Soil information is a vital input for crop models applications in various large area studies including climate change impact and food security. One of the global soil databases that provide full information for crop models is HC27 of IFPRI. The quality of the database has not been assessed for crop modeling so far. A tested crop simulation model (SSM-iCrop2) was used for this purpose that needs soil water related properties (i.e., depth, albedo, curve number for runoff, drainage coefficient, and soil water limits at saturation, drained upper limit and lower limit) for the simulation of crop properties. Actual data of two soil profiles from three different climate zones (locations) were used as model inputs to simulate potential yield, evapotranspiration (under rainfed conditions) or net irrigation water requirement (under irrigated conditions) of some important plant species (alfalfa, sugar beet, sugar cane, wheat, olive, soybean, apricot and chickpea) under rainfed and irrigated conditions of Iran. Results showed that the application of HC27 soil information in the SSM-iCrop2 model resulted in model output that was not different from the model output with actual soil information with respect to mean, variance, and distribution. No statistically significant difference was found in the simulation of various combinations of soil profiles-plant species-locations. It was concluded that HC27 information can be used in simulation studies with SSM-iCrop2 or other similar simple models for the simulation of potential yield, net irrigation water, or evapotranspiration that are commonly required for food security and climate change studies.
Plant simulation models have become an important tool for investigating different agricultural-related scenarios like climate change. Soil information is one of the key input information in most plant simulation models as soil provides an environment for root growth, water storage and uptake and nutrient exchange. The use of reliable soil information will minimize the uncertainty of model predictions. For example, Folberth et al. (2016) reported that soil type variations relative to weather conditions in an area that could have a greater impact on model prediction. Other studies have also highlighted the importance of soil information (Woli et al. 2013; Nouri et al. 2016; Sharda et al. 2017). However, it is difficult to obtain soil information at a large area scale because accurate soil information is often available for the small number of locations where soil samples have been collected and analyzed.
The International Union of Soil Sciences (IUSS) at its Seventh Congress in Madison, USA, 1960 recommends that maps of continents and large areas should be prepared. To this end, FAO and UNESCO decided in 1961 to produce a world soil map with a 1:5,000,000 scale. This project was completed in 20 years. This map was the result of the effort of soil scientists from around the world (FAO-UNESCO 1974), and it was the only available overview of the world’s soil resources until recently. From other World Soil Maps, WRB, Soil Regions and Zobler maps can be mentioned that their basis is FAO-UNESCO (WRB 1998; WRB 2006; Zobler 1986; http://www.fao.org).
Soil data from web-based databases are useful for preparing soil input data for plant models. So far, various soil databases have been developed to assist modeling on a global and regional scale. For example, the SoilGrids (Hengl et al. 2014), developed by the International Soil Reference and Information Center (ISRIC), contains global maps of soil information. The HWSD soil map is the result of the collaboration between FAO and IIASA, ISRIC World Soil Information, the Soil Institute, the Chinese Academy of Sciences, and the European Commission Joint Research Center (JRC) (Wieder et al. 2014; http://www.isric.org). In the WISE project, a wide range of worldwide soil databases (point and network-based) have been collected to make the data more accurate at regional and global levels. These data are related to the FAO-UNESCO soil map data (Batjes 1995; http://www.isric.org). As mentioned above, today there are various sources for obtaining soil information globally. Although FAO-UNESCO, HWSD, and SoilGrids soil maps are world-class, they lack some of the information needed for simulation models. In addition, WISE soil profiles include the information required for simulation models, but it covers a small area in the world (Koo and Dimes 2013).
To solve the problem of the limitation of soil information and spatial limitation of information for use in crop simulation models, Koo and Dimes (2013) prepared general soil profiles based on three criteria of soil texture, root depth, and organic carbon for use in crop simulation models. This map is known as HC27 soil map (Koo and Dimes 2013). It is the result of the collaboration between FAO, IIASA, ISRIC, ISSCAS and JRC. This map combines more than 15,000 updated regional and national soil maps around the world with information available on FAO and UNESCO maps. All the regions of the world are networked with a size of 5 min (about 10 × 10 km) for mapping the HC27 soil.
For food security studies, it is generally necessary to estimate potential yields under rainfed and irrigated conditions as well as irrigation water requirements under irrigated conditions (Van Ittersum et al. 2013; Fereres et al. 2011; Godfray et al. 2010). Plant simulation models have been identified as one of the best methods for estimating these variables on a large scale (Van Ittersum et al. 2013). These models typically require soil properties including soil depth, soil water limits, soil drainage coefficient, and soil curve number to calculate runoff (Soltani and Sinclair 2012). This study was conducted to evaluate the quality of HC27 soil map information for use in the SSM-iCrop2 simulation model (Soltani and Sinclair 2012; Soltani et al. 2020a) by comparing the simulated potential yield, irrigation water requirements and potential evapotranspiration using actual soil information and soil information from HC27 for the same locations.
The model used in this study was SSM-iCrop2 (Soltani and Sinclair 2012; Soltani et al. 2020a). The model includes daily phenology progress, leaf area development and senescence, dry matter production, yield formation, and soil water balance. Responses of crop processes to solar radiation, temperature, water availability, and cultivar differences are included in the model. Soil water sub-model accounts for soil water additions from precipitation or irrigation, and increasing rooting depth and water removal via deep drainage, run-off, soil evaporation, and plant transpiration. The soil profile is divided into two layers: one top layer of 15–20 cm thickness and a second layer that includes the first layer and its depth increases by root growth. Soil water balance of both layers is calculated separately. The effect of water deficit and excess on leaf area development and senescence, dry mass accumulation, and phenological development are simulated. The model also accounts for the effect of freezing temperatures on plant leaf area that might take place in early spring sowings or winter sowings. The model has been tested extensively for a wide range of plant species, including those selected for the current study, and proved to be robust (Soltani et al. 2020a).
HC27 Soil Database
In the HC27 soil map, general soil profiles are based on three criteria of soil texture, root depth, and organic carbon (Koo and Dimes 2013): based on soil texture, three clay, silt, and sand groups are defined. Based on soil depth, they are classified into three groups of deep, medium, and shallow and based on fertility into three groups of high, medium, and low fertility. In total, general soil profiles include 27 types of profiles numbering from 1 to 27 that are prepared based on the compatible format to DSSAT (Jones et al. 2003) and APSIM (McCown et al. 1995) simulation models. Out of 27 soil profiles of HC27, only 9 soil profiles (2, 5, 8, 11, 12, 13, 14, 17, and 26) are found in Iran’s agricultural lands (Table S1).
Evaluation of HC27
SSM-iCrop2 model was run for several diverse plant species under rainfed or irrigated conditions of three different different locations for a period of 10 years from 2005 to 2014. The selected weather stations were Hashemabad in Golestan province, Tabriz in East Azarbaijan province, and Dezful in Khuzestan province (Table 1). According to GYGA-ED climate zonation (http://www.yieldgap.org), climate zone is 6102 for Hashemabad, 4103 for Tabriz and 8003 for Dezful, so that Hashemabad has a moderate sub-humid climate, Tabriz has a cold, semi-arid climate, and Dezful has a warm and dry climate.
The impact of HC27 soil information quality was investigated on yield (expressed as dry weight per unit area, net irrigation water requirement, and evapotranspiration simulated with the model. To do this, the model was fed with actual or HC27 soil information, but identical weather data. Diverse plant species were selected which are different with respect to plant type (C3 vs. C4, annual vs. perennial and cool-season vs. warm season) and growing season. The plant species were simulated including wheat, soybean, chickpea, alfalfa, sugar beet, sugar cane, olive, and apricot. Not all selected plant species are grown under rainfed and irrigated conditions of all the selected locations. Table 2 presents plant species-farming type-location combinations that were evaluated in the current study.
Depending on the location and selected species, information of six actual soil profiles was used (Table 3). For each of the actual soil profiles, corresponding information for the same location was extracted from the HC27 soil map (Table 3).
Mean, variance, and distribution of simulated variables using actual soil information and soil information extracted from HC27 for the same locations were compared by t test, F test, and Kolmogorov–Smirnov (K–S) test, respectively. In addition, the deviation between and correlation of simulated variables with actual or HC27 soil profiles was evaluated using root mean square error (RMSE), correlation coefficient (r), and relative RMSE (RMSE divided by variable mean using actual profile, i.e. coefficient of variation, CV). If the difference between these variables is not statistically significant, it can be concluded that HC27 soil information will lead to the same result and will be used for crop model applications.
Table 4 includes simulated yield, net irrigation water requirement and evapotranspiration for the selected plant species, farming conditions and locations using actual or HC27 soil information. The results show that differences were not significant except for one case, i.e. yield of rainfed soybean at Hashemabad station. Under irrigated conditions, it is expected both soil information result in identical yield because automatic irrigation defined in the model for this condition removes any effect of water deficit. Negligible differences are due to the effect of water logging differently simulated using actual or HC27 soil information. Net irrigation water requirement is highly depended on plant dry mass accumulation and resultant transpiration, so non-significant difference are normal. Under rainfed conditions, however, soil water limites are important as they determine soil water storage capacity. One single significant case belonged to rainfed soybean where simulated yield using actual soil information was 8% higher than simulated yield using HC27 soil data.
Simulated yield using actual soil information and soil information from HC27 are shown in Fig. 1. Most of the data are scattered around 1:1 line and the discrepancy is lower for irrigated conditions (Fig. 1). Under irrigated conditions, RMSE was 30 g/m2, which was equivalent to 2% of the mean simulated yield based on actual soil information. There was also a high correlation coefficient (r = 0.99) between simulated yield using actual or HC27 soil information. Under rainfed conditions, RMSE was 30 g/m2, which was 10% of the simulated mean using actual soil data. The correlation coefficient was 0.99 (Fig. 1). K–S, t and F tests were not able to detect any significant difference between the simulated yield with actual soil information and the simulated yield with data derived from HC27 in terms of distribution, mean or variance (Tables S2, S3, and S4). Therefore, using HC27 soil profiles instead of actual soil profiles results in similar yields using the SSM-iCrop2 simulation model.
Similar results were obtained for simulated net irrigation water requirement (for irrigated conditions) and evapotranspiration (for rainfed conditions) based on HC27 soil data versus actual soil data (Fig. 2). Most of the data are scattered around 1:1 line, especially for net irrigation water. For net irrigation water, RMSE was 42 mm, which was about 6% of the average net irrigation water requirement based on real soil information (Fig. 2). There was also a high correlation coefficient (r = 0.99; P = 0.01) between these two. For evapotranspiration, based on HC27 soil information against actual soil information under rainfed conditions, RMSE was 14 mm, which was 6% of average evapotranspiration using actual soil information, and the correlation coefficient was 0.99 (Fig. 2). According to K-S, t, and F tests, there was no significant difference between net irrigation water requirement and evapotranspiration simulated using actual soil information or soil information derived from HC27 with respect to distribution, mean or variance (Tables S2, S3, and S4). Thus, using the HC27 soil profile instead of the actual soil profile information would not cause a significant difference in the outputs of the SSM-iCrop2 simulation model.
Simulated soil water balance components were also evaluated to justify the results obtained for simulated yield, net irrigation water requirement or evapotranspiration (Tables 5, 6). Simulated irrigation water, soil evaporation, crop transpiration and deep drainage under irrigated conditions using actual or HC27 soil information were similar with no significant difference (Table 5).
The results of soil water balance simulated under irrigated and rainfed conditions showed that, with the exception of a few cases, differences were not significant for actual soil profile information and soil information derived from HC27 (Tables 5, 6). Under rainfed conditions, more cases of significant differences were found (Table 6). No significant difference was found for soil evaporation, but two differences were significant for plant transpiration; one belonged to rainfed soybean in Hashem Abad, i.e. the case with significant difference for grain yield, and the other belonged to rainfed chickpea in Tabriz. But, in this later case, the difference for plant transpiration did not result in significant simulated yield. Some difference were significant for deep drainage and run-off but these component are rather small compared to soil evaporation or plant transpiration.
Because of the need to assess the effects of environmental change on agriculture and food security for use in policymaking, it is necessary to use plant simulation models at regional and global scales (Han et al. 2019). Model application in large areas requires input data related to weather, soil, plant, and management. Access to suitable soil information is one of the fundamental problems in such applications.
In the models, depending on the processes that they simulate, the soil input parameters can vary, but in most of these models, such as APSIM and DSSAT models, there is considerable overlap in this respect (Gijsman et al. 2007). The soil information required for the models is often not obtained from data from routine soil measurements, and on the other hand, it is difficult to obtain large-scale data in the appropriate format and the model users cannot easily access such information. This is a constant problem, especially in developing countries.
Several world soil maps have been prepared to address this problem (FAO-UNESCO 1974; WRB 1998, 2006; Zobler 1986; Hengl et al. 2014; Batjes 1995; Wieder et al. 2014). Gijsman et al. (2007) created a set of soil input sets for the DSSAT model using the WISE soil database. Romero et al. (2012) developed a global dataset for soil input sets for the DSSAT model. Chaves and Hoogenboom (2014) extended this data to 9613 profiles based on WISE version 3.1. Unfortunately, some soil databases are not world-class in terms of size and others lack all the soil information required for simulation models. But, the HC27 soil map has been developed globally containing various soil parameters such as soil water limits, organic carbon, acidity, water storage capacity, soil depth, exchangeable cation capacity, total exchangeable nutrients, salinity, and soil texture. This map allows the use of soil data in many plant models. While HC27 soil profiles have been used in several crop modeling studies (Alizadeh-dehkordi et al. 2020; Dadrasi et al. 2020; Kafatos et al. 2017; Liu et al. 2020; Nehbandani et al. 2020; Soltani et al. 2020b, c), it has not been evaluated for use in crop models so far.
Our evaluations of the HC27 under irrigated and rainfed conditions indicated that using HC27 information or using actual soil profile information would not cause any statistically significant difference in SSM-iCrop2 model simulation of crop yield and net irrigation water or evapotranspiration. These variables are commonly simulated in large area assessments related to food security and climate change. Under irrigated conditions, automatic irrigation was activated in SSM-iCrop2 simulations, so the plant is not stressed and there will be no significant difference in yield based on actual soil profiles or HC27 soil profiles. However, differences in soil information can cause differences in irrigation water requirements. The results also showed that this difference was not significant. Under rainfed conditions, soil texture and depth information and soil water limits are important as they affect the amount of water available to the plant. This potentially makes soil information more important under rainfed conditions, but the results showed that the simulated differences in crop yield for rainfed conditions were not significant except for few cases. Thus, it can be stated that the application of HC27 soil information is satisfactory and can be used to simulate yield, net irrigation water, and evapotranspiration in crop modeling studies over a large area using SSM-iCrop2 simulation model. Of course, actual soil information is preferred, if available.
One limitation of the current study is that it does not evaluate HC27 soil chemical properties which are important when crop models are used to simulate soil and crop nutrients. This needs to be assessed in future studies. Furthermore, SSM-iCrop2 only consider one soil layer and the results presented here may not be applicable for crop models that consider several soil layers in simulation of soil water balance.
Using HC27 soil information compared to actual soil information in the SSM-iCrop2 simulation model produced results that were statistically similar in terms of distribution, variance, and mean. No significant difference was found in any of the location-plant species-soil profile combinations tested in this study. Therefore, the use of this soil database for simulation of potential yield and water-related variables is verified at least for SSM-iCrop2 and other models with similar complexity. Obviously, for other purposes and in models with different complexity levels, the same results may not be obtained and more tests need to be done.
Alizadeh-dehkordi, P. A., Nehbandani, A., Hassanpour-bourkheili, S., & Kamkar, B. (2020). Yield gap analysis using remote sensing and modelling approaches: Wheat in the Northwest of Iran. International Journal of Plant Production, 14, 1–10.
Batjes, N. (1995). A homogenized soil data file for global environmental research: A subset of FAO, ISRIC and NRCS profiles (Version 1.0). ISRIC.
Chaves, B., & Hoogenboom, G. (2014). Strengthening soil databases for climate change and food security modeling applications.Technical Report. Copenhagen, Denmark: CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS).
Dadrasi, A., Torabi, B., Rahimi, A., Soltani, A., & Zeinali, E. (2020). Parameterization and evaluation of a simple simulation model (SSM-iCrop2) for Potato (Solanum tuberosum L.) growth and yield in Iran. Potato Research, 1–19. https://doi.org/10.1007/s11540-020-09456-y
FAO-UNESCO. (1974). Legend of the soil map of the world. FAO: Rome, Italy.
Fereres, E., Orgaz, F., & Gonzalez-Dugo, V. (2011). Reflections on food security under water scarcity. Journal of Experimental Botany, 62(12), 4079–4086.
Folberth, C., Skalský, R., Moltchanova, E., Balkovič, J., Azevedo, L. B., Obersteiner, M., et al. (2016). Uncertainty in soil data can outweigh climate impact signals in global crop yield simulations. Nature Communications, 7(1), 1–13.
Gijsman, A. J., Thornton, P. K., & Hoogenboom, G. (2007). Using the WISE database to parameterize soil inputs for crop simulation models. Computers and Electronics in Agriculture, 56(2), 85–100.
Godfray, H. C. J., Beddington, J. R., Crute, I. R., Haddad, L., Lawrence, D., Muir, J. F., et al. (2010). Food security: The challenge of feeding 9 billion people. Science, 327(5967), 812–818.
Han, E., Ines, A. V., & Koo, J. (2019). Development of a 10-km resolution global soil profile dataset for crop modeling applications. Environmental Modelling & Software, 119, 70–83.
Hengl, T., de Jesus, J., MacMillan, R., Batjes, N., & Heuvelink, G. (2014). SoilGrids1km—Global soil information based on automated. PLoS One, 9, e105992.
Jones, J. W., Hoogenboom, G., Porter, C. H., Boote, K. J., Batchelor, W. D., Hunt, L., et al. (2003). The DSSAT cropping system model. European Journal of Agronomy, 18(3–4), 235–265.
Kafatos, M. C., Kim, S. H., Lim, C.-H., Kim, J., & Lee, W.-K. (2017). Responses of agroecosystems to climate change: Specifics of resilience in the mid-latitude region. Sustainability, 9(8), 1361.
Koo, J., & Dimes, J. (2013). HC27 generic soil profile database. Harvard Dataverse Ver. 4. Washington, DC: International Food Policy Research Institute.
Liu, H., Pequeno, D. N., Hernández-Ochoa, I. M., Krupnik, T. J., Sonder, K., Xiong, W., et al. (2020). A consistent calibration across three wheat models to simulate wheat yield and phenology in China. Ecological Modelling, 430, 109132.
McCown, R., Hammer, G., Hargreaves, J., Holzworth, D., & Huth, N. (1995). APSIM: An agricultural production system simulation model for operational research. Mathematics and Computers in Simulation, 39(3–4), 225–231.
Nehbandani, A., Soltani, A., Rahemi Karizaki, A., Dadrasid, A., & Nourbakhsh, F. (2020). Determination of soybean yield gap and potential production in Iran using modelling approach and GIS. Journal of Integrative Agriculture, 19, 2–14.
Nouri, M., Homaee, M., Bannayan, M., & Hoogenboom, G. (2016). Towards modeling soil texture-specific sensitivity of wheat yield and water balance to climatic changes. Agricultural Water Management, 177, 248–263.
Romero, C. C., Hoogenboom, G., Baigorria, G. A., Koo, J., Gijsman, A. J., & Wood, S. (2012). Reanalysis of a global soil database for crop and environmental modeling. Environmental Modelling & Software, 35, 163–170.
Sharda, V., Handyside, C., Chaves, B., McNider, R. T., & Hoogenboom, G. (2017). The impact of spatial soil variability on simulation of regional maize yield. Transactions of the ASABE, 60(6), 2137–2148.
Soltani, A., & Sinclair, T. R. (2012). Modeling physiology of crop development, growth and yield. Wallingford: CABi.
Soltani, A., Alimagham, S., Nehbandani, A., Torabi, B., Zeinali, E., Dadrasi, A., et al. (2020a). SSM-iCrop2: A simple model for diverse crop species over large areas. Agricultural Systems, 182, 102855.
Soltani, A., Alimagham, S., Nehbandani, A., Torabi, B., Zeinali, E., Zand, E., et al. (2020b). Modeling plant production at country level as affected by availability and productivity of land and water. Agricultural Systems, 183, 102859.
Soltani, A., Alimagham, S., Nehbandani, A., Torabi, B., Zeinali, E., Zand, E., et al. (2020c). Future food self-sufficiency in Iran: A model-based analysis. Global Food Security, 24, 100351.
Van Ittersum, M. K., Cassman, K. G., Grassini, P., Wolf, J., Tittonell, P., & Hochman, Z. (2013). Yield gap analysis with local to global relevance—A review. Field Crops Research, 143, 4–17.
Wieder, W., Boehnert, J., Bonan, G., & Langseth, M. (2014). Regridded harmonized world soil database v1.2. ORNL DAAC.
WRB. (1998). World reference base for soil resources. World soil resources report no. 84 (p. 88). Rome: FAO.
WRB. (2006). A framework for international classification, correlation and communication. In World Reference Base for Soil Resources 2006 (142 pp). Rome: World Soil Resources Report No. 103, FAO.
Woli, P., Paz, J. O., Hoogenboom, G., y Garcia, A. G., & Fraisse, C. W. (2013). The ENSO effect on peanut yield as influenced by planting date and soil type. Agricultural Systems, 121, 1–8.
Zobler, L. (1986). A world soil file for global climate modeling, NASA Tech. Memo., 87802.
The present paper is part of a joint study by Agricultural Research, Education and Expansion Organization of Iran (AREEO) and Gorgan University of Agricultural Sciences and Natural Resources (GUASNR), Gorgan, Iran. The support from AREEO is highly appreciated. We are also thankful to Dr. Hamid Rahimian and Dr. Farhad Khormali for their support and discussion.
Conflict of interest
The authors declared that they have no conflict of interest.
Electronic Supplementary Material
Below is the link to the electronic supplementary material.
About this article
Cite this article
Nehbandani, A., Soltani, A., Taghdisi Naghab, R. et al. Assessing HC27 Soil Database for Modeling Plant Production. Int. J. Plant Prod. (2020). https://doi.org/10.1007/s42106-020-00114-4
- Food security
- Potential yield
- Irrigation water requirement
- Soil information