Skip to content

This repository contains python script for predicting cloud droplet concentration based on aircraft observations. It involves data cleansing, manipulation and visualization to better understand the droplet formation trends.

Notifications You must be signed in to change notification settings

adityabiyani97/MS-Research-cloud-drop-prediction

Repository files navigation

Predicting cloud droplet concentration using machine learning methods.

Clouds are the basic building blocks of all the hydrometeor species in the atmosphere like rain, snow, ice and graupel. Their concentration in the atmosphere determines the concentration of subsequent hydrometeors. Thus, it is important to be able to accurately predict the concentration of cloud droplets in the atmosphere and any time instance and location.

The repository consists of my one year of research on how to use machine learning methods to predict the drop conc. Currently, it has two python files written in Jupyter notebook explaining the data cleaning and analysis done on files retrieved from the ARM wing of U.S Department of Energy. First file contains code on how the data was reterieved, cleaned, engineered (time-series, spatio-temporatal analysis) and analysed using Regression algorithms like Linear Regression, Neural Networks and Random Forests using sklearn packages.

The second file contains python scrip that using cartopy library along with other python visualization packages to understand the airplane trajectory over the Amazon Rainforest (in Manaus, Brazil) using Radar reflectivity data (extracted from netcdf files) and NASA MODIS Satellite images.

Additionally, some plot images are also added to substantiate the research undertaken.

Acknowledgement:

  1. ARM: https://www.arm.gov/working-with-arm/acknowledging-arm
  2. Dr. Hamish Gordon: https://engineering.cmu.edu/accelerator/research/faculty/gordon-group.html

About

This repository contains python script for predicting cloud droplet concentration based on aircraft observations. It involves data cleansing, manipulation and visualization to better understand the droplet formation trends.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published