CONUS5 Dataset

The CONUS5 dataset provides: Inputs: GOES-16 ABI Band 02 0.64 micron Red Visible Reflectance (256 x 256), GOES-16 ABI BAnd 13 10.3 micron Clean Longwave Infrared Window Brightness Temperature (64 x 64) Output: MRMS Convection Flag (256 x 256) The data consist of 483 cases with 25 images per case with a 15-minute time step. Cases were chosen to maximize the number of SPC storm reports. A reverse parallax adjustment was applied to MRMS assuming a cloud top height of 10 km. The MRMS Convection Flag was computed from the MRMS PrecipFlag using the mapping: 6 (convection), 7 (hail), 96 (tropical convective rain) --> 1 = convection, 0 (no precip), 1 (warm stratiform rain), 3 (snow), 10 (cool stratiform rain), 91 (tropical stratiform rain) --> 0 = not convection, -1 (missing), -3 (no coverage) --> -1.E30 (fill value). Pixels with missing GOES data have no-data fill value of -1.E30. The data preparation methodology is described in: Hilburn, K. A., I. Ebert-Uphoff, and S. D. Miller, 2020: Development and interpretation of a neural-network-based synthetic radar reflectivity estimator using GOES-R satellite observations. J. Appl. Meteor. Climatol., 60, 3-21, https://doi.org/10.1175/JAMC-D-20-0084.1. The dataset is meant to extend the Lee et al. dataset described in: Lee, Y., C. D. Kummerow, and I. Ebert-Uphoff, 2021: Applying machine learning methods to detet convection using Geostationary Operational Environmental Satellite-16 (GOES-16) advanced baseline imager (ABI) data. Atmos. Meas. Tech., 14, 2699-2716, https://doi.org/10.5194/amt-14-2699-2021.