当前位置: X-MOL 学术J. Oper. Oceanogr. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Forecasting the severity of the Newfoundland iceberg season using a control systems model
Journal of Operational Oceanography ( IF 3.1 ) Pub Date : 2019-06-20 , DOI: 10.1080/1755876x.2019.1632128
Grant R. Bigg 1 , Yifan Zhao 2 , Edward Hanna 3
Affiliation  

ABSTRACT

The iceberg hazard for the Grand Banks area to the east of Newfoundland varies dramatically from one year to the next. In some years no icebergs penetrate south of 48°N, while in others well over 1000 icebergs enter the main shipping lanes between Europe and NE North America. Advance knowledge of this seasonal hazard would have major implications for ship routing, as well as the resources required for maintaining an effective ice hazard service. Here, a Windowed Error Reduction Ratio control system identification approach is used to forecast the severity of the 2018 iceberg season off Newfoundland, in terms of the predicted number of icebergs crossing 48°N, as well as to hindcast iceberg numbers for 2017. The best estimates are for 766 ± 297 icebergs crossing 48°N before the end of September 2017 and 685 ± 207 for 2018. These are both above the recent observed average of 592 icebergs for that date, and substantially so for 2017. Given the bimodal nature of the annual iceberg number, this means that our predictions for both 2017 and 2018 are for a high iceberg season, with a 71% level of confidence. However, it is most likely that the 2018 iceberg numbers will be somewhat less than 1000, while our higher hindcast for 2017 is consistent with the observed level of 1008. Our verification analysis, covering the 20-year period up to 2016, shows our model's correspondence to the high or low nature of the 48°N iceberg numbers is statistically robust to the 0.05% level, with a skill level of 80%.



中文翻译:

使用控制系统模型预测纽芬兰冰山季节的严重程度

摘要

纽芬兰东部大银行区的冰山危害一年到一年变化很大。在某些年份中,没有冰山穿透48°N以南,而在另一些国家中,则有超过1000座冰山进入欧洲和北美东北部之间的主要运输通道。对这种季节性危害的预先了解将对船舶航行以及维持有效的冰害服务所需的资源产生重大影响。在这里,采用开窗误差减少率控制系统识别方法来预测纽芬兰省2018年冰山季节的严重性,以预测穿越48°N的冰山数量以及2017年的后冰山数量为依据。估计为2017年9月底之前穿越48°N的766±297个冰山,2018年为685±207。这些均高于该日期最近观测到的592个冰山的平均值,2017年也是如此。鉴于年度冰山数量的双峰性质,这意味着我们对2017年和2018年的预测都是针对冰山旺季, 71%的置信度。但是,2018年的冰山极有可能会略低于1000,而我们2017年的较高后兆与观测到的1008水平保持一致。我们的验证分析涵盖了截至2016年的20年,显示了模型的统计上,与48°N冰山数字的高低性质相对应,在0.05%的水平上具有统计稳健性,技能水平为80%。这意味着我们对2017年和2018年的预测都是高冰山季节,信心水平为71%。但是,2018年的冰山极有可能会略低于1000,而我们2017年的较高后兆与观测到的1008水平保持一致。我们的验证分析涵盖了截至2016年的20年,显示了模型的统计上,与48°N冰山数字的高低性质相对应,在0.05%的水平上具有统计稳健性,技能水平为80%。这意味着我们对2017年和2018年的预测都是高冰山季节,信心水平为71%。但是,2018年的冰山极有可能会略低于1000,而我们2017年的较高后兆与观测到的1008水平保持一致。我们的验证分析涵盖了截至2016年的20年,显示了模型的在统计上,与48°N冰山数字的高低性质相对应,在0.05%的水平上具有鲁棒性,技能水平为80%。

更新日期:2019-06-20
down
wechat
bug