Evaluation of Advanced Microwave Sounding Unit Tropical-Cyclone Intensity and Size Estimation Algorithms
Evaluation of Advanced Microwave Sounding Unit Tropical-Cyclone Intensity and Size Estimation Algorithms
复制标题
先进微波探测装置热带气旋强度和大小估计算法的评估
DOI：
--
复制
发表时间：
2004
期刊：
影响因子：
--
通讯作者：
T. Haar
中科院分区：
文献类型：
--
作者：
J. Demuth;M. DeMaria;J. Knaff;T. Haar
先进微波测深装置（AMSU）数据用于客观估计热带气旋在东北、东南、西南和西北象限的1分钟最大持续地面风、最小海平面压力和34、50和64-kt （1 kt [0.5144 m s 21]）风半径。这些算法来源于AMSU 1999 - 2001年期间大西洋和东太平洋盆地所有热带气旋的温度、压力和风的反演。在多元回归方法中，国家飓风中心的最佳路径强度和操作半径估计值被用作因变量。利用折刀法和2002年飓风季节的独立案例对发展样本的强度算法进行了评估。最大风和最小海平面气压估算的Jackknife结果的平均绝对误差（MAE）分别为11.0 kt和6.7 hPa， rmse分别为14.1 kt和9.3 hPa。在有相应观测资料的情况下，MAE分别为10.7 kt和6.1 hPa， rmse分别为14.9 kt和9.2 hPa。2002年的独立案例的误差仅略大于发展样本的误差。对34、50和64 kt半径的折刀评估结果显示，平均误差分别为30、24和14 n mi。2002年独立样本的结果与发展样本大体相当，但64-kt风半径误差较大。利用飞机侦察数据得到的2002年样本的半径误差都与小刀样本的误差相当，包括64-kt半径。
Advanced Microwave Sounding Unit (AMSU) data are used to provide objective estimates of 1-min maximum sustained surface winds, minimum sea level pressure, and the radii of 34-, 50-, and 64-kt (1 kt [ 0.5144 m s 21) winds in the northeast, southeast, southwest, and northwest quadrants of tropical cyclones. The algorithms are derived from AMSU temperature, pressure, and wind retrievals from all tropical cyclones in the Atlantic and east Pacific basins during 1999‐2001. National Hurricane Center best-track intensity and operational radii estimates are used as dependent variables in a multiple-regression approach. The intensity algorithms are evaluated for the developmental sample using a jackknife procedure and independent cases from the 2002 hurricane season. Jackknife results for the maximum winds and minimum sea level pressure estimates are mean absolute errors (MAE) of 11.0 kt and 6.7 hPa, respectively, and rmse of 14.1 kt and 9.3 hPa, respectively. For cases with corresponding reconnaissance data, the MAE are 10.7 kt and 6.1 hPa, and the rmse are 14.9 kt and 9.2 hPa. The independent cases for 2002 have errors that are only slightly larger than those from the developmental sample. Results from the jackknife evaluation of the 34-, 50-, and 64-kt radii show mean errors of 30, 24, and 14 n mi, respectively. The results for the independent sample from 2002 are generally comparable to the developmental sample, except for the 64-kt wind radii, which have larger errors. The radii errors for the 2002 sample with aircraft reconnaissance data available are all comparable to the errors from the jackknife sample, including the 64-kt radii.