Detection of Tropical Overshooting Cloud Tops Using Himawari-8 Imagery

Abstract

Overshooting convective cloud Top (OT)-accompanied clouds can cause severe weather conditions, such as lightning, strong winds, and heavy rainfall. The distribution and behavior of OTs can affect regional and global climate systems. In this paper, we propose a new approach for OT detection by using machine learning methods with multiple infrared images and their derived features. Himawari-8 satellite images were used as the main input data, and binary detection (OT or nonOT) with class probability was the output of the machine learning models. Three machine learning techniques-random forest (RF), extremely randomized trees (ERT), and logistic regression (LR)-were used to develop OT classification models to distinguish OT from non-OT. The hindcast validation over the Southeast Asia and West Pacific regions showed that RF performed best, resulting in a mean probabilities of detection (POD) of 77.06% and a mean false alarm ratio (FAR) of 36.13%. Brightness temperature at 11.2 µm (Tb11) and its standard deviation (STD) in a 3 × 3 window size were identified as the most contributing variables for discriminating OT and nonOT classes. The proposed machine learning-based OT detection algorithms produced promising results comparable to or even better than the existing approaches, which are the infrared window (IRW)-texture and water vapor (WV) minus IRW brightness temperature difference (BTD) methods.

Rapidly Evolving Cirrus Clouds Modulated by Convectively Generated Gravity Waves

“…While early modeling approaches were restricted to two dimensions (Alexander et al, 1995;Holton & Alexander, 1999;Lane & Knievel, 2005), later three-dimensional models (Grimsdell et al, 2010;Trier et al, 2010) were used in tandem with observations to study gravity waves excited from individual or clusters of convective systems. More recently, Su and Zhai (2017) found close agreement in the propagation Kim et al (2017) convection.…”

Icing Detection over East Asia from Geostationary Satellite Data Using Machine Learning Approaches

“…The multinomial log-linear model (MLL) is an improved version of logistic regression that incorporates an artificial neural network approach for parameter optimization [42]. Logistic regression has often been used in meteorological satellite applications [30,31] and operational systems, such as to monitor rapid development thunderstorms (RDT) using Spinning Enhanced Visible and Infrared Imager (SEVIRI) satellite data [43]. Thus, in this study, the two machine learning classification approaches-RF and MLL via neural networks-were used to develop icing detection models.…”

“…Random forest (RF) has proven very effective for several meteorological satellite-based applications such as the detection of convective initiation and overshooting tops, and rainfall rate estimation [29][30][31]41]. The multinomial log-linear model (MLL) is an improved version of logistic regression that incorporates an artificial neural network approach for parameter optimization [42].…”

“…The performance of icing detection algorithms was evaluated using five-fold cross-validation based on a 2 × 2 contingency table, helping to quantify the detection probability of models [29][30][31][59][60][61] (Table 5). Based on the components of the contingency table, four performance metrics were calculated: the probability of detection (POD; calculated as H/(H + M) × 100%); the probability of false detection (POFD; calculated as F/(N + F) × 100%); overall accuracy (OA; calculated as (N + H)/(N + F+ M + H) × 100%); and total skill statistics (TSS; calculated as (H × N−F × M)/[(H + M) × (F + N)]).…”

“…Unlike typical statistical approaches, machine learning is generally free from data assumptions and has been proven to be effective in modeling non-linear behaviors. For meteorological applications, machine learning has been applied to detect phenomena that pose a risk to aviation, such as overshooting tops and cloud convective initiation [29][30][31].…”