Session 4 AI Applications for the Detection of Earth Science Phenomena

Tuesday, 14 January 2020: 10:30 AM-12:00 PM
156A (Boston Convention and Exhibition Center)
Hosts: (Joint between the 19th Conference on Artificial Intelligence for Environmental Science; and the Events )
Cochairs:
Christina Kumler, University of Colorado, CIRES, Boulder, CO; Aaron Kaulfus, Univ. of Alabama in Huntsville, Earth System Science Center, Huntsville, AL and Vladimir Krasnopolsky, NOAA, EMC, College Park, MD

With the increase in data volume of Earth Science data over the last decade and the projected increase over the next several years, efficient, automated algorithms are required for the identification of phenomena.  The use of artificial intelligence to address this issue is currently in early development and requires further collaboration between physical, data, and computer scientists. Using AI, specifically deep learning methods, Earth Science phenomena can be identified in data and imagery to develop dynamic event databases for detailed scientific process studies, and analyze climatological trends in climate model output.  Such databases would serve to provide pathways to improve our current understanding of the physical atmosphere and how it may change in the future. The purpose of this session is to provide an interdisciplinary forum to showcase existing methods and recent advancements for automated detection of Earth Science phenomena using AI techniques. This session seeks submissions from projects that leverage AI techniques for phenomena detection across Earth Science Big Data repositories, promote the development of phenomenon-based climatologies or databases for further scientific investigation, extract/detect rare phenomena in large datasets or imagery, and overcome challenges related to computational limitation or the data representation of  various phenomena for specific datasets. 

Papers:
10:30 AM
4.1
Detecting Cloud Cover in Webcam Images Using Neural Networks: A Nowcasting Application
Thomas Nipen, Norwegian Meteorological Institute, Oslo, Norway; and E. Myrland, M. Pejcoch, C. Lussana, and I. A. Seierstad
10:45 AM
4.2
Rapid Hailstone Characterization: A 3D Computer Vision Shape Analysis Model
Stan Biryukov, Understory Weather, Madison, WI; and K. Jero, A. Kubicek, E. Hewitt, and J. Leonard
11:00 AM
4.3
Topological Data Analysis and Machine Learning Methods for Pattern Detection in Spatiotemporal Climate Data
Karthik Kashinath, LBNL, Berkeley, CA; and G. Muszynski, M. F. Wehner, V. Kurlin, M. Prabhat, and J. Balewski

11:15 AM
4.4
Analysis and Application of Mesoscale Radar Scenes during Severe Weather Events
Alex M. Haberlie, Louisiana State Univ., Baton Rouge, LA; and W. S. Ashley, V. A. Gensini, and M. Karpinski
11:30 AM
4.5
Deep Learning Approach for the Detection of Areas Likely for Convection Initiation
Jebb Q. Stewart, NOAA, Boulder, CO; and C. Kumler, D. Hall, and M. W. Govett
11:45 AM
4.6
Using Deep Learning to Create a Long-Term Climatology of Warm and Cold Fronts
Ryan A. Lagerquist, CIMMS, Norman, OK; and J. T. Allen and A. McGovern
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner