An Artificial Neural Network Model Function (AMF) for SARAL-Altika Winds

Abstract

High-quality winds over the ocean surface, at an enhanced spatio-temporal resolution are required for a better understanding of the dynamics of the ocean and atmosphere. Altimetry helps in increasing the frequency of satellite observations. Traditional algorithms for wind speed retrievals from altimeter consider only the backscatter (sigma-0) and possibly the significant wave height (SWH). In this study, we propose an artificial neural network (ANN) model function for AltiKa on board Satellite for ARgos and ALtiKa (SARAL) to relate wind speed to sigma-0, SWH, the width of the waveform leading edge, the two brightness temperatures (TBK and TBKa), and the amplitude of the 1-Hz echo. These parameters influence either the backscatter from the ocean or the propagation of the altimeter radar signal. The wind estimates have significantly improved by incorporating these parameters.Index Terms-AltiKa, artificial neural network (ANN), geophysical data records, wind speed.

Tropical Cyclone Wind Speed Estimation From Satellite Altimeter‐Derived Ocean Parameters

“…The training epoch stops when the error converged to the minimum. Many studies have applied ANN for normal conditions U 10 estimation (Ali et al, 2015;Gourrion et al, 2002;Jiang et al, 2020), but challenging data filtering and extensive training samples requirement limit the ANN for U 10 tropical cyclone application.…”

“…The recent altimeter missions with radiometers onboard provide simultaneously brightness temperatures (T b ) that can be used to estimate this response to the signal. Integrating the T b in U 10 model have improved the accuracy of altimeter U 10 in low-to-moderate wind (Ali et al, 2015;Bushair & Gairola, 2019;Jiang et al, 2020), and scatterometer U 10 inside the tropical cyclone (Stiles et al, 2014;Xu et al, 2018). Lillibridge et al (2014) reported an improved accuracy of SARAL Ka-band U 10 when integrating atmospheric parameters such as pressure, temperature, water vapor, and liquid water content.…”

Predicting Indian Ocean Cyclone Parameters Using an Artificial Intelligence Technique

“…This non-dynamic numerical model has been used in many oceanographic [10][11][12][13][14][15] (and meteorological studies [16][17][18]. The ANN technique is also useful for satellite-parameter retrievals [19][20][21][22]. Multiple linear regression (MLR) is a method dealing with linear dependencies, whereas neural networks deal with nonlinearities.…”

Predicting Indian Ocean Cyclone Parameters Using an Artificial Intelligence Technique

“…This non-dynamic numerical model has been used in many oceanographic ( (Liu et al, 1997, Sharma and Ali., 2012. This technique is also found useful for satellite parameter retrievals (Krasnopolsky and Schiller., 2003, Ali et al, 2015, Sharma et al, 2013, Ali et al, 2016. ANN requires three sets of data, one for training, one for veri cation and the other for validation.…”