Based on projections of the Intergovernmental Panel on Climate Change (IPCC, 2014) it is expected that the average surface temperature of the world will increase by 1.5–4.5°C by 2100 if the recent greenhouse gas emission rates continue. Hansen et al. (2012) estimated an increase in average universal temperature of 0.18°C per decade. IPCC (2014) reported 2000s as the warmest decade and 2014 as the warmest year. Agricultural productivity is under threat due to climate change all over the globe as predicted by current and future scenarios (van Ogtrop et al., 2014; Ahmad et al., 2017; Dettori et al., 2017). Increased temperature significantly influences crop productivity and can lead to dramatic yield loss. Climate change in the form of increased temperature has reduced cereal yield up to 17% (Lobell et al., 2011). Pakistan is also facing impacts of climate change in different socio-economic sectors (Akram and Hamid, 2015; Ahmad et al., 2017). The warming trend, predicted to be 2–4°C at the end of century, will drastically affect agriculture (Rasul et al., 2012). Dwivedi et al. (2017) reported yield loss associated with high temperatures in Eastern Indo-Gangetic Plains (EIGP). In their experiments the average temperature during anthesis of late and very late sowing conditions were 2.7°C and 5.2°C higher than that of the optimum sowing which resulted in the reduction in grain filling duration by 8.6–12.6 days, and average yield loss of 18 and 34% for late sown and very late sown conditions, respectively.

Global warming is producing a shift in crop phenology over the globe (Root et al., 2003; Parmesan, 2006; Ahmad et al., 2017). A warmer climate accelerates crop phenology and affects crop yield (Madan et al., 2012; Zhang et al., 2013; Ahmed et al., 2016). Craufurd and Wheeler (2009) reported earlier flowering and crop maturity under high temperature conditions. Higher temperature resulted in a shorter growing season, accelerated leaf aging/senescence and reductions in biomass and yield (Tao et al., 2013; Figueiredo et al., 2015). High temperature had a negative effect on biomass, spike length, grain yield and harvest index (Plaut et al., 2004).

Wheat production in the rainfed area of Pothwar Pakistan is highly vulnerable to higher temperature (Ahmed and Hassan, 2015). The wheat-growing season in this area starts with the onset of winter and ends in April-May. Due to low soil moisture, wheat sowing is delayed which results in high temperature at anthesis and during grain filling. This high temperature can reduce final yield, and the probability of high temperature is being exacerbated by global warming (Dias and Lidon, 2009, 2010; Modarresi et al., 2010). Joshi et al. (2007) reported heat stress as a major abiotic stress due to delayed sowing affecting wheat cultivation in South Asia. The optimum temperature range for wheat during anthesis and grain filling is between 12 and 22°C. Wheat exposure to above 30°C at pre- and post-anthesis stages reduces the grain filling rate and thereby decreases grain yield and quality (Barnabas et al., 2008). High temperature (>30°C) during anthesis is deleterious (Langer and Olugbemi, 1970; Tashiro and Wardlaw, 1990), while short exposure to very high temperature (>35°C) during anthesis and grain filling can result in substantial reduction of grain yield in wheat (Randall and Moss, 1990; Stone and Nicolas, 1994). Hussain and Mudasser (2007) conducted an experiment in the rainfed area of Pakistan to investigate the effect of high temperature on growing season length. They reported that a 1.5°C increase in temperature shortened the crop life cycle and lowered grain yield by 7%. However, a temperature increase of 3°C reduced wheat yield by 24%.

A prerequisite for modeling crop growth is good forecasting of plant development, with air temperature recognized as the principle regulator of plant development (Bewick et al., 1988). Growing degree day (GDD) values are utilized for predicting crop development (Yang et al., 1995; McMaster and Wilhelm, 1997), which usually integrate temperature values on a daily basis. With increasing global temperature, faster accumulation of GDDs will take place. In addition to temperature, photoperiod is also a major environmental factor determining time to flower initiation and first flower appearance in plants (Daba et al., 2016), and flowering in wheat is sensitive to photoperiod. Thus, plant development is a function of temperature and daylength (Slafer and Rawson, 1994; Cao and Moss, 1997), and both should be included in predictive algorithms.

To investigate temperature and photoperiod effects on wheat development, we artificially created different temperature and daylength conditions for wheat by varying the sowing times. Early-sown wheat encountered more cold days but late-sown wheat experienced higher temperatures, especially during anthesis to maturity. To support the projection of changes in phenology in responses to future global warming, we used the experimental results to evaluate the ability of GDD and photoperiod algorithms to predict wheat crop phenology.

In this study, we saw a reduced number of days to maturity due to higher temperature, which has been reported by other authors, e.g., Sadras and Monzon (2006), who concluded that wheat phenology has a strong correlation with temperature. Temperature effects at three different study sites resulted in reduced number of days to maturity. The differences in phenology were due to variation in climatic parameters that prevailed during the crop life cycle at these sites (Table 2), in addition to the effect of higher seasonal temperature resulting from delayed sowing. Our results were in agreement with the findings of Asseng et al. (2011), Deryng et al. (2014), Hawkins et al. (2013), and Lobell et al. (2011). Wolkovich et al. (2012) reported that a change in the phenology of a crop is the most important indicator of climate change. Furthermore, Wang et al. (2015) suggested adjusting the sowing date to minimize the impact of heat stress, as well as using late-maturing cultivars to maintain crop phenology patterns that utilize resources more effectively. Using crop models to design ideotypes could be a good approach to predict the best genotypes for a target environment under climate change (Gouache et al., 2015). Since phenology is a key component in ideotype design, matching crop growth stages to optimal growing conditions is necessary to utilize resources effectively and to avoid climatic stresses such as the terminal heat stress observed in our studies (Foulkes et al., 2011; Sylvester-Bradley et al., 2012). Gomez-Macpherson and Richards (1995) concluded that phenology is an important aspect of adaptation and maximization of yield as phenology controls the duration of critical growth stages of the crop which has associations with abiotic stresses and further partitioning of assimilates. Accurate prediction of phenology is a very important step to calibrate different crop models (Ceglar and KajfeŽ-Bogataj, 2012; Archontoulis et al., 2014; Ahmed et al., 2016). Models can capture cultivar variations which affect leaf area development, biomass production and grain yield only if they can capture crop phenology accurately. Phenological parameters such as days to maturity are directly linked to crop productivity. Under normal conditions the more days to maturity, the higher is the yield. The higher seasonal temperature obtained in this study by sowing wheat late reduced days to maturity during both years. The Wang and Engel (1998) GDD method with photoperiod adjustments predicted days to maturity accurately, suggesting that spring wheat phenology is sensitive to both temperature and photoperiod. Zhang et al. (2015) argued that wheat genotypes were highly sensitive to photoperiod.

Crop biomass is closely associated with life cycle duration, and temperature is a dominant environmental factor influencing crop development (Liu et al., 2016). Camargo et al. (2016) reported that the relative timing of flowering and grain filling has a direct influence on wheat productivity. In our studies, a decrease in biomass due to higher temperature is consistent with earlier findings that higher heat accelerates crop phenology and crop senescence, which results in earlier crop maturity, lower biomass and lower crop yield (Zhao et al., 2007; Prasad et al., 2008; Hatfield and Prueger, 2015; Prasad and Jagadish, 2015). With higher temperature leading to earlier maturity and lower biomass production, it is critical to consider changing both sowing date and genotype under future warming. In all earlier studies a reduction in growing season length was associated with higher temperature, so current varieties may not be suitable in the future due to shortened growing seasons (Xiao et al., 2010; Asseng et al., 2015; Wang et al., 2015).

An increase in daily atmospheric temperature will accumulate more degree days and hasten maturity. Observed days to maturity varied significantly with atmospheric temperature. An increase in seasonal atmospheric temperature of 2–4°C (e.g., Wheeler et al., 2000) can shorten the crop life cycle, which ultimately reduces yield (Batts et al., 1997; Ahmed and Hassan, 2015). Semenov (2009) used a modeling approach and future climate projections and found that phenology is reduced to 14–16 days due to higher temperature. In our study, higher temperature had a negative correlation with biological and grain yield under all treatments.

For wheat, the Wang and Engel GDD method with photoperiod adjustment (for the period from emergence to anthesis) was effective in predicting days to anthesis and days from anthesis to maturity as a function of changing sowing dates and associated seasonal temperatures. This provides indication that this approach can be useful for predicting crop phenology under global warming scenarios.

Spring wheat phenology is driven by both temperature and photoperiod. Increasing temperatures have a significant effect on wheat crop phenology, biomass production, grain yield, and harvest index. High temperatures occurring in a period 7 days before and after anthesis affected grain numbers (data not presented) and yield but it could not be isolated from the effect of seasonal high temperatures. The Wang and Engel method for the calculation of GDD, coupled with an algorithm to adjust GDD as a function of day length extracted from the APSIM crop growth model, provided good estimates of number of days to anthesis across all sites, sowing dates and years. The Wang and Engel GDD method also provided excellent estimates of the number of days between anthesis and maturity. The good predictive capability of the algorithms tested suggests that they are adequate for estimation of global warming effects on wheat phenology.

MAA conducted field experiments, collected, and analyzed the data, CS and MA designed the idea. MAA, MA, CS, SH, and RH wrote and finalized the manuscript; MA, CS, and SH conceived, directed the project, and approve of the final version of the manuscript. FH did editing of the article.