Effect of intra-seasonal temperature on wheat at different locations of India: A study using CERES-Wheat model

Wheat (Triticumaestivum L.) is the second main source of world’s food energy and nutrition. It supplied about 19 and 21% of the total calories and proteins, respectively, required by the world’s population (Anonymous, 2011). Although, wheat can be grown in a wide range of climatic conditions but many abiotic and biotic factors limits its yield. Most of the wheat growing areas of the world experience above optimum temperatures at some point in their life cycle and have a large negative impact on productivity. Continual heat stress affects nearly 7 million hectares of wheat in developing countries, whereas terminal heat stress is a problem in 40% of the temperate world that encompasses 36 million hectares (Reynolds et al., 2001). During March 2004, temperatures were higher in the Indo–Gangetic plains by 3-6oC and as a result, the wheat crop matured earlier by 10-20 days and wheat production dropped by more than 4 million tonnes in the country (Aggarwal, 2008). The objectives of the present study was first to investigate the effect of rise in intra-seasonal temperature on wheat productivity in different wheat growing zones of India. Secondly, to identify the most critical phases of crop growth during which the above normal temperature causes maximum damage. Thirdly, to identify the optimum sowing time as mitigation strategy against above normal temperature.

Wheat (Triticumaestivum L.) is the second main source of world's food energy and nutrition. It supplied about 19 and 21% of the total calories and proteins, respectively, required by the world's population (Anonymous, 2011). Although, wheat can be grown in a wide range of climatic conditions but many abiotic and biotic factors limits its yield. Most of the wheat growing areas of the world experience above optimum temperatures at some point in their life cycle and have a large negative impact on productivity. Continual heat stress affects nearly 7 million hectares of wheat in developing countries, whereas terminal heat stress is a problem in 40% of the temperate world that encompasses 36 million hectares (Reynolds et al., 2001). During March 2004, temperatures were higher in the Indo-Gangetic plains by 3-6ºC and as a result, the wheat crop matured earlier by 10-20 days and wheat production dropped by more than 4 million tonnes in the country (Aggarwal, 2008).
The objectives of the present study was first to investigate the effect of rise in intra-seasonal temperature on wheat productivity in different wheat growing zones of India. Secondly, to identify the most critical phases of crop growth during which the above normal temperature causes maximum damage. Thirdly, to identify the optimum sowing time as mitigation strategy against above normal temperature.

MATERIALS AND METHODS
The wheat yield and meteorological data of seven major wheat growing locations were collected from respective station (Table 1). It may be seen that there was large spatial variation in maximum and minimum temperatures across the locations (Fig. 1).
The CERES-Wheat model was calibrated using the field experimental data for the most popular wheat variety (Table 2) using the recommended crop management practices Effect of intra-seasonal temperature on wheat at different locations of India: A study using CERES-Wheat model for the respective stations.The genotypic coefficients were derived by repeated iterations until a close match between simulated and observed phenology and yield was obtained.
The performance of the model was validated with the experimental data from the years which were not used for calibration of the model and was used for calculation of these parameters/ indices and for validation of the model.
The statistical measures such as the root mean square error (RMSE) and normalized root mean square error (NRMSE) were calculated following Jamieson et al., (1991).The d-stat or index of agreement (D-index) was estimated following Willmott et al.,(1981).
After validation, the CERES-Wheat model was used as a research tool to study the effect of rise in intra-seasonal temperature on wheat productivity. At all the locations a similar set of treatments, i.e., three dates of sowing (early, normal and late) with a difference of 15 days were selected. The normal date of sowing corresponds to the time when most of the farmers undertake sowing of wheat crop at the respective locations. The model was used to simulate normal yield using normal weather data along with soil and crop data of the respective locations. Above normal temperature scenarios were created by increasing both maximum and minimum temperatures from normal in the increments of 1.0°C up to 3.0°C. The temperature was increased for 15 days duration, keeping the normal temperature during rest of the crop growth period. The increase in temperature was imposed at fortnightly intervals between 1 st December to 30 th April. The model output in terms of wheat yield was used to calculate per cent deviation in wheat productivity from the respective time of sowing under different above normal temperature scenarios.

Validation of the CERES-Wheat model
The results (Table 3) shows that the phenological   events matched quite well. The difference between observed and simulated days taken for anthesis at various locations varied from -6 to +4 days, except at Faizabad. Similarly, the difference in observed and simulated days taken for maturity was -6 to +3 days. The r 2 and d-statvalueswere acceptable.
The NRMSE values at all the locations for days taken for anthesis and maturity, were in excellent range (<10%) as proposed by Jamieson et al., (1991). The grain yield for wheat crop was simulated quite satisfactorily and the differences between observed and simulated values were in the range of -498 to +754 kg ha -1 . The values of r 2 was above 0.73 and was nearly one for some places except Palampur (r 2 =0.28). The value of d-Stat was quite acceptable and was above 0.60 at all the locations.The validation results showed that the model calibration was quite satisfactory and may be used for further applications.

Effect of temperature rise on wheat yield
The calibrated model was used to study the effect of rise in intra-seasonal temperature and the results are presented in the following sections.
Northern hills zone : At Palampur, representing northern hill zone, the rise in temperature by 1.0°C from normal was harmful during mid January to end March for early, mid February to end March for timely sown wheat and mid January to mid April for late sown wheat (Table 4). Temperature rise during January to mid February was beneficial for the timely sown wheat. In this zone the timely sown crop was benefited / less affected due to rise in temperature as compared to other sowing.
The rise in temperature by 2.0°C from normal during February and March, second fortnight of March and mid February to mid April in case of early, timely and late sown wheat, respectively, reduced its productivity ( Table 5). The rise in temperature during January to mid March for the timely sown and during January to mid February for late sown wheat increased their productivity. In this zone, with a rise in temperature by 2.0°C from normal, the productivity of timely sown wheat was more benefited while that of late sown wheat was less adversely affected.
The rise in temperature by 3.0°C from normal was harmful during second fortnight of January and March in case of early sown, second fortnight of March in timely sown and mid March to mid April in case of late sown wheat ( Table 6). The increase in temperature during January to mid March increased the yield of the timely and late sown wheat. The temperature rise by 3.0°C in case of late sown wheat during initial stages proved beneficial, but that towards end proved harmful for productivity of wheat, however, timely sown wheat remained less affected by the rise in temperature.
North western plain zone : At Ludhiana, representing north western plain zone, the rise in temperature by 1.0°C form normalduring January to February, mid January to mid March and from February to mid March in case of early, timely and late sown wheat, respectively, led to reduction in wheat yield (Table 4). However, temperature rise by 1.0 °C during December to January was beneficial (1.8-5.2%) for late sown crop. In North Western plain zone the productivity of late sown wheat was increased/less affected due to increase in temperature by 1.0°C from normal.
The rise in temperature by 2.0°C from normal during January and February in early sown, mid January to mid March in timely sown and February to midMarch in late sown wheat, decreased the productivity by 2.93-5.68% (Table 5). However, a rise of 2.0°C during 16-31 December for timely sown and during December to January for late sown led to an increase in productivity. In this zone, the productivity of late sown wheat was less affected by a rise of 2.0°C from normal temperature.
The rise in temperature by 3.0°C from normal during December to end February, mid January to mid March and February to end March reduced the yield of early, timely and late sown wheat, respectively (Table 6). However, temperature rise of 3.0°C during December to mid January for timely sown crop and during December to end January for late sown crop was beneficial. In this zone, the late sown wheat was benefited by the rise in temperature during early phases and was adversely affected during later phases of crop growth.
North eastern plain zone : Kanpur, Faizabad and Ranchi were the three stations under this zone. At Kanpur, the temperature increase by 1.0°C from normal during mid January to end February for early, 1-15 December and mid January to end February for timely and from February to mid March for late sown wheat led to reduction in wheat productivity (Table 4). However, temperature rise during December to mid January for timely, mid December to mid January for timelyand mid December to end January for late sown crop was beneficial.At Faizabad, temperature rise by 1.0°C from normalduring January in case of early, first fortnight of December and January and whole of February in case of timely and duringsecond fortnight of December and February in case of late sown wheat led to a reduction in its productivity.At Ranchi, the rise in temperature Effect of temperature on wheat at different location in India December 2016 by 1.0°C from normal, during December to March for timely sown and February to mid March for late sown wheat reduced its yield. However, wheat yield was increased with rise in temperature during December to mid Januaryandmid Decemberto end January in early and late sown wheat, respectively. In this zone the late, early and early sown crop at Kanpur, Faizabad and Ranchi, respectively, will be benefited/ less affected due to rise in temperature by 1.0°C than normal.
The temperature rise by 2.0°C form normal during second fortnight of December and during mid January to end February in early sown crop and from February to mid March for timely and late sown wheat led to yield reduction (1.0-5.78%) (Table 5). However, the rise in temperature during first fortnight of January increased wheat productivity (2.10-9.15%) of wheat sown on all dates. At Faizabad, the rise in temperature by 1.0°C form normal during December to mid February in early sown wheat, from December to end February in timely sown wheat and during mid December to mid March in late sown wheat decreased its productivity. At Ranchi, the rise in temperature by 1.0°C from normal during January to mid February in early sown, first fortnight of December and mid January to end February in timely sown and mid January to mid March in late sown crops, reduced the wheat productivity. In this zone, at Kanpur and Faizabad, with a rise in temperature by 2.0°C from normal, the late sown wheat was benefited / less affected and at Ranchi the productivity of early sown crop was less affected as compared to timely and late sown wheat.
The temperature rise by 3.0°C from normal during December and mid January to end February in early sown, mid January to mid March in timely sown and mid January to end March in late sown wheat led to a reduction in its productivity. However, temperature rise during December to mid January increased the yield of timely sown wheat and that during mid December to mid January enhanced the productivity of late sown wheat.At Faizabad, the yield of early sown wheat was reduced if the temperature was increased by 3.0°C from normal during December and mid January to mid February. The yield of timely sown wheat decreased if the temperature increased during December and mid March. At Ranchi, temperature increase by 3.0°C from normal during January to mid February in early sown, during 1-15 December and January to end February in timely sown and during January to mid March in late sown wheat reduced the its productivity. However, the productivity of late sown wheat was benefited with temperature increase during second fortnight of December. In this zone, at Kanpur, Faizabad and Ranchi the yield of timely, late and early sown wheat, respectively, was benefited / less affected by increase of 3.0°C as compared to other sowings.
Central zone : Raipur and Udaipur were the two station & under this zone. At Raipur, the wheat yield decreased with temperature rise by 1.0°C from normal during December to February in early sown,mid December to mid March in timely sown and during mid December to end March in late sown wheat (Table 4). At Udaipur, the temperature rise during December and January in early sown, December to mid February in timely sown and during January and February in late sown crop led to a reduction in wheat productivity (Table  4). However, the increase in temperature during December was beneficial for the late sown crop.In this zone, with an increase in temperature by 1.0°C from normal,the timely sown wheat at Raipur and late sown wheat at Udaipur was less affected as compared to other dates of sowing.
The productivity of early sown wheat was reduced if the temperature increased by 2.0°C from normalduring December to February (Table 5). The productivity of timely sown wheat was reduced with rise in temperature during December to mid March at Raipur. The yield of late sown wheat was reduced with rise in temperature during 16-31 December, mid January to end March. At Udaipur the wheat productivity of early, timely and late sown crop was reduced by increase in temperature during December and January, January to mid February and January to end February, respectively (Table 5). However, the increase of 2.0°C from normal temperature during December benefited the timely sown crop. In this zone, the yield of early sown wheat at Raipur and that of late sown wheat at Udaipur remained less affected with a rise in temperature by 2.0°C from normal.
The temperature increase by 3.0°C from normal during December to mid January and February in case of early sown, during December to end March in case of timely sown and during December to mid January and February to mid April in case of late sown wheat reduced its productivity.At Udaipur, the temperature rise by 3.0°C from normal from December to end January, January to end February and December to mid March reduced the yield of early, timely and late sown wheat, respectively (Table 6). Temperature rise during December increased the yield of timely sown wheat. In this zone, the early and timely sown wheat at Raipur and Udaipur, respectively, remained less affected as compared other dates of sowings due to temperature increase by 3.0°C from normal. Simulation results revealed that above normal temperature did not affect the wheat yield in the same way in all the four zones under study. The effects also varied widely within the zones and also with the sowing time. The positive / negative impacts of high temperature on wheat yield increased with itsintensity.
Amongst the zones, central zone seems to be the most vulnerable to rise in temperature (3.0°C from normal), as wheat productivity decreased almost throughout the crop growth period for wheat sown on all the dates. Maximum reduction in yield (10.2%) was also noticed in this zone in case of late sown wheat at Raipur. On the other hand, in Northern hillszone the productivity of wheat increased upto 4.07% with temperature rise by 3.0°C from normal during early phases of crop growth. Thereafter, the yield decreased (3.46-6.84%) with rise in temperature by3.0°C, which was less as compared to other zones. Amongst, the seven locations the differential response of wheat yield to similar (3.0°C) rise in temperature may be due to varietal and temperature difference (normal growing season temperature of timely sown wheat at Raipur was 21.0°C while at Palampur was 12.7°C). Talukder et al., (2014) reported heat stress can cause wheat yield reduction upto 25% and the effect varies with the varieties. Early heading wheat genotypes with slower rate of leaf senescence after heat exposure and longer post-heading duration could be more tolerant to heat stress.
Amongst the sowing dates, the yield of late sown wheat in Northern hills and North Western plain zones increased more due to heat stress (3.0°C above normal) during early phases of crop growth and were less reduced with heat stress in later phases of growth.The yield of timely, late and earlysown wheat at Kanpur, Faizabad and Ranchi, respectively, in North Eastern plain zone and early and late sown wheat at Raipur and Udaipur in the Central zone suffered to a lesser extent as compared to other date of sowings. Asseng et al., (2011) found that a variations in average growing-season temperatures (±2 °C) in Australia can cause reductions in wheat production up to 50% and this can be due to increased leaf senescence as a result of temperatures >34°C. Lobell et al (2012)in northern India used nine years of satellite measurements of wheat growth to monitor rates of leaf senescence due to exposure to temperatures > 34°C and found a statistically significant acceleration of senescence from extreme heat. In our study at some places the increased temperature scenario was also near or above 34°C.
Among the times of above normal temperature by 3.0°C,second fortnight of March in Northern hills zone, February in and North Western plain zone,February at Kanpur, mid January to February at Faizabad and mid January to mid February at Ranchi (North Eastern zone) was most critical as during this period temperature rise caused reduction in productivity of wheat sown on any date. In Central zone,February at Raipur and January at Udaipur was most critical as maximum reduction in productivity of wheat sown on any date occurred during this period.
Over the zones, except Northern hills zone first fortnight of February month was most critical as above normal temperature by3.0°C decreased the productivity of wheat sown on any date.In North western plain zone above normal temperature by3.0°C during early vegetative stages favoured the yield of timely and late sown wheat.The optimum temperature for growth and yield of wheat is about 18-24°C, even short periods (4-6 days) of very high temperature (35-40°C) significantly decreases grain yield (Stone and Nicolas 1994;1995). In many wheat growing areas of India, wheat experiences 35°C during grain development which reduce theyield. Lobellet al., (2008) had also estimated yield losses of 3-17% for each degree rise in temperature in northwest India and Pakistan. Prasad and Djanaguiraman (2013) reported that mean daily high temperatures >30°C for short periods (5days), when imposed from start of heading, caused a linear decrease in grain number and when the stress was imposed after seed-set, it caused a quadratic decrease in grain weight. High nighttime temperatures >20°C during the reproductive phase decreased grain filling duration and grain weight.

CONCLUSIONS
The present study indicates that the temperature rise had differential effects on wheat yield in different zones and the effects also varied with the sowing time. The results showed that the positive and negative effects of above normal temperature on wheat productivity increased with the level of stress. Amongst the zones, Central zoneis the most susceptible and North hillszone being the least susceptible to temperature rise (3.0°C from normal).Amongst the sowing dates, the productivity of late sown wheat in Northern hills and North Western plain zones; timely, late and early sown wheat at Kanpur, Faizabad and Ranchi, respectively, in North Eastern plain zone and early and timly sown wheat at Raipur and Udaipur, respectively, in the Central zone suffered to a lesser extent as compared to other date of sowings.Over the zones, barring Northern hills zone the temperature during first fortnight of Februarywas most critical to wheat productivity. In Northern hills zone, second fortnight of March was most important as during this period the above normal temperature caused a reduction in productivity of wheat irrespective of dates of sowing under study.