Detection Uncertainty Matters for Understanding Atmospheric Rivers
The Berkeley Lab CASCADE SFA hosted over 30 participants from multiple universities and research institutions for the 3rd Atmospheric River Tracking Method Intercomparison Project (ARTMIP). Presentations and discussions by participants helped to summarize the state of current knowledge related to identifying atmospheric rivers (ARs) in atmospheric datasets.
The resulting workshop summary article enumerates emerging directions for AR research. Further, participants contributed to the first-ever dataset of ARs that are manually identified by experts. This will have long-term, positive impacts on the use of machine learning for AR research.
The Berkeley Lab CASCADE SFA hosted over 30 participants from multiple universities and research institutions for the 3rd Atmospheric River Tracking Method Intercomparison Project Workshop: October 16-18, 2019, Lawrence Berkeley Lab. Presentations and discussions by participants helped to summarize the state of current knowledge related to identifying atmospheric rivers (ARs) in atmospheric datasets. The workshop discussions also helped identify emerging research directions for AR research, such as:
- Understanding drivers of the AR lifecycle
- Categorizing different types of AR
- Categorizing classes of AR detection algorithm
- Leverage machine learning techniques for AR research