considering its soil hydraulic characteristics.
