Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

MachineLearningInClimateAndWeather

N/A
by

Tianyu Jiang

on 10 May 2016

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of MachineLearningInClimateAndWeather

Visit:
http://prezibase.com
Prezi Templates
Custom Design
Prezi To Video
Add your own content
Prezi Tips & News
Create a unique Presentation
http://prezibase.com
prezibase
prezibase.com
www.prezibase.com
prezi templates
prezibase templates
Bridging the Gap between Weather and Climate:
Machine Learning Approaches


By Tianyu Jiang
Computer Science and Mathematics Division
Oak Ridge National Lab

Weather
: (regional impact, short-term)

Short-term (minutes to days) variations in the atmosphere.
Thought of in terms of
temperature
, humidity,
precipitation
, cloudiness, visibility, and wind. ( < 2 weeks).














Weather & Climate
Climate
: (global impact, long-term)
climate is the slowly varying aspects of the earth (atmosphere, hydrosphere, land surface) system.






Motivation
Using climate information to improve weather predictablity is important to general public, especially on energy comsuption.

Undestanding change of weather under global warming and making recommandation is also critical to policymakers.




The Gap:
Long-term v.s. short-term, global v.s. regional
Technical challenges in automatically detecting the weather phenomena from massive collected data.


Data
Various observation
Local report
Climate model generated data
Tools:
Climate (numerical) model
Machine learning model

Weather Phenomena of Interest:
Winter Storm
Atmospheric Blocking
Atmospheric River

I .
Winter Storm tracking
:
Responsible for most winter blizzard, associated with strong wind and low-temperature.

Data: (global gridded)
sea level pressure
2-D wind field
temperature.

Machine learning methods:
anomaly detection
clustering
feature tracking
Winter Storm Tracks
II. Atmospheric Blocking:
A weather phenomenon that blocks the normal west-east flow of airmass and water vapor transport.
Responsible for west coast drought, heat waves in the U.S. and Europe, and cold air outbreak.
Classify blocking phenomenon with SVM:
Data:
Gridded wind and height data, and historical records are used to train a
support vector machine (SVM) model
with large margin.
The classified pattern are further clustered and verified through the cross-validation and connected labelling.
Atmospheric River:
a long and narrow water vapor band on the weather map or satellite image.
Rich in water vapor.
Producing extreme precipitation (rainfall and snowfall) and creating floods.
Data:
2-D wind
total water vapor from satellite
Weather station record
Local report

Machine learning algorithm:
Random Forest
image processing
Identified Atmospheric River
Percentage of Triggered Flooding
Summary:
Weather patterns are successfully detected from large datasets through machine learning methods.

The machine learning models are able to predict impact of future climate change.

The insight is critical to policymakers and relevant industries on decision making.
Winter Storm in satellite image
Full transcript