基于气象灾害的ORYZA(V3)模型对双季稻发育期模拟的适应性评价

摘要

To evaluate the simulation accuracy of the ORYZA(V3)model for development stage of the double cropping rice in Hainan Province,the parameters of the ORYZA(V3)model were calibrated and validated based on daily meteorological observations,meteorological disaster data,soil properties and observed rice phenology data from four typical experiment sites(Haikou,Danzhou,Ledong and Qionghai)in the double cropping rice planting area during 2005-2014.Different types of meteorological disasters and their frequencies during the development stages of the double cropping rice were counted,and the disasters with the highest frequency in each growing phase were selected.Then the simulation accuracy of the model was evaluated based on each single meteorological disaster.The results show that the ORYZA(V3)model could reproduce the development stages of the double cropping rice in Hainan island well,the deciding coefficient R2 was more than 0.90 and NRMSE(normalized root mean square error)ranged from 3.97％to 9.80％.High temperature occurred most frequently during the development stages of double cropping rice,followed by typhoon and drought disaster.Under typhoon condition,the simulation accuracy for flowering stage of the late rice was good with R2=0.90 and NRMSE=3.90％,while it was unacceptable for other growing stages.Under high temperature condition,R2 ranged from 0.87 to 0.89 for early rice and from 0.18 to 0.61 for late rice,and NRMSE ranged from 3.49％to 5.71％for the double cropping rice.The R2 was larger than 0.87 under drought condition with NRMSE ranging from 3.11％to 9.73％.The evaluation results can be valuable for better application and optimization of the model in the future.%为评价ORYZA(V3)模型在海南岛双季稻发育期模拟的适应性,利用2005-2014年海南岛双季稻区4个站点(海口、儋州、乐东、琼海)的逐日气象数据、气象灾害资料、土壤、水稻发育期等观测资料,对模型进行调参与验证,本地化不同品种水稻发育期参数;统计双季稻各个发育期出现的气象灾害及其次数,筛选出各个发育期内出现次数较多的气象灾害.以单独的气象灾害为背景,对各个发育期的模拟与实测结果进行对比验证.结果表明:ORYZA(V3)模型对海南岛双季稻发育期的模拟精度较高,决定系数R2>0.90,归一化均方根误差NRMSE为3.97％～9.80％;双季稻发育期内出现的气象灾害次数由多到少依次为:高温、台风、干旱;ORYZA(V3)模型对气象灾害的敏感性从大到小依次为:台风、高温、干旱.在台风背景下,仅晚稻开花期的R2为0.90,NRMSE为3.90％,其他发育期的模拟均在误差范围外;在高温背景下,早稻的R2为0.87～0.89,晚稻的R2为0.18～0.61,双季稻的NRMSE为3.49％～5.71％;在干旱背景下,R2>0.87,NRMSE为3.11％～9.73％.评价结果在模型应用和优化方面具有一定的参考价值.