Authors:

Henry Peng

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Heat waves can have devastating effects on crop growth, infrastructure, and human life. Therefore, understanding and predicting heat waves are integral to a wide variety of industries. Machine learning techniques including convolutional neural networks (CNN) have seen widespread use in many aspects of atmospheric science, and this study aims to evaluate if those techniques can be applied to predict heat waves considering that large-scale atmospheric circulation may be conductive to the occurrence and persistence of heat waves. To do so, a CNN is trained on geopotential height at 500 hPa data obtained from a large scale simulation dataset (LENS) before transfer learning is applied using the same variable from an observed reanalysis dataset (ERA5). This study will look at the results of the CNN and determine the relation of heat waves and atmospheric variables along with application of CNN in heat wave prediction. Using this model, this study will provide conclusions on the viability of CNN in heat wave prediction and identifying large-scale atmospheric circulation patterns preceding heat waves.

Source:

Purdue University / 2023

Topics:

No topics listed

Co-authors:

Henry Peng