Data archive for: Exploring the use of machine learning to improve vertical profiles of temperature and moisture

Vertical profiles of temperature and dewpoint are useful in predicting deep convection that leads to severe weather that threatens property and lives. Currently, forecasters rely on observations from radiosonde launches and numerical weather prediction (NWP) models. Radiosonde observations are, however, temporally and spatially sparse, and NWP models contain inherent errors that influence short-term predictions of high-impact events. This work explores using machine learning (ML) to postprocess NWP model forecasts, combining them with satellite data to improve vertical profiles of temperature and dewpoint. We focus on different ML architectures, loss functions, and input features to optimize predictions. Because we are predicting vertical profiles at 256 levels in the atmosphere, this work provides a unique perspective at using ML for 1-D tasks. Compared to baseline profiles from the Rapid Refresh (RAP), ML predictions offer the largest improvement for dewpoint, particularly in the mid-..., This dataset was collected for the corresponding publication in Artificial Intelligence for the Earth Systems, and the processing methodology is outlined in that publication., , # Title of Dataset: Data Archive for Exploring the Use of Machine Learning to Improve Vertical Profiles of Temperature and Moisture

This dataset contains a combination of four data sources that allow for the training and testing of using machine learning to predict temperature and dewpoint vertical profiles. The data consists of matched Radiosonde Observations (RAOB), Rapid Refresh (RAP) output, Real-Time Mesoscale Analysis (RTMA), and Geostationary Operational Environmental Satellite (GOES)-16 data at sites over the central U.S. Tornado Alley from January 2017 through May 2020. All data sources have been collocated to corresponding site, height (256 vertical levels), and time.

## Description of the data

The data are in a single netCDF file, mlsoundings\_dataset.nc. The file consists of five different groups. The groups correspond to the different data sources, with an additional group that has the indices that were used for training, validation, and testing.

## Sharing/Access infor...