Regional Frequency Analysis of Extreme Winds at Parana State, Brazil

Abstract Gusts from 26 meteorological stations at Paraná (Brazil) were analyzed to estimate their magnitude and frequency. This analysis integrated the homogeneous regions identification, using the principal components analysis, and regional frequency analysis, since the available dataset were too short to apply conventional methods. Similar data patterns were identified in five regions: Coast, Central, Mid-West, West and North. This results were geographical consistent, because the stations are at contiguous sites and located at regions with similar characteristics of relief and coast distance. The Wakeby distribution was selected as the most adequate for the regional frequency analysis. The regional quantiles estimated by the adjusted distribution were coherent with the spatial differences verified in the descriptive statistic. For a given return period of 50 years, the region composed by Cascavel and Toledo stations showed the most severe winds, with 41,9 and 44,8 ms-1, respectively. In the West winds vary from 35,3 to 39,7 ms-1, in the Central from 28,8 to 38,2 ms-1, in the North from 26,4 to 36,6 ms-1 and in the Coast from 21,4 to 30,6 ms-1. The results of this study are innovator since the wind probabilities were estimated using a short dataset, however with good quality and temporal resolution.

摘要：本研究针对巴西巴拉那州（Paraná）26个气象站点的阵风数据展开分析，旨在估算其强度与发生频率。鉴于可用数据集时长过短，无法直接应用常规分析方法，本次分析整合了基于主成分分析（principal components analysis）的同质区域识别流程与区域频率分析手段。研究在五大区域中识别出相似的数据模式，分别为沿海区、中部区、中西部区、西部区与北部区。该结果具备地理一致性，因所有站点均位于相邻区域，且所处区域的地形特征与距海岸的距离均较为相似。最终选定Wakeby分布作为区域频率分析的最优拟合分布。经校正分布估算得到的区域分位数，与描述性统计中验证得到的空间差异结果相一致。在50年重现期下，由卡斯卡韦尔（Cascavel）与托莱多（Toledo）站点组成的区域出现了最强阵风，风速分别达41.9米/秒与44.8米/秒。其中，西部区域风速区间为35.3~39.7米/秒，中部区域为28.8~38.2米/秒，北部区域为26.4~36.6米/秒，沿海区域则为21.4~30.6米/秒。本研究结果具备创新性：其风速概率通过短时数据集估算得到，但该数据集具备良好的质量与时间分辨率。