Regional modeling of winter wheat yield and water productivity under water-saving irrigation scenarios

Wheat is a major cereal crop in the world due to its importance as a main source of food and protein in human diets (Hernandez-Ochoa et al. 2018). Wheat (Triticum aestivum L.) is also the major and strategic field crop in Iran because approximately 45% of the total arable lands are under irrigated wheat production and contribute to approximately 70% of the total wheat production. Irrigated wheat has also the highest grain production among the cereals with an average yield of 3.45 Mg ha⁻¹ (Ministry of Jihad of Agriculture 2022). Therefore, improving the irrigated wheat production practices plays a critical role in food security and nationwide sustainable wheat production (Mehrabi et al. 2020; Davarpanah & Ahmadi 2021). In this regard, proper field management is essential to sustain wheat yield under limited water resources that could also be impacted by spatial and temporal changes in climate and environmental factors (e.g. rainfall) and their influence on wheat yield production (Hernandez-Ochoa et al. 2018; Li et al. 2022). Therefore, reliable grain yield estimation and modeling at a field scale could be a useful decision-making strategy for the agricultural sector's planners and stakeholders for narrowing or closing the yield gap (Dehkordi et al. 2020; Zhang et al. 2022). In this approach, accurate quantification of the relationship between water, soil, crop, and climate remains the main challenge to predict the yield at the field scale and develop field management strategies.

Achievement of the optimum yield is greatly affected by different agronomic practices and irrigation scheduling in various regional and climatic conditions (Mehrabi & Sepaskhah 2018; Solgi et al. 2022). Dynamic crop growth models are, however, widely used to determine the crop responses to various climates, crop species, and field managements (Hernandez-Ochoa et al. 2018). Most crop models are generally used to simulate soil water balance, plant phenology, root growth, biomass production, and crop yields (Krishna 2004; Kisekka et al. 2017). Crop models make it possible to predict the yield and have the ability to assess the factors affecting the yield under different scenarios (Krishna 2004; Hernandez-Ochoa et al. 2018). The Food and Agriculture Organization of the United Nations (FAO) developed and released the AquaCrop in 2009 as one of the most widely used crop models for a wide range of field and horticultural crops (Steduto et al. 2009; Ahmadi et al. 2015). AquaCrop is a water-driven, low-cost, and easy-to-use model to determine the response of the crop growth yield in different conditions. This model requires low input information and provides a reliable prediction of the related aspects of the crop yield (Steduto et al. 2009). There have been several researches on the successful and reliable application of the AquaCrop model to simulate the wheat yield during the last decade (e.g. Andarzian et al. 2011; Zhang et al. 2013; Iqbal et al. 2014; Jin et al. 2018; Davarpanah & Ahmadi 2021; Amiri et al. 2022; Solgi et al. 2022). These studies have confirmed the great and reliable ability of the AquaCrop model for simulating the in-season growth and final grain yield of wheat in diverse geographical locations and climatic conditions.

An important challenge in the AquaCrop application is its accuracy in large-scale and regional modeling of different crop cultivars via precise model parameterization since calibration and validation steps should provide consistency, uniformity, and repeatability in model predictions (Moriasi et al. 2012; Shirazi et al. 2021). Still, one essential feature of a crop growth model is its capability to reflect crop responses to water and environmental dynamics under diverse site conditions to properly assess climate impact on a regional scale (Hernandez-Ochoa et al. 2018; Wallor et al. 2018). In this regard, the application of crop models such as AquaCrop in climate studies requires large-scale investigations, and sufficient local and regional data of water, soil, and crop are essentially required in modeling (Beveridge et al. 2018; Wallor et al. 2018). So far, few studies have been done using AquaCrop on different wheat cultivars (Abrha et al. 2012; Zeleke & Nendel 2019; Shirazi et al. 2021) and climate (Manivasagam & Rozenstein 2020; Davarpanah & Ahmadi 2021; Solgi et al. 2022). Nevertheless, due to the complexities of crop responses to climatic conditions, crop model predictions still impose some degrees of uncertainties and challenges (Hernandez-Ochoa et al. 2018). Thus, a robust methodology is required to determine the model accuracy in the use of climate, soil, and crop management data over large areas (Webber et al. 2018; Manivasagam & Rozenstein 2020). One approach is to assess the accuracy of a calibrated crop model in different locations where the four major components of a cropping system, such as climate, cultivar, management, and soil, are different (Liu & Basso 2020).

Khuzestan and Fars provinces as the top-ranked irrigated wheat production areas are located in the southwest and south of Iran, which together produce approximately 28% of the national wheat production (Ministry of Jihad of Agriculture 2022). Recent studies have shown that climate change has led to increased air temperature and decreased rainfall in the southern parts of Iran, which has hugely reduced the amount of renewable groundwater and surface water resources (Mansouri Daneshvar et al. 2019; Vaghefi et al. 2019). Moreover, recent studies in these regions have shown that adapting appropriate field management policies under water deficit and climatic variability is essential for improving water productivity (WP) and sustainable wheat production in these areas (Davarpanah & Ahmadi 2021; Solgi et al. 2022). The objectives of this study were defined according to the concerns raised by the local farmers in the Khuzestan and Fars provinces regarding the impact of irrigation water shortages and rising air temperatures on wheat production. To address this concern, grain yield and irrigation WP of three commercially widely cultivated wheat cultivars were simulated in the south and southwest of Iran subject to different irrigation managements, sowing dates, and climatic conditions. Due to the current concerns of the farmers about the water resources in wheat production, climatic conditions in this study were defined based on the rainfall amount according to the statistical analysis of the long-term rainfall data in the study regions, and we did not intend to study the impact of future climate change using the scenarios of the Coupled Model Intercomparison Project (CMIP) (Touze-Peiffer et al. 2020).

The simulations were done using the fully calibrated and validated AquaCrop model in a large-scale modeling scheme to develop proper wheat cultivation policies. Recent studies have emphasized on shifting the sowing date and crop cultivars of winter wheat as a practical strategy for adaptation to climate change and climatic variability (Tao et al. 2014; Ojara et al. 2022; Rezaie et al. 2022). Moreover, regarding a more practical approach, instead of focusing on the extensive dataset collected from controlled small-scale research plots that are normally routine in scientific crop modeling, our modeling approach emphasized on small but spatialized datasets collected from large real irrigated winter wheat fields under local farmer managements (Harou et al. 2021; Pasquel et al. 2022). The results based on this modeling approach seem to resemble the real field conditions, and therefore, the outcomes of this study could offer a practical and useful guideline for the regional and local decision-makers, stakeholders, and practitioners in data-poor environments. In addition, the present study focuses on regional and large-scale modeling of irrigated winter wheat in multiple locations with diverse climatic conditions that offer the local farmers’ new practical approaches to deal with the water and heat stresses.

The wheat cultivars sown in the S1, S2, S3, and S4 fields were Mehregan, Chamran, Sirvan, and Sirvan, respectively. All these cultivars are locally bred for the hot to semi-hot climates. Mehregan is an early-maturing cultivar, is suitable for delayed sowing dates, and is resistant to terminal water and heat stresses. Chamran is a medium-maturing cultivar and is a high-yield cultivar under normal and sufficient irrigation. Sirvan is a semi-early-maturing and high-yield cultivar and is resistant to terminal water stress. Wheat seeds were sown at the rate of 200 kg ha⁻¹ from 20th Nov. until 5th Jan. in the studied fields. The Granstar herbicide was applied in 30–40 g ha⁻¹ in the studied fields around 80 days after the sowing date. The triple super phosphate fertilizer was used at the rate of 150–200 kg ha⁻¹ between the sowing date and flowering. Five soil samples per field were taken at the depths of 0–0.3, 0.3–0.6, and 0.6–0.9 m to determine soil textures and soil hydraulic properties (Table 1).

Irrigation is inevitable in the study regions due to erratic rainfall and its large spatial and temporal variations that do not address the wheat water requirement during the growing season (Davarpanah & Ahmadi 2021). Irrigation water was supplied from surface water in the S1 and S2 fields and from deep groundwater in the S3 and S4 fields. The electrical conductivity of irrigation water was less than 2.5 dS m⁻¹ in the four locations. Surface irrigation, as the common irrigation method in the four studied regions, was practiced to irrigate the fields. The local irrigation schedule fulfilled by the farmers is four to six irrigation events during the growing season depending on the water availability and rainfall amount and frequency in the regions (Table 2) (Davarpanah & Ahmadi 2021; Solgi et al. 2022). For instance, the growing season 2018–2019 was a wet year, and therefore lower irrigation amounts and less irrigation events were needed (Solgi et al. 2022). Moreover, all irrigation practices were carried out according to the local irrigation strategies of the farmers. The total applied irrigation volume during the whole growing season is given in Table 2. The average depth of applied irrigation water in each irrigation event varied between about 90 and 120 mm resulting from the applied irrigation volume and the number of irrigation events (Table 2).

Since at least 80% of the wheat root mass is distributed in the top 30 cm of the soil surface (Mehrabi et al. 2020), soil samples were taken from 0.3 m of the top soil to determine soil water content. Soil samples were collected six to seven times during the first growing season of 2016–2017. There were no soil water measurements in the second and third growing seasons. Aboveground biomass was measured three to four times during each growing season at two 0.5 × 0.5 m² random sampling areas per hectare. The biomass samples were oven-dried at 105 °C for 24 h and weighed. The final grain yield was measured based on four 1 × 1 m² random sampling areas in each field during each growing season.

The digital camera (model: SX700HS, Canon Inc., Japan) was used to monitor the canopy cover (CC). Photos were taken from 2.0 m height above the soil surface four to five times during each growing season in four replications per hectare. The ENVI image processor (version 5.1) was used to calculate the CC captured in each photo.

Soil properties, irrigation data, crop parameters, field conditions, and management parameters were used in AquaCrop simulation based on the calendar-day mode. Soil data were obtained from the local information (Zareian 2010). Daily measurements of temperature, humidity, and solar radiation were used in the model to determine daily crop evapotranspiration. We used the fully parameterized AquaCrop model that has been already calibrated and validated for wheat in the southern part of Iran (Andarzian et al. 2011). The conservative and non-conservative crop parameters are shown in Table 3. The parameterized AquaCrop model was used to simulate soil water content, CC, biomass, and grain yield for different wheat cultivars during the study period.

Based on the concerns raised by the local experts and farmers in the study regions, three scenario priorities were considered according to the existing local conditions and the major constraints of lack of water resources, seasonal variations in rainfall amount, and the increasing trend in air temperature. Therefore, the grain yield and WP were simulated under three scenarios of deficit irrigation levels, sowing dates, and climatic conditions. In total, three deficit irrigation managements, nine sowing dates, and three climate conditions were modeled in the four locations, which resulted in 324 combinations of scenarios and locations.

To schedule irrigation scenarios, it was necessary to consider the irrigation events, times and amounts of irrigation water in simulating the scenarios. Sowing, tillering, stem elongation, flowering, and seed filling stages are the sensitive growth stages that are affected by deficit irrigation (Goosheh et al. 2019; Davarpanah & Ahmadi 2021). Based on the previous studies, three irrigation levels, including 100% (d1), 75% (d2), and 50% (d3) of full irrigation, were considered as irrigation scenarios (Fooladmand 2012; Goosheh et al. 2019). The d2 and d3 are the water-saving and deficit irrigation management strategies.

Based on the local data in the four studied locations, nine sowing dates at 10-day intervals from 10 October to 30 December were considered as the earliest (T1) to the most delayed (T9) sowing dates, respectively (T1: Oct. 10; T2: Oct. 20; T3: Oct. 30; T4: Nov. 10; T5: Nov. 20; T6: Nov. 30; T7: Dec. 10; T8: Dec. 20; T9: Dec. 30).

Due to erratic rainfall variability under climatic conditions, the probabilities of exceedance of 20, 50, and 80% of annual rainfall were considered to represent wet, normal, and dry years, respectively, according to the historical rainfall data in each location. Rainbow software version 2.2 (Raes et al. 2006) was used to calculate the rainfall probabilities. The rainfall distribution in Figure 2 was used for generating the climatic scenarios.

The findings clearly show that the parameterized AquaCrop model could be a strong decision-making and analytical tool for simulating soil water content during a growing season. Since AquaCrop is structurally a water-driven model, accurate simulation of soil water content is essential for precise simulation of crop growth, biomass, and grain yield (Ahmadi et al. 2015). Geerts et al. (2009) and Ahmadi et al. (2015) reported that accurate simulations of soil water content would result in accurate simulations of crop evapotranspiration and biomass. Moreover, precise simulation of crop evapotranspiration is essential for reliable application of AquaCrop for irrigation water management purposes in water stress conditions to increase WP (Fernandez et al. 2020). Therefore, accurate simulation of soil water content will help farm managers and local experts to clearly monitor soil water dynamics for regional crop yield simulations.

A high canopy growth coefficient of Dezful (CGC = 8.65) caused high CC growth (Figure 4). Zhang et al. (2013) found that AquaCrop has more accurate results in simulating CC than biomass. Simulations of transpiration and biomass are followed by CC, and consequently, the errors in determining the CC are also affected on biomass (Raes et al. 2009). Therefore, this shows that accurate simulation of CC compared with the measured field data is useful for simulating biomass and yield. So, this is a strong skill of AquaCrop that precise parameterization of CC growth will be a strong step in accurate yield prediction. Since CC can be easily obtained through remote sensing methods and satellite images, this makes it very feasible to monitor the CC over the growing season to predict biomass and yield over large areas (Dalla Marta et al. 2019; Solgi et al. 2022).

The model had reliable and excellent final grain yield simulations for the three growth years (R² = 0.90; NRMSE = 0.08) (Figure 6). Previous studies have also reported the R² values ranging from 0.77 to 0.99 in the simulation of winter wheat grain yield by AquaCrop (Xiangxiang et al. 2013; Iqbal et al. 2014). In our study the maximum grain yield was observed in Dezful (S1), where the air temperature and air humidity were higher than in the other locations (Figure 2).

This is an important point that application of the AquaCrop model, which has been already parameterized (calibrated and validated) based on the small-scale experimental fields, could be up-scaled and used successfully for large areas. In this study, the AquaCrop model could simulate the main components of crop growth very well in different locations in the southern part of Iran, where wheat and other crops are cultivated extensively. This makes it feasible and applicable to model different field managements and environmental scenarios that would potentially help in improved management of the farmers’ fields to achieve higher grain yields and WP.

The AquaCrop showed that changing the sowing date would affect the wheat grain yield under different deficit irrigation levels and climate conditions. It is also reported recently that shifting sowing dates is a proper field management strategy to adapt to climatic conditions to maintain wheat grain yield production (Tao et al. 2014; Rezaie et al. 2022). Figure 7 shows that by shifting the sowing date from 10th Oct. to 10th Nov., grain yield increased in S1 and S4; and from 10th Nov. to 30th Dec. grain yield decreased. In S2 and S3, grain yield increased from 10th Oct. to 10th Dec. and decreased from 10th Oct. to 30th Dec. Thus, the optimum sowing date is close to the 10th Nov. in S1 and S4 locations, and the 10th Dec. in S2 and S3 locations. It is noteworthy that the sowing date scenarios were in line with the traditional practice of farmers in the study region. Therefore, it seems that delayed sowing compared with the current sowing practices would be beneficial for the farmers. A reason could be that crop establishment will coincide with cooler weather conditions and also a higher chance of effective rainfall during germination.

Former studies have shown the ability of AquaCrop in simulating grain yield under sowing date scenarios (Izadfard et al. 2016). Davarpanah & Ahmadi (2021) reported that 8th Nov. until 22nd Nov. is the optimal sowing date period for achieving the highest simulated wheat grain yield in different climate conditions. In this regard, Raja et al. (2018) reported the ability of AquaCrop to simulate the canopy cover, biomass, grain yield, and crop water use of maize with good accuracy under different sowing dates. In another study, sowing dates of sorghum were simulated in a regular 15-day interval based on the collected data from five meteorological stations from 1979 to 2013 (Alshikh et al. 2017). The model was used to investigate the optimum sowing date to maximize the grain yield. They indicate that the farmers’ choice of sowing date is an important adaptation strategy to the impacts of climate change and should be considered in the future studies of climate change impact on agriculture. It is suggested that future crop modeling studies consider the sowing dates under the scenarios of the CMIP (Touze-Peiffer et al. 2020).

However, it should be highlighted that delayed sowing may cause inappropriate pollination and seed germination under high temperatures and water deficits (Izadfard et al. 2016). Delay in sowing might also encounter the main crop growth stages with heat and water stress, which is a current real threat for crop production in the study area (Nouhi et al. 2008; Nazari et al. 2021). Raja et al. (2018) showed that the AquaCrop had moderately insufficient accuracy in predicting grain yield, biomass, WP, and crop water use with delayed sowing. This could be due to the effect of heat stress on senescence, which has been reported that AquaCrop cannot simulate senescence well (Hsiao et al. 2009; Zeleke et al. 2011; Ahmadi et al. 2015).

Figure 7 shows that normal and dry years had similar effects on simulations of grain yield under different deficit irrigations. There were significant differences between d2 (75%) and d3 (50%) in all cases (p-value = 0.015). These differences increased with reduced regional rainfall scenarios. Unlike wet and normal years, d1 compared with d2, and d2 compared with d3, had significant differences in dry years (p-value = 0.002). Simulations implied the higher negative impact of deficit irrigation levels on decreasing grain yield in dry years than in the other years. This shows that in dry years, it is essential to adapt regulated deficit irrigation rather than the conventional deficit irrigation to prevent significant yield loss. In this regard, recent studies have shown that adequate irrigation during the sensitive wheat growth stages may maintain grain yield in dry years (Davarpanah & Ahmadi 2021; Solgi et al. 2022). However, our simulation results showed that grain yield decreased under deficit irrigation managements in the study locations.

Mhizha et al. (2013) showed that the highest maize grain yields depended on the climate, variety, and soil water capacity. They showed that the late variety gave higher mean grain yields for all sowing dates. Hekmatzadeh et al. (2020) concluded that the warm seasons (spring and summer) are getting longer in the Khuzestan province (S1 and S2 locations), while the cold seasons (autumn and winter) are getting shorter. For example, at Ahvaz which is located between S1 and S2, autumn, winter, spring, and summer were 100, 90, 85, and 90 days, respectively, in 1967–1976, but they changed to 80, 90, 105, and 90 days in 2007–2016. These variations highlight the appropriate management of the interaction of irrigation management and sowing dates in terms of climatic variations in order to prevent grain yield loss. Our study pinpointed this issue that it is possible to model the interaction of several field managements and environmental conditions on crop growth and yield production. Due to high variability in large-scale field management, this is essential to model multiple factors for improved crop yield production under sustainable agriculture (Ewert et al. 2015).

In wet years, grain yield decreased from 10 to 15% in the 50% deficit irrigation (d3). The reduction of 25 and 50% of full irrigation reduced the grain yield by 10–18% and 18–27%, respectively, in normal years. In dry years, the grain yield decreased around 20–28% from d1 to d2 and around 30–45% from d1 to d3. It is noteworthy that the cultivated wheat cultivars are nationally bred for the arid and semi-arid areas, which should have suitable stability and high resistance in dry conditions. So, it seems that in plant breeding activities, a breeding target should be also the interaction of genotype × environment × management, which shows the importance of setting breeding targets based on future climatic conditions (Peng et al. 2020). The model simulations showed that 25% irrigation reduction had not a significant reduction in grain yield compared with the full irrigation. However, in the study regions, 50% deficit irrigation could be implemented with careful consideration of economic issues of yield impairment and minor losses.

Further analysis showed that optimum grain yield simulation in wet years is more than dry years. This could be due to accurate simulation of root growth pattern that would lead to better simulation of water uptake (Xue et al. 2003). Tsakmakis et al. (2019) showed that AquaCrop has an acceptable simulation of root growth pattern under different soil water conditions. Therefore, it seems that under wet and dry years, the model could have simulated root growth pattern well according to the simulated soil water in soil, which would influence grain yield under wet and dry years. In line with this argument, Thapa et al. (2020) found that wheat extracted more water from the deeper soil profile in wet years and increased the grain yield by about 64% than dry years. In the dry season, high vapor pressure deficit (VPD) would have increased soil surface evaporation, thereby reducing the amount of irrigation water infiltrating into the deeper root zone. Consequently, water extraction by plants and grain yield production is decreased. In wet seasons, however, irrigation could contribute more effectively to the grain yield. Not only the soil water storage at the sowing date, but also the uniformly distributed rainfall in the growing season is essential for gaining higher grain yield. Xiangxiang et al. (2013) found that the highest grain yield in AquaCrop was simulated for the scenarios in which irrigation was applied after the cold season and turning into green coverage stages. The simulations indicated that water can be withheld during green coverage or stem elongation stages without high reduction in the grain yield.

As a general rule, irrigation WP under deficit irrigation management is a critical balance between the grain yield (numerator) and the applied irrigation water (denominator). Therefore, WP would increase if grain yield is maintained (or with minor loss) under deficit irrigation. In this regard, AquCrop revealed diverse responses of the WP in the studied locations. Unlike grain yield (Figure 7), there were no regular changes in the WP under different climates and deficit irrigation scenarios. Based on Figure 8, there are 36 box plots attributed to the sowing dates in different climatic conditions and irrigation levels. The WP increased from d1 (100%) to d2 (75%) and from d2 to d3 (50%) in all cases. Figure 8 indicates the AquaCrop's skill in determining the optimum sowing date with higher WP at a regional basis. As it is discussed below, there are scenarios that changing the sowing date improved WP with increasing deficit irrigation levels.

In dry years, irrigation management is an important factor to reach high WP. The highest WP was achieved in Darab (S3) where d3 (50%) substantially increased WP. Decreasing irrigation level to d2 decreased WP compared with d1. However, WP increased in d3 over the study locations compared with the d2. Comparing the S1 and S2 locations in Khuzestan province with the S3 and S4 locations in Fars province implies that higher WP could be achieved under d3 in Fars province. However, a mutual comparison of Figures 7 and 8 reveals that although higher WP could be obtained under d3, a substantial decrease in grain yield would also happen. Therefore, field managers should tackle with this trade-off between high WP and low grain yield under 50% of deficit irrigation through proper field management. Economic WP would play a major role in proper decision-making and field management (Fernandez et al. 2020).

Nevertheless, WP was very different among the different sowing dates. Analyses showed that the 20th Dec. in Omidieh and Darab, and the 30^th Oct. in Dezful and Zarghan could be recommended as the sowing dates to obtain higher WP in dry years. Shifting the sowing date to 30th Dec. resulted in minimum WP in the study locations.

In normal years, the four locations showed very diverse responses to sowing dates and irrigation levels. Dezful (S1) and Omidieh (S2) showed different trends of WP fluctuations in deficit irrigation scenarios. While there were not many differences between WP values in the d1 and d2 in both locations, the WP increased sharply under d3 in S2 but decreased in S1 compared with d1 and d2. On the other hand, while deficit irrigation (d2 and d3) gradually increased WP in S3, the WP gradually decreased in S4 under deficit irrigation. These diverse and interactive responses to deficit irrigation clearly indicate the importance of proper planning for wheat cultivation under limited water resources and climate variability. Analyses showed that 10th Dec. in Omidieh and Darab, and 20th Oct. in Dezful and Zarghan could be recommended as the recommended sowing dates to obtain the highest WP in normal years.

In wet years, WP is influenced by rainfall distribution in these areas (Figure 2). AquaCrop showed an increasing trend of WP with decreasing irrigation water in Dezful (S1), Omidieh (S2), and Zarghan (S4). However, Darab (S3) did not follow this trend under d3 probably because of lower rainfall than the other three locations during initial growing stages and high grain yield loss under severe deficit irrigation (d3). A sharp decline in grain yield under deficit irrigations coupled with a sharp increase in WP in wet years shows that irrigation still plays the major role in maintaining grain production and rainfall cannot compensate for reduced irrigation water. This argument has been also confirmed and discussed in wheat field experiments (Mehrabi & Sepaskhah 2018).

The highest WP could be achieved on 10th Dec. in Omidieh and Darab, and 30th Oct. in Dezful and Zarghan in wet years. In general, the optimum sowing date is close to early December in Omidieh and Darab and late October in Dezful and Zarghan.

This study was basically initiated and conducted based on the queries raised by the local farmers and their concerns regarding irrigation water shortage and rising air temperature on wheat production. Therefore, the main practical challenges would be applying and transferring the knowledge in this study to the farmers through proper and extensive outreach and extension activities by the local stakeholders and managers. The local farmers are skillful enough to adapt to new sowing dates and apply deficit irrigations depending on the predicted rainfall conditions. In fact, in areas with erratic and uncertain annual rainfalls, irrigation strategies should be synchronized based on the amount and distribution of rainfall during the growing season in order to increase WP and sustainable grain production. In this regard, weather and rainfall forecast technologies would help in advance for the real-time irrigation scheduling (Solgi et al. 2022).

The parameterized AquaCrop model was used to simulate soil water content, canopy cover, biomass, and grain yield of different wheat cultivars in four different locations across the south and the southwest of Iran with different weather and climatic conditions. The model showed acceptable and reliable simulations of soil water content, canopy cover, biomass, and grain yield during different growing seasons. The results apparently showed that the model has high skills in simulating the major governing soil and crop factors (soil water content, canopy cover, biomass, and grain yield) of irrigated wheat growth, which makes it a strong tool for decision-making analysis in wheat production at multiple locations. A detailed scenario-based analysis using the parameterized AquaCrop showed that the model could be reasonably used to simulate wheat grain yield and WP under different sowing dates and irrigation managements subject to different rainfall amounts. Due to rising air temperature and reduced water resources, defining appropriate sowing dates and irrigation levels can help in sustaining wheat production. Since water and heat stresses are becoming a real threat for sustainable crop production, it is essential to employ crop models such as AquaCrop to study different management scenarios suitable for each location.

Overall, the analysis implied that AquaCrop is a strong analytical decision-making tool to determine the optimum sowing date in different climatic conditions for achieving the optimum WP of irrigated winter wheat. Future large-scale applications of AquaCrop could be aimed at up-scaling from regional-scale to national-scale simulation of irrigated and rainfed winter wheat.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.