Identification of climate change induced heat stress sensitive 1 environments and prediction for diverse representative concentration 2 pathways – A novel approach for tracking hotspots

54 Climate change is unequivocal across economies and India owing to its distinct geography has been 55 exposed to several climatic risks, especially in agriculture. Heat stress is a serious environmental 56 problem posed by climate change, wherein mean temperature is expected to increase relatively more 57 during wheat growing season affecting production and food security. In the milieu , the investigation is 58 pioneered to predict heat stress sensitive wheat growing environments in India for research 59 prioritization using a long term (30 years) historical daily data. The study has developed a 60 methodological approach by integrating statistical downscaling of climate information and principal 61 component analysis for computing heat stress intensity index (HSII) for 17 experiment locations across 62 wheat growing environments. HSII were estimated for existing locations post testing for Levene’s 63 homogeneity of variance, followed by prediction for three periods’ viz., early-future (2026-2050), mid- 64 future (2051-2075) and far-future (2076-2100) under two emission scenarios namely RCP4.5 and 65 RCP8.5. The results alarmed a radical shift in HSII of experiment locations from one period to another 66 in both scenarios. Experiment locations with high index values for the existing environment has moved 67 almost to lower category in the early future and subsequently shifted to higher position in the mid-future 68 and far-future. The investigation also found that under projected RCP4.5, trial locations in peninsular 69 zone need more emphasis, whereas in RCP8.5, peninsular zone coupled with central zone and north 70 eastern plains zone have to be focused. Overall, the study develops a pragmatic approach in location 71 prioritization across predicted periods which can be replicated to other regions. On policy front, rational 72 allocation of research funds has been suggested to carry out field trials on climate change induced heat 73 stress sensitive environments for sustaining the national wheat production apart from developing micro- 74 level adaptation strategies to counter adverse effects of climate change.


Introduction 79
Intergovernmental Panel on Climate Change (IPCC, 2007) predicts an increase in temperature 80 from 0.3°C to 0.7°C over the next two decades and an increase of 0.3°C to 4.8°C by the end of this 81 century (Collins et al., 2013; Kirtman et al., 2013). Wheat is a cold adored cereal plant and 82 temperature plays a detrimental role for its productivity. Despite the effects of climatic aberrations 83 vary with location, some studies reported a yield reduction of 10% with only 1°C increase in 84 temperature above normal (Brown, 2009), while some reports indicate 3-7% reduction in wheat 85 yield for every 1°C increase in temperature in the range of 15°C-21°C (Wardlaw et al., 1989;86 Hatfield et al. 2008). The concern here is the future pattern of climate is much uncertain to plan the 87 investment decisions for research. A galore of literature indicates that the crop is highly sensitive to 88 the change in climatic variables in different growth stages . Relative higher 89 temperature during the crop growth season generally accelerate the phenological progress, 90 affecting the plant's routine metabolic activities that impacts the yield (Lobell and Gourdji, 2012;91 Rezaei et al., 2015). Increase in the atmospheric demand for water has also been reported when 92 the temperature is high, further causing reduction in the water-use efficiency of the crop (Ray et al., 93 2002). 94 95 Stress tolerance is a very complex trait and gets aggravated with the changing climate. It not only 96 depends on the severity of the cause, but also on the sensitivity of developmental stage and 97 duration of stress. Climate change induced heat stress reduces the plant growth and development 98 by disrupting the process of photosynthesis, respiration, net assimilation rate as well as total plant 99 dry weight (Wahid et al., 2007). Photosynthetic inhibition is a universal indicator of environmental 100 stress. Temperature variations beyond a threshold level also causes growth inhibition occurred due 101 to impaired stomatal activity, respiratory losses, bleaching of plant pigments etc. In general, 102 change in plant architecture, coleoptiles elongation, stunted growth, scorching of leaves and stems 103 and senescence are some well documented responses which have been observed under elevated 104 temperatures leading to reduced productivity (Tian et al., 2009). During initial stage early vigor, 105 coleoptiles length, canopy cover plays an important role in stabilization of a field crop. But, 106

137
The study is marshalled by the daily timeseries climate data (January 01, 1984 to December 31, 138 2013) obtained from the India Meteorological Department (IMD) for selected locations and tested 139 for the homogeneity of variance following Levene's (1960) approach. In India, wheat is being 140 cultivated across five diverse agro-climatic zones (Figure 1)  *Growing degree days (GDD) has been estimated till the crop maturity adopting the approach of Shaykewich (1995). 170 171 As a preliminary step, quality check was done for the collected daily data on maximum and 172 minimum temperature as well as rainfall for 17 experiment locations. It was carried out by visual 173 observation of time plot, climatology, outliers identification (if any) followed by computing the 174 summary statistics and then tested for homogeneity of variance to proceed for further analysis. 175

Testing the Homogeneity of Variance (Levene's approach)
Levene's approach was employed to check for the homogeneity in variance since the test is 176 relatively less sensitive to departure of the time series from normal distribution. The test (Levene, 177 1960) was carried out with the following assumption: 178 Null Hypothesis (H0) : σ1 2 = σ2 2 = … = σk 2 179 Alternate Hypothesis (Ha) : σ1 2 ≠ σ2 2 ≠… ≠ σk 2 180 Assume a climate time series data (Y) with sample size 'N' that comprise 'k' sub-groups with Ni 181 being the i th sub-group's sample size. The test statistic (W) given by Levene is as follows:    The trend in climate variables viz., maximum temperature, minimum temperature and rainfall was 218 estimated using Mann-Kendall's non-parametric test approach. It is a widely adopted method to 219 identify the long-run trends in time series data (Ahmad et al., 2015) by comparing the relative 220 magnitude of sample observations against the actual observations. It is a rank-based procedure 221 with an assumption that only one data exists at a point of time, robustness to the influence of 222 outliers as well as suitable for skewed time series possessing non-linear trend (Hamed, 2008;223 Helsel and Hirsch, 1992). The advantage of the technique is that it does not require any particular

231
The collected data on climate variables viz., rainfall, maximum and minimum temperature and 232 rainfall were normalisedto make them unit and scale freefor comparison (

..(1) 236
It is well known that rainfall reduces the heat stress. Owing to the negative association, the formula 237 presented in Eq. 2 has been used for normalising the rainfall variable. 238

Normalisation = (Maximum value -Actual value) / (Maximum value -Minimum value)
...(2) 239 where, Ej represents the Eigen value of the j th factor, and Lij indicates the loading value of the i th 256 variable on j th factor. 257

258
The weights for each variable for all the selected 17 trial locations were fitted in the following 259 formula (Eq. 5) to arrive the composite heat stress intensity index (HSII) value for each location. Levene's test results indicated that all variables turned to be significant (Table 2) rejecting the null 275 hypothesis to establish the fact that there exists a significant difference in variance among 276 observations. This led to proceed with the estimation of trend coefficient (Mann-Kendall's test). 277

279
* denotes the statistical significance at one per cent level of probability.

Level of Heat Stress Intensity in Wheat Producing Environments: Existing Scenario 281
Globally, wheat production is expected to fall by 6 per cent for each degree of temperature

317
The long-term trend in climatic variables as explained in Table 2 (Table  368 7). However, with respect to rainfall, Durgapura, Hisar, Indore, Niphad and Pune registered 369 negative statistic. Low level of rainfall will lead to drought stress and affects the crop productivity 370    (Table 9)

411
Perusal of Table 9 and 10 indicates that the trend in climatic variables is expected to undergo a 412 major shift both in RCP 4.5 and 8.5 scenarios. Barring a few locations like Dharwad, Durgapura, 413 Hisar, Indore and Kanpur, the rest found to have an increasing trend in minimum temperature for 414 all the periods (Table 10). Dharwad, Jabalpur, Kalyani, Parbhani, Pune and Ranchi have shown a 415 raising maximum temperature in all periods considered for the study. In the case of rainfall, 416 Dharwad, Junagadh, Kanpur, Karnal, Ludhiana, Niphad and Pune had positive trend in the existing 417 scenario and they are expected to have higher quantum of rainfall for the predicted period as well 418 (Table 9). Barring Dharwad, all locations registered positive trend in minimum temperature all the 419 periods under consideration (Table 11)