What is it about?

Winds blowing over the ocean generate waves that grow with time. Until they reach a significant height, where they a part of the highest 1/3 of the waves, there is a noticeable time delay between waves and winds that generated them. In this work, we take advantage of this time delay to model a neural network that can predict, by running a time series backwards, the wind speed that generated significant wave height. We compare the method to other wind speed estimation techniques and try filling the gap in wind speed generated by a storm in Placenta Bay, Newfoundland and Labrador, Canada.

Featured Image

Why is it important?

Wind speed estimation from significant wave height has been studied for decades, and have been used in hindcasting studies and remote sensing applications. The proposed method achieved a high level of correlation between true and estimated wind speeds if compared to other estimation techniques, as well as a good correlation with other local wind speed measurements when trying to estimate wind speeds within a measurement gap during a storm. This shows that the proposed method is robust enough to estimate wind speeds in both high and low sea states.


Writing this article helped me understand the challenges we face when dealing with local measurements, and how the model of estimation techniques impact prediction results. I hope it can be seen as an alternative approach to the estimation of wind speeds in remote sensing applications.

Mr. Murilo Teixeira Silva
Memorial University of Newfoundland

Read the Original

This page is a summary of: An Improved Estimation and Gap-Filling Technique for Sea Surface Wind Speeds Using NARX Neural Networks, Journal of Atmospheric and Oceanic Technology, July 2018, American Meteorological Society,
DOI: 10.1175/jtech-d-18-0001.1.
You can read the full text:




The following have contributed to this page