A Preliminary Evaluation of Advanced Dvorak Technique-derived Intensity Estimate Errors and Biases During the Extratropical Transition of Tropical Cyclones Using Synthetic Satellite Imagery
Real-time and historical tropical cyclone (TC) intensity estimates during extratropical transition (ET) are derived mainly from satellite-based methods such as the Dvorak Technique (DT) and Advanced Dvorak Technique (ADT). However, the empirical relationships developed between cloud organization pat...
| Main Author: | |
|---|---|
| Format: | Text |
| Language: | unknown |
| Published: |
UWM Digital Commons
2014
|
| Subjects: | |
| Online Access: | https://dc.uwm.edu/etd/533 https://dc.uwm.edu/context/etd/article/1538/viewcontent/Manion_uwm_0263m_10821.pdf |
| Summary: | Real-time and historical tropical cyclone (TC) intensity estimates during extratropical transition (ET) are derived mainly from satellite-based methods such as the Dvorak Technique (DT) and Advanced Dvorak Technique (ADT). However, the empirical relationships developed between cloud organization patterns and cyclone intensity that underlie the DT and ADT are primarily tropical in nature and thus become less reliable during ET. Preliminary analyses suggest that ADT-derived intensity estimates are weak-biased during ET; however, due to the lack of direct observations of cyclone intensity during ET, the extent to which this is true is unknown. Herein, an attempt to quantify errors during this process is evaluated. Synthetic satellite imagery obtained from numerical simulations of five representative North Atlantic ET events between 1990 - 2013, spanning five microphysics schemes, are used to quantify ADT-derived intensity estimate errors during ET. Model-derived hourly time series of minimum sea level pressure and maximum sustained surface wind speed obtained from each simulation serve as a proxy for "observed" TC intensity to which ADT-derived intensity estimates are compared. Results suggest that ADT-derived composite intensity estimates are weak-biased, with mean errors that are largest during the onset of ET and become somewhat smaller as ET ends. An alternative means of obtaining intensity estimates during ET, utilizing an empirical orthogonal function-based linear regression model, is developed and evaluated. The linear regression model shows promise in providing improved intensity estimates during ET; however, the small sample size of cases used in its development precludes any meaningful conclusions from being derived from the analysis. |
|---|