User Defined Machine Learning Determined Tropical Cyclone Intensity
Input one of the following (if more than one inputted, farthest left will be used):
Date:
--/--/---- --:--:--
Location:
--N --W
Machine Learning Intensity:
- kts
Estimated Radius:
- n mi
Previous Intensity:
- kts
Sensitivity (Gradient)
Red indicates parts of the satellite imagery where warmer brightness temperatures would increase the estimated intensity. Blue indicates parts of the satellite imagery where colder brightness temperatures would increase the estimated intensity.
Attribution (Input*Gradient)
Red indicates parts of the satellite imagery where values contributed positively to the estimated intensity. Blue indicates parts of the satellite imagery where values contributed negatively to the estimated intensity.
CAM 1
This captures the smallest scale features such as cloud edges.
CAM 2
This captures the largest scale features such as overall cyclone structure.
Note: This model largely adjusts the intensity from the previous intensity. Rapid changes in intensity are not captured well.
Download Graphic