Machine Learning for Soundings (Vertical Profiles of Temperature and Dewpoint)
To accurately forecast thunderstorms, it is crucial to have frequent and accurate vertical profiles of the temperature and dewpoint, both of which affect the instability of the atmosphere. This instability is typically measured by the quantity convective available potential energy (CAPE), and the likelihood that the instability will be realized is given by the quantity convective inhibition (CIN). CAPE and CIN are highly sensitive to the temperature and dewpoint in the lowest layer of the atmosphere, usually from the surface to 1-2 km in altitude.
Radiosondes (weather balloons) provide the best thermodynamical profiles, as they measure atmospheric variables with high resolution while ascending in the atmosphere. However, unfortunately balloon releases are conducted routinely only twice a day and are separated by several hundred kilometers. As a result, forecasters leverage vertical profiles generated by numerical weather prediction (NWP) models, which have high spatial and temporal resolution. Unfortunately, these models are not perfect, and this work uses machine learning (ML), with input from satellite data, to improve vertical profiles of temperature and dewpoint produced by an NWP model.
As part of this work, we
- Apply ML to 1-D temperature and dewpoint profiles,
- Incorporate physically-inspired loss functions,
- Utilize a variety of physic-based metrics to evaluate and improve the models, and
- Provide well-calibrated uncertainty estimates at every level in the vertical column.
For the ML model, we explored the use of different ML model architectures, including linear regression, fully-connected neural networks, convolutional neural networks, and deep residual UNets. Using data from 2017-2020, we found that all ML model architectures improve the NWP profiles; however, the UNet models are the best-performing since they can fully-utilize the vertical spatial information. The ML models also substantially improve the CAPE and CIN predictions.
In addition, ML model’s are capable of predicting uncertainty estimates that correctly match the corresponding error for the majority of the predictions without suffering any degradation to the temperature and dewpoint profiles (see figure). Having uncertainty estimates at all vertical levels helps identify where the ML model may have the largest errors, providing forecasters not only with a heads-up for potential errors, but also with information regarding where in the profile these errors are occurring so that they may use their domain expertise to help overcome these errors. The uncertainty estimates are well-calibrated, meaning that removing the most uncertain samples results in improved model performance; thus, if specific applications require more accurate results, separate thresholds on both temperature and dewpoint can be set up to show only the most-certain profiles with reasonable confidence.
NOAA-Funded Future Work
- Update datasets (2021-2022 RAP, RTMA, and GOES data)
- Augment ML algorithms to include epistemic uncertainty estimates of temperature and dewpoint (at every vertical level)
- Explore adding additional features (e.g., wind profiles, surface fluxes)
- Expand prediction area, create hourly gridded predictions, and evaluate temporal and spatial variability
- Develop pipeline for creating real-time predictions
- Haynes, K., J. Stock, J. Dostalek, C. Anderson, and I. Ebert-Uphoff, Exploring the use of machine learning to improve vertical profiles of temperature and moisture. Artif. Intell. Earth Syst., 2023, conditionally accepted.
Katherine Haynes (CIRA)