东北地区夏季旬土壤水分推算模型的初步探讨

利用东北地区1993—2009年6—8月旬土壤湿度资料,首先采用CAST聚类法对该区域土壤湿度进行气候分区,再以土壤水量平衡原理为依据,在考虑有物理意义的影响因子基础上,分别对每个分区建立统计回归模型,并验证模型的适用性。结果表明:模拟精度在78.8%～88.2%之间,该模型对于东北地区夏季的旬土壤湿度具有一定的推算能力,可以作为一种可应用的模拟方法。     

基于聚类分析的中国东北地区气温和降水时空变化特征

利用1961-2010 年中国东北地区81 个代表站逐日气温和降水资料,采用旋转主分量分析（REOF）和聚类分析（CAST）相结合的方法,对东北地区年平均气温和年降水的时空变化特征进行了分析。结果表明：东北地区年平均气温呈显著的增加趋势,但区域性较为明显,北部和东部地区的升温幅度较小,西部和南部升温幅度较大;年降水量呈现减少的趋势,西部地区降水减少的趋势较为显著。     

祁连山及周边降水分布聚类检验和典型流域增雨效果评价

祁连山及周边区域地形复杂,降水分布极不均匀,如何科学地分区分型,对把握研究区域降水分布特征具有极为重要的意义。利用该区域31个国家基本站1961—2020年5—9月降水量资料,采用主成分分析方法（PCA）对该区域降水量进行分析,再通过聚类分析（CAST）对该结果进行显著性检验,最后将分区分型结果应用于人工增雨作业效果评估。结果表明：（1） 祁连山及周边可分成7个区域（Z1~Z7）,累积方差贡献率超过78%。（2） 这7个区域分别以乐都、海晏、野牛沟、武威、高台、临泽、刚察为中心点,划分结果与降水量、地理地形、海拔高度显著关联。（3） 依托区域历史回归统计方法对人工增雨作业效果进行评价,发现1992—2020年石羊河流域5—9月绝对增雨量、相对增雨率分别为8.91 mm、6.51%,其中7月最高（6.30 mm、21.86%）,8月次之（5.44 mm、16.11%）。基于地面降水量的作业效果评价往往受对比区选择的影响,聚类检验方法不仅有助于复杂地形下降水量的分区研究,还为科学选择对比区提供了客观依据。     

Towards an explanation of why onboard fires happen: The case of an engine room fire on the cruise ship “Le Boreal”

As the cruise ship industry enjoys continuous growth and penetration into new markets, good safety records must be maintained to achieve business objectives. Unfortunately, serious incidents and accidents reoccur on modern cruise ships and hence a better understanding of why this continues to happen is needed. This paper contributes towards a better understanding of underlying causal factors behind onboard fires, focusing on consequential engine room fires in particular. By analysing a recent fire incident on an almost new cruise ship built to the latest standards, we reveal potential deficiencies in maritime safety control and suggest how they can be rectified. The incident analysis was guided by systemic method CAST, which enables to explore a complex socio-technical system responsible for safety control. The analysis provides a different, more comprehensive explanation of, and response to, the incident from that in the official accident investigation report. Given the systemic nature of suggested causal scenarios, they are capable to explain and preclude other incidents. From a practical point of view, the findings would allow developing and maintaining robust safety management systems, which are currently required onboard.