How is d-iap built?
Delve into the methodology used to evaluate the effects of drought
Simulation grid and associated database
The simulation grid selection was based on the Global Land Cover - SHARE (GLC-SHARE) database from FAO. GLC-SHARE, with a spatial resolution of 30 arc-seconds (approximately 1 km² per pixel), combines the "best available" high-resolution national, regional, and sub-national land cover databases. The harmonization of these diverse databases is achieved using the Land Cover Classification System (LCCS). GLC-SHARE provides eleven major thematic land cover layers, each representing the proportion of the 1-km pixel in a specific land cover class. The data, provided in GeoTIFF format, utilizes the World Geodetic System 1984 (WGS 84) coordinate system and is accessible via the FAO Geonetwork site. For the platform's purposes, class number 2, 'Cropland,' was selected from the 11 aggregated land cover classes, and defined as: “Herbaceous Crops: The class is composed of a main layer of cultivated herbaceous plants (graminoids or forbs). It includes herbaceous crops used for hay. All the non-perennial crops that do not last for more than two growing seasons and crops like sugar cane where the upper part of the plant is regularly harvested while the root system can remain for more than one year in the field are included in this class”.
As mentioned earlier, GLC-SHARE has a spatial resolution of 30 arc-seconds. To align with the grid scale defined for d-iap (0.1 degrees), the percentage share values of the land cover classes were aggregated using QGIS. A zonal statistics operation was applied to the raster layer to calculate the average value for each land cover class within the designated grid cells. The simulation grid was then defined by selecting cells where the percentage of the 'Cropland' class was equal to or greater than 10%. In addition, after a preliminary analysis of the resolutions of the available databases for each AquaCrop input data (mainly climate and soil), a balance between data volume and the highest possible resolution for the simulation grid was sought, considering computational capacity and processing times. Ultimately, the spatial resolution was set at 0.1° x 0.1° (approximately 81 km²). This decision resulted in a global grid of 1,586,653 cells. After reclassifying the soils (as detailed in the “Soil data” subsection within the “Input data” section), the number of cells was finally reduced to 370,674.
Input data
Dive into the input data used for the development of d-iap
Generation of aquacrop input files
Understand how Aquacrop files are generated by selecting the different file types
Estimation of indicators
The targeted indicators for assessing the impacts of drought (refer to the “What can you get from d-iap” section) are derived from the outputs of projects simulated by AquaCrop. A dedicated Python script was developed to estimate and store these indicators in a Geopackage database for display within the d-iap interface. The following outlines the procedure used for their estimation:
At the core of d-iap is the AquaCrop v 7.1 model (Steduto et al., 2009; Salman et al., 2021), which was used to simulate the effects of drought on crop and water productivity, as well as irrigation water requirements, under present and future climate scenarios on a global scale. Crop simulation models are excellent tools for studying the effects of water availability on crop production and also on water balance components. The structure of the AquaCrop model allows for the assessment of the combined effects of water stress on crop canopy cover development and senescence, root development, stomatal closure, and building up of the harvest index (Raes et al., 2023). Additionally, in AquaCrop, elevated atmospheric CO concentrations induce stomatal closure, thereby reducing crop transpiration (Raes et al., 2023). All this makes it one of the best models for simulating the effects of water stress on crop under both current and future conditions (Tenreiro et al., 2020).
2
Rainfed management
Three levels of drought stress were categorized as follows:
-
Mild stress: Yield is reduced by less than 30%.
-
Moderate stress: Yield reduction between 30% to 70%.
-
Severe stress: Yield is reduced more than 70%.
To estimate yield reduction, a reference yield (Yo) was defined for each project. Yo represents the average dry yield of the lowest 20% of years in terms of water stress (20th percentile). To select the years with the least water stress, a water stress indicator (StW) was as the sum of ExpStr (water stress for canopy expansion) and StoStr (water stress for stomatal closure), both outputs of AquaCrop. To exclude thermal stress from the yield reduction estimation, a thermal stress threshold (TSTh) was calculated using the average "TemStr" (temperature stress), output from AquaCrop, and its standard deviation for the years used to determine Yo. Excess TemStr above TSTh adjusts dry yield through a relationship established between biomass reduction and temperature stress. Once the yield reduction associated with drought was estimated for each year, the probability of occurrence for the three stress levels was calculated.
The potential income loss due to drought-induced stress (refer to the 'What can you get from d-iap' section) was estimated based on the minimum and maximum reductions in dry yield for each stress level. After converting dry yield to fresh yield, these reductions were multiplied by the producer price (refer to the 'Economic data' subsection within the 'Input data' section). In cases where producer price data were unavailable, 'NO DATA' will be displayed.
The maximum net irrigation requirements, minimum net irrigation requirements, median net irrigation requirements, and net irrigation requirements at the 80th percentile will be displayed for users interested in irrigating to mitigate the effects of drought. This information is crucial for making necessary investments and planning irrigation water supply effectively. These indicators are derived from the 'Irr' variable, an output of AquaCrop, excluding the lowest 20% of years with the least irrigation requirements. This exclusion is a standard criterion in irrigation system design to ensure a safety margin. To estimate the maximum income gain due to irrigation and the minimum income gain due to irrigation, the difference in fresh yield between rainfed and irrigated management was calculated, and then multiplied by the producer price.
The maximum net irrigation requirements, minimum net irrigation requirements, median net irrigation requirements, and net irrigation requirements at the 80th percentile are estimated similarly to rainfed management. Moreover, d-iap provides the option to assess the probability of meeting irrigation requirements with user-provided water supply.
Additionally, crop water productivity per unit of evapotranspiration is estimated by dividing the yield by the sum of soil evaporation and crop transpiration, outputs provided by AquaCrop. For crop water productivity per unit of irrigation, the yield increase due to irrigation is divided by the net irrigation requirements. These variables are calculated annually, and d-iap displays their average and standard deviation values.
Irrigation management
References
Geerts, S., Raes, D., Garcia, M., Miranda, R., Cusicanqui, J. A., Taboada, C., Mendoza, J., Huanca, R., Mamani, A., Condori, O., Mamani, J., Morales, B., Osco, V., and Steduto, P. (2009). Simulating yield response of quinoa to water availability with AquaCrop. Agronomy Journal, 101(3), 499-508.
Jones, M. R., Singels, A., Chinorumba, S., Poser, C., Christina, M., Shine, J., Annandale, J., and Hammer, G. L. (2021). Evaluating process-based sugarcane models for simulating genotypic and environmental effects observed in an international dataset. Field Crops Research, 260, 107983.
Raes, D., Steduto, P., Hsiao, T.C., and Fereres, E. (2023). Chapter 3. Calculation procedures. In AquaCrop Version 7.1. Reference manual, D. Raes, P. Steduto, T. C. Hsiao, and E. Fereres, eds. (FAO), pp. 1–167.
Salman, M., Garcia-Vila, M., Fereres, E., Raes, D., and Steduto, P. (2021). Enhancing Crop Water Productivity - The AquaCrop Model -Ten years of development, dissemination and implementation, 2009 – 2019. FAO, Rome.
Saxton, K.E., and Rawls, W.J. (2006). Soil water characteristic estimates by texture and organic matter for hydrologic solutions. Soil Science Society of America Journal, 70, 1569-1578.
Steduto, P., Hsiao, T.C., Raes, D., and Fereres E. (2009). AquaCrop – the FAO crop model to simulate yield response to water: I. Concepts and underlying principles. Agronomy Journal, 101(3), 426-437.
Tenreiro, T.R., García-Vila, M., Gómez, J.A., Jimenez-Berni, J.A., and Fereres, E. (2020). Water modelling approaches and opportunities to simulate spatial water variations at crop field level. Agricultural Water Management, 240, 106254.
Tsegay, A., Raes, D., Geerts, S., Vanuytrecht, E., Abraha, B., Deckers, J., Bauer, H., and Gebrehiwot, K. (2012). Unravelling crop water productivity of tef (Eragrostis Tef (Zucc.) Trotter) through AquaCrop in northern Ethiopia. Experimental Agriculture, 48(2), 222-237.
Wellens, J., Raes, D., Fereres, E., Diels, J., Coppye, C., Adiele, J. G., Ezui, K. S. G., Becerra, L.A., Selvaraj, M.G., Dercon, G., and Heng, L. K. (2022). Calibration and validation of the FAO AquaCrop water productivity model for cassava (Manihot esculenta Crantz). Agricultural Water Management, 263, 107491.