Impact Of Climate Change On The Phenology Andyield Of Rice In China And The Optimization Of Sowing Dates

China is the largest producer and consumer of rice and the rice production accounting for 40%of China’s total grain production in 2021.At present,the annual average temperature in China showed a significant upward trend,higher than the global average level during the same period,which is a sensitive area for global climate change.It is predicted that climate change,mainly characterized by global warming,will have adverse effects on rice.Planting date shifts of rice is an adaptive strategy that low-cost and easy to implement,regarded as an effective measure to mitigate the impact of climate change.Therefore,study the changes in rice growth and yield in China,identified key variables that affect rice yield,analyzed the impact of planting date shifts on rice phenology and yield have great significance to ensure the food supply of China.This paper divides the main cultivation area of rice in China into four single season rice regions and two double season rice regions.The season rice regions include Dongbei region,Changjiang region,Chuanyu region and Xinan region,two double season rice regions include Jiangnan region and Huanan region.The historical period（1961-2020）and two future periods（2021-2060,2061-2100）under the scenarios of SSP1-2.6,SSP2-4.5,SSP3-7.0,and SSP5-8.5 was considered,study the changed characteristics of climate resource.Based on the Decision Support System for Agricultural Technology Transfer（DSSATv4.7.5）,to simulate the phenology and yield of rice,analyze its spatiotemporal changes,then use Copula entropy（CE）to determine the correlation between rice yield and thirteen characteristic variables.Based on partial least squares（PLS）,generalized linear model（GLM）,generalized additive model（GAM）and random forest model（RF）,the relationship between characteristic variables and rice yield was established.Shapley value（SHAP）,partial functional dependency diagram（PFD）and variable importance（VIP）were used as evaluation indicators to explain the process of model construction,to infer the impact of climate change on rice yield statistically.We set four different irrigation schemes to compare the impact of different irrigation amounts on rice,then based on Irrigation Scheme II,to simulate the growth and yield of rice on a 5-day planting gradient using the DSSAT model,to explore the impact of planting date shifts on the growth and yield of rice.The main results are as follows:（1）The daily maximum temperature(T_max),daily minimum temperature(T_min),and daily average temperature(T_mean)all showed an upward trend,and the increasing range from largest to smallest is T_max,T_mean and T_min.Under the scenarios of SSP1-2.6,SSP2-4.5,SSP3-7.0 and SSP5-8.5,the temperature increased during the rice growth period became increasingly significant,with values of 0.21℃/10a,0.29℃/10a,0.36℃/10a and0.46℃/10a,respectively.The annual precipitation(P_re)in each region showed an upward trend,and the increased range of P_re in the rice growth period under the four future scenarios is 11.2 mm/10a,9.8 mm/10a,8.8 mm/10a and 14.6 mm/10a,respectively.During the historical period,the daily solar radiation（SRAD）of rice growth period showed a decreased trend,while in the future,the SRAD fluctuant increased.（2）Compared with historical periods,the anthesis date and maturation date of single cropping rice and early rice showed a significant trend of advance,while the anthesis date of late rice showed a significant trend of delay,and the days of change increased with the increase of scenario radiation forcing.Without considering the impact of CO₂,under the SSP1-2.6,SSP2-4.5,SSP3-7.0 and SSP5-8.5 scenarios,the yield of single season rice decreased by 3.0%,9.1%,16.8%and 18.4%,respectively,while the yield of early rice decreased by 2.7%,9.0%,18.5%and 17.9%,and the yield of late rice decreased by 10.9%,19.6%,30.5%and 33.8%,respectively.When considered the CO₂ efficiency of fertilizer,the yield of single season rice increased by 3.2%,2.0%-0.2%and 0.2%,respectively,while the yield of early rice increased by 7.9%,7.8%,4.6%and 9.5%,respectively,and the yield of late rice decreased by 2.3%,3.9%,8.3%and 8.8%,respectively.The fertilizer effect of CO₂ can basically compensate for the negative impact of climate change on single cropping and early rice,but it cannot compensate for the negative impact on late rice.（3）The simulation effect of nonlinear models on rice yield was significantly better than that of linear models and the simulation effect of RF was best.The model had lower simulation result for the extreme value of rice yield,and higher simulation result for the peak yield part.CO₂,SRAD,LAIXS and GDD contributed significantly to rice yield.The soil moisture content,CO₂,SRAD,and LAIXS are positively correlated with yield,while GDD,KDD_high,T_min,T_mean and T_max are negatively correlated with yield,and the correlation on rice yield increased with the increase of radiation forcing.The contribution of KDD_lowand P_re to the yield was not significantly.（4）The yield of single season rice and early rice increased with the advance of planting date,while the yield of late rice increased with the delay of planting date.The change tend to the optimal planting date will prolong the growth period of rice.The extension of the growth period for early rice and single season rice mainly comes from the extension of the vegetative growth stage,while late rice mainly comes from the extension of the reproductive growth stage.The optimal planting dates for single season rice under SSP1-2.6,SSP2-4.5,SSP3-7.0 and SSP5-8.5 scenarios were 40 days,30 days,30 day and 40 days in advance,respectively.The optimized planting windows were（-45,15）,and the yield increased by 2.5%,2.6%,2.8%and 3.1%,respectively.Compared with historical periods,they increased by 6.0%,4.9%,2.8%and 3.5%,respectively.The optimal planting dates for double season early rice were 30 days,35 days,30 days and 35 days in advance,respectively.The optimized planting window is（-45,5）,（-45,5）,（-45,5）,（-45,5）and（-40,-10）,respectively.The yield increased by 2.5%,3.5%,4.5%and 7.4%,respectively,which are 10.1%,10.9%,8.8%and 16.8%higher than historical periods.The optimal planting dates for double season late rice were postponed by 10 days,20 days,25 days,and 30 days,respectively.The optimized planting dates windows were（-10,30）,（-5,35）,（5,35）and（15,40）,respectively.The yield increased by 1.4%,3.1%,4.8%and 11.7%,respectively and decreased by 0.2%,0.3%,2.7%and-4.2%compared to historical periods.The variation for rice yield caused by planting date shifts is in descending order of late rice,early rice,and single cropping rice,the yield change increased with the increase of radiation forcing.After adjusting to the optimal planting date,the yield of single cropping rice and early rice increased compared to historical periods.Except for the SSP5-8.5 scenario,the yield of late rice still decreased in other scenarios.Adjusting to the optimal planting date can compensate for the negative impact of climate change on single cropping and early rice,while cannot compensate for late rice.At the national level,adjusting to the optimal planting date can increase rice yield.This study provides adaptive measures and recommendations for government departments and local decision-makers to address the impact of future climate change.