基于改进权重湿润指数的东北地区春玉米涝渍灾害评估方法

Evaluation Method of Spring Maize Waterlogging Disaster in Northeast China Based on Improved Weight Moist Index

【目的】研究东北地区春玉米涝渍灾害评估方法,为气候变化背景下东北地区春玉米合理生产和管理提供科学依据。【方法】以权重湿润指数为基础,采用作物需水量代替参考作物蒸散量,构建改进权重湿润指数;运用1986—2015年东北地区春玉米生育期内涝渍灾害事件与改进权重湿润指数耦合方法、基于K-S检验的正态分布显著性检验方法和t-分布区间估计方法等,确定基于改进权重湿润指数的东北地区春玉米不同生育阶段涝渍强度分级阈值;从农田土壤相对湿度与改进权重湿润指数相关性和拟合关系检验,改进权重湿润指数、权重湿润指数评估结果与历史灾情结果一致性检验,以及典型年份涝渍灾变过程动态评估准确性检验3个方面,对改进权重湿润指数评估春玉米涝渍灾害的精度进行验证;分析东北地区春玉米不同生育阶段涝渍强度空间格局,揭示其规律特征。【结果】(1)改进权重湿润指数与不同深度的土壤相对湿度均达到极显著相关(P<0.001),除玉米快速发育期外,其他生育期内二者相关系数最大值均出现在20 cm深度土壤相对湿度;将不同生育阶段改进权重湿润指数与20 cm深度土壤相对湿度间进行Slogistic曲线拟合,决定系数(R2)最大值出现在生育中期(0.46),其后依次是生育后期(0.34)、快速发育期(0.31)和初始生长期(0.21),均方根误差(RMSE)最小值出现在生育后期(0.49),其后依次是生育中期(0.51)、初始生长期(0.52)和快速发育期(0.56),不同生育阶段拟合曲线中90%的20 cm土壤相对湿度所对应的改进权重湿润指数值最小值出现在初始生长期(0.77),其后依次是快速发育期(1.12)、生育中期(1.21)和生育后期(1.25)。(2)与随机预留的春玉米涝渍灾害事件测试样本中灾情判定结果相比,权重湿润指数评估结果的完全准确率为26.7%,部分准确率为66.7%;改进权重湿润指数评估结果的完全准确率为66.7%,部分准确率为93.3%。(3)以嫩江、白城和沈阳3个代表站点1998年春玉米生育期内涝渍灾害为例,对比降水、权重湿润指数和改进权重湿润指数随时间变化的过程曲线,发现改进权重湿润指数所反映的当年涝渍发生时间和强度等级与灾情资料更为相符,而其大小还可直观表征涝渍强度动态变化。(4)春玉米不同生育阶段各强度等级的涝渍频率差别较大,黑龙江中部、吉林东南部、辽宁东南部为中度及以上涝渍频率的高发区,生育中期是中度及以上涝渍频率高发期。【结论】基于改进权重湿润指数的东北地区春玉米不同生育阶段涝渍强度分级阈值划分方法结果可靠,可表征农田涝渍实际发生状况;相对于权重湿润指数而言,改进权重湿润指数不仅对春玉米涝渍灾害评估准确率更高,而且更适于对涝渍灾变过程的跟踪和动态评估,对于揭示气候变化背景下东北地区春玉米涝渍灾变机制和时空演变规律具有重要意义。

【Objective】The aim of this study investigated the evaluation methods of spring maize waterlogging disasters in Northeast China,so as to provide a scientific basis for the rational production and management of spring maize in Northeast China under the background of climate change.【Method】Based on the Weight Moist Index(WMI),the crop water demand,instead of the reference crop evapotranspiration,was used to construct an Improved Weight Moist Index(IWMI).The coupling method of waterlogging disaster events and IWMI during the growth period of spring maize in Northeast China from 1986 to 2015,the normal distribution significance test method based on K-S test,and the estimation method of t-distribution interval were used to determine the thresholds for waterlogging intensity grading of different growth stages in spring maize.The correlation and fitting relationship of soil relative humidity and IWMI,the consistency of the evaluation results on IWMI and WMI and waterlogging disaster statistics results,and the accuracy of the assessment on dynamic process of waterlogging disasters in typical years were analyzed to evaluate the IWMI precision.The spatial pattern of waterlogging at different growth stages of spring maize in Northeast China was analyzed,and then its regular characteristics were revealed.【Result】(1)IWMI was significantly correlated with soil relative humidity at different depths(P<0.001).Except for the rapidly developmental period of maize,in the other stages,the maximum correlation coefficient of both appeared at 20 cm depth.Through fitting the IWMI and the soil relative humidity at 20 cm depth based on a Slogistic model,it was found that the maximum coefficient of determination(R2)appeared in the middle period(0.46),followed by the end period(0.34),the rapidly developmental period(0.31)and the initial growth period(0.21).The minimum root mean square error(RMSE)appeared in the end period(0.49),followed by the middle period(0.51),the initial growth period(0.52),and the rapidly developmental period(0.56).The values of the IWMI corresponding to soil relative humidity of 90%at 20 cm depth were 0.77 in the initial growth period,1.12 in the rapidly developmental period,1.21 in the middle period,and 1.25 in the end period.(2)Compared with the disaster diagnosis results in the test samples of the spring maize waterlogging disaster events randomly reserved,the complete accuracy rate and partial accuracy rate of the WMI evaluation results were 26.7%and 66.7%,respectively,while the complete accuracy rate and partial accuracy rate of the IWMI were 66.7%and 93.3%,respectively.(3)Taking the waterlogging disasters during the growth period of spring maize in 1998 in Nenjiang,Baicheng and Shenyang stations as examples,changes of the precipitation,WMI and the IWMI with the days of the year were compared,and it was found that the time of occurrence and intensity level of the waterlogging evaluated by the IWMI were more consistent with the disaster records,and IWMI could also be used to represent the dynamic change of the waterlogging intensity.(4)The frequency of waterlogging of the different intensity levels in spring maize had a great difference.The central part of Heilongjiang,southeastern Jilin,and southeastern Liaoning were high-incidence areas with moderate and above waterlogging,and the middle period was a period of high frequency of moderate and above waterlogging.【Conclusion】Based on the IWMI,the method of determining the thresholds of waterlogging intensities at different growth stages of spring maize in Northeast China was reliable,which could be used to characterize the actual occurrence of farmland waterlogging.Compared with the WMI,IWMI not only had a higher accuracy,but also was more suitable to track and evaluate the dynamic assess of waterlogging disasters,which was of great significance to reveal the catastrophic mechanism and spatio-temporal evolution of spring maize waterlogging disasters in Northeast China under the background of climate change.