Analyzing the impact of climatic conditions on rainfed wheat yield in northwest Iran: a parametric and nonparametric approach

The sustainable development of agricultural production is essential because of factors like population growth and the depletion of resources available for agriculture in highly populated countries. Reducing food losses as a result of extreme weather and climate change is one way to reach this goal. Nonetheless, yearly fluctuations in the climate pose a special risk to agricultural productivity (Gourdji et al. 2015; Wang et al. 2019; Eini et al. 2023b). However, as a result of the numerous physical, biological, and environmental systems that have been impacted by global warming brought on by human activity, agriculture has always been regarded as the most vulnerable industry (De Costa et al. 2007; Halkos & Tsilika 2017; Zhang et al. 2018; Zarei & Khaledi-Alamdari 2023).

One of the challenges facing the agricultural community is figuring out how to use weather information to implement risk management strategies, which will increase preparedness and decrease vulnerability to climate change (Meinke et al. 2003). Agriculture is a very risky business because of the significant weather fluctuations in arid and semi-arid regions, particularly in developing nations. To manage the impact of risks, climate risk management in these areas entails forecasting potential risks associated with climate change and offering solutions that can significantly reduce these risks (Ramesh 2010; Tomczyk et al. 2022).

To determine the optimal timing for agricultural operations and thereby maximize output, it is crucial to evaluate the risks posed by climate variability and extreme weather events to agricultural production (Andre et al. 2007). Therefore, researchers are increasingly interested in studying how climate-related factors impact the production of agricultural products. Consequently, there is a need to raise public awareness of the evaluation and prediction of these factors. Many factors that affect the climate are relevant to agricultural management and food security, particularly when it comes to rainfed agriculture in arid regions.

Among the most significant rainfed crops, wheat is a strategic plant vital to human nutrition and a staple food in many societies. Due to its widespread use and cultivation, this plant makes up the majority of all cultivations worldwide. As a result, its cultivated area is estimated to be around 17% of all cultivated area, and it provides 40% of the world's food supply (Giraldo et al. 2019; Tolwani & Shukla 2022). Climate variability has a major global impact on crop productivity. Remarkably, a number of researchers have reported that this variability causes a 60% change in yield (Ray et al. 2015; Matiu et al. 2017; Eini et al. 2023a).

Furthermore, rainfed agriculture is extremely susceptible to extreme weather events and changes (Zampieri et al. 2017; Ribeiro et al. 2019a), including drought, global warming, and heavy rainfall. As a result, extreme weather conditions like drought are thought to be a major threat to agricultural systems, with the effects being especially apparent in rainfed crops (Páscoa et al. 2017). Significant economic ramifications can result from quantitative and qualitative changes in variables linked to the success of strategic investments, which may also give rise to pervasive social issues. Rainfed crop yields are influenced by weather conditions and CO₂, which are the most unpredictable (Szyga-Pluta et al. 2023; Marcinkowski & Piniewski 2024). Accurately estimating the weather in target areas aids farmers and planners in planting on schedule and providing for the plant's needs throughout the growing season. The development of quantitative and qualitative agricultural products may be particularly affected by this.

Furthermore, changes in agricultural crop yield in the region can be accurately predicted by determining and quantifying the impact of each climate variable. Plant performance, development, and growth conditions are significantly impacted by water stress. This is a severe issue in arid and semi-arid regions, where rainfall varies annually, and plants must withstand extended periods of water scarcity (Boyer 1982). As a result, it is evident that understanding how weather affects variations in crop yield is crucial (Faghih et al. 2021).

Analyzing climate trends using temperature and precipitation integration indices to identify and categorize droughts is essential for understanding how weather affects changes in crop yield (Vangelis et al. 2013). The most popular technique for researching droughts uses indicators that allow one to assess the drought's spatial and temporal characteristics as well as its relative intensity (Heim Jr 2002; Zargar et al. 2011). Agriculture's risk of drought, defined as the potential for crop failure due to drought conditions, frequently causes considerable financial losses (Skakun et al. 2016; Xie et al. 2018; Khaledi-Alamdari et al. 2023). Drought in the context of agriculture is characterized as a severe and ongoing lack of rainfall that drastically lowers agricultural output relative to normal (Warren et al. 1996), as well as meteorological drought, which is the outcome of the region's natural decline in rainfall over a specific period of time. It can occur over several years in a row or over a short period and is frequently represented by the difference between the annual and average precipitation totals. Compared to other natural disasters, drought is one of the most costly natural disasters in the world, impacting a large number of people and societies (Keyantash & Dracup 2002). Governments, regional and international organizations, and non-governmental organizations have focused especially on drought preparedness and policy during the past few decades (Wilhite 2000). Consequently, large-scale and occasionally local plant performance has been predicted using meteorological variables like temperature and precipitation (Bannayan et al. 2003; Palosuo et al. 2011). Farmers can take advantage of favorable climatic conditions to double production or make better decisions in the face of weather-related risks by accurately modeling multivariable weather distribution.

This study examined the variables that better justify the changes in rainfed wheat yield using meteorological and climatic parameters, including precipitation and temperature. The variables used were analyzed parametrically and nonparametrically to clarify the effects. The role of droughts during the studied period was assessed using the widely recognized standardized precipitation index (SPI) and the standardized precipitation–evaporation index (SPEI), which were compared to rainfed wheat yield. For the first time, this type of analysis provides very important and significant results for agricultural decision-making and management, as well as the discussion about food security.

To evaluate the role of each phase of wheat growth in relation to yield, the values of heat units during six phases are used, which consist of (1) germination, (2) completion of germination to onset of flowering, (3) flowering stage, (4) completion of flowering to the start of seed filling, (5) seed filling stage, and (6) total growing season.

Meteorological droughts occur before other types of droughts (Eini et al. 2023c), and their occurrence is considered a necessary condition for other droughts (Mirabbasi et al. 2012; Eini et al. 2023d). In the case of meteorological drought, the basis for calculating the degree of dryness is determined by comparing the amount of precipitation with the long-term average or its normal value. To assess and monitor climatic droughts, researchers worldwide use different methods, which are not accessible everywhere and are sometimes not efficient enough for other regions.

In the current study, the SPI was used to assess meteorological droughts. The SPI proposed by McKee et al. (1993) was calculated for each region based on the long-term precipitation occurring in each region (Edwards 1997). Basically, the SPI was developed to detect the lack of precipitation at multiple time scales. These time scales show the specific impacts of drought on the accessibility of different water sources. Instead of using just precipitation, as the SPI does, there seems to be a recent trend toward using the SPEI, which makes use of both temperature and precipitation (Dikshit et al. 2021). The SPEI is one of the indicators of drought, which can be determined at different scales by the difference between precipitation (P) and potential evapotranspiration or reference transpiration (ET0). This includes the influence of temperature on soil moisture, which affects plant growth.

To check the presence of trends in the variables under study such as temperature, precipitation, and yield, the Mann–Kendall test (Mann 1945; Kendall 1948) is considered the most widely used nonparametric testing method (Nikhil Raj & Azeez 2012; Zarch et al. 2015; Leng & Hall 2019; Szyga-Pluta et al. 2023).

Kendall's τ is a coefficient that indicates ordinal correlation and is used to determine the dependence criterion between random variables. This coefficient is generally used to determine copula functions.

According to Figure 6, the proportion assigned to each stage, in descending order, refers to the following: the stages from the completion of germination to the beginning of flowering, from the completion of flowering to the beginning of seed filling, the flowering phase, the seed filling phase, and germination.

The changes in annual precipitation, especially during the growing season of rainfed wheat, are evident when comparing the yield of rainfed wheat in each growing season during periods when precipitation decreased significantly. The yield of rainfed wheat also reacted to this and fell significantly. Figure 7 shows the yield compared to growing season precipitation during the study period. In the growing season 2007–2008, when the lowest precipitation was recorded during the growing season, the yield of rainfed wheat (detrended) in this region was also lower at the lowest level, 103 kg/ha. This is despite the fact that in the case of the highest precipitation of the growing season in the region, which was measured at 348.5 mm, the yield of rainfed wheat was not the highest, which could indicate a very strong response of the yield of rainfed wheat to small amounts of atmospheric precipitation.

A recent study focused on the response of wheat and barley plants to different drought indicators in Spain. It showed the differing efficiencies of several drought indicators and highlighted the importance of the multiscale feature of droughts, especially the SPEI (Peña-Gallardo et al. 2019). The study by Ribeiro et al. (2019b) demonstrates the efficiency and appropriateness of using the SPEI in jointly analyzing the risk of crop yields and drought events.

Since many parameters play a role in determining the yield of rainfed wheat, it is essential to specify the parameters that play a significant role in the variations of rainfed wheat yield in the region. In Table 1, according to the parameters mentioned in the previous sections, their parametric and nonparametric correlations with the yield of rainfed wheat in three different states of normal yield, detrended yield, and standardized yield are given, to which the values refer to annual rainfall. Precipitation during six growing seasons of rainfed wheat, including germination (S1), completion of germination to start of flowering (S2), flowering stage (S3), completion of flowering to start of seed filling (S4), seed filling stage (S5), and the entire growing season (GP), is given in the table above. Based on the results obtained, the highest parametric and nonparametric (ordered) correlation of precipitation is related to the total growing season precipitation at normal yield and detrended yield. The Kendall correlation values for growing season precipitation relative to normal yield and detrended yield are estimated at 0.396 and 0.366, respectively, both significant at the 1% level.

Furthermore, the studied indices have an appropriate and acceptable correlation with respect to rainfed wheat yield, and Kendall's correlation for the SPI and SPEI with respect to normal yield are 0.396 and 0.253, respectively. These values suggest a detrending yield of 0.366 and 0.361, respectively. Based on the results presented in Table 1, four time series of total growing steps, seasonal precipitation, and SPI and SPEI were obtained.

This study provides a novel approach to rainfed wheat yield estimation by analyzing the specific impacts of precipitation during different growth stages, which is crucial for regions that rely heavily on rainfed agriculture. The key finding is the identification of the second growth phase – from the completion of germination to the beginning of flowering – as critically influential, with fall precipitation emerging as a particularly strong predictor of final yield. This early-stage precipitation monitoring offers a new perspective for agricultural planning and food security management.

The study's use of SPI and SPEI to correlate drought events with yield changes further contributes to the understanding of how seasonal precipitation patterns influence rainfed wheat production. This research is innovative in its detailed dissection of precipitation's role across different growth stages and its implications for more accurate yield forecasting in the face of climate variability.

The novelty of this manuscript lies in its focused examination of precipitation effects across distinct growth stages of rainfed wheat, particularly emphasizing the critical role of early-stage precipitation. This study introduces a novel methodology by separating precipitation occurrences based on the GDD and using both parametric and nonparametric correlation analyses to evaluate their impact. The implications of these findings are significant, offering actionable insights for early-stage agricultural interventions and more effective food security strategies in regions where rainfed agriculture is predominant.

Numerous factors influence rainfed wheat yield, and understanding these factors is crucial for improving crop yield estimation and forecasting. By analyzing precipitation during different growth stages using commonly used drought indicators, researchers can determine the impact of precipitation on rainfed wheat yield. This study highlights the significant role of precipitation during specific growth stages, particularly from the completion of germination to the beginning of flowering. However, it is important to acknowledge that all growth stages are critical; insufficient precipitation in the early stages cannot be compensated by more precipitation later on, underscoring the importance of the early stages.

The study also examined the seasonality of precipitation – specifically fall, winter, and spring precipitation – during the growing season of rainfed wheat in the Tabriz region. The results indicate that autumn precipitation plays a more significant role than other seasons, showing strong correlations with the final rainfed wheat yield. Early monitoring of fall rainfall can, therefore, be a valuable tool for predicting wheat yield, which is essential for planning and managing food security.

To account for drought events during the growing season, the study utilized the SPI and SPEI, both of which effectively explained variations in rainfed wheat yield in the region. Given that rainfed crops rely heavily on precipitation, drought events directly impact yield, and changes in these indices correspond to changes in crop yield.

Future research could benefit from comparing the findings of this study with similar studies in other regions or climates to assess the broader applicability of the results. Additionally, exploring new or improved methods for yield prediction, such as integrating remote sensing data or advanced machine learning techniques, could provide more accurate forecasts. Studies could also investigate the potential of other environmental factors, technological advancements, and socio-economic variables in influencing rainfed wheat yield, offering a more comprehensive understanding of the factors at play.

The authors thank the University of Tabriz for supporting this research. M.R.E. was supported by National Science Centre (Narodowe Centrum Nauki – NCN, PRELUDIUM BIS-1 project, UMO-2019/35/O/ST10/04392, Poland).

M.K.-A. and A.F.-F. conceptualized, analyzed, and prepared the figures. M.K.-A. and A.M.-H. prepared and provided a literature review and methodology investigation. A.F.-F. and M.R.E. supervised and validated the results. All authors reviewed and participated in the writing of the manuscript.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.