DL-FRONT MERRA-2 weather front probability maps over North America, 1980-

<p>DL-FRONT is a Deep Learning Neural Network (DLNN) that was trained to detect weather fronts using spatial&nbsp;grids of near-surface atmospheric variables. The dataset is composed of hourly spatial grids containing probability maps for each of five front-type categories&mdash;cold front, warm front, stationary front, occluded front, and no front.</p> <p>This dataset is the product of processing data from the National Aeronautics&nbsp;and Space Administration (NASA)&nbsp;<a href="https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/">Modern-Era Retrospective analysis for Research and Applications, Version 2</a> (MERRA-2). DL-FRONT processed MERRA-2 hourly data grids of instantaneous measures of air pressure reduced to mean sea level, air temperature at 2 meters, specific humidity at 2 meters, and wind velocity at 10 meters over the time span 1980 - 2018&nbsp;to produce this dataset. The original MERRA-2 data were resampled at 1 degree resolution over the spatial range 31W - 171W x 10N - 77N using bicubic interpolation.</p> <p>At each hourly time step&nbsp;the network produced a set of spatial grids with the same resolution and spatial range as the input, one for each of the five categories mentioned above. Each cell in a spatial grid for a given category records the network-assigned probability (from 0.0 to 1.0) that the cell is in a weather front boundary region of that category (or, for&nbsp;the &quot;no front&quot; category, the probability that the cell is not in any weather front boundary region).</p> <p>The DLNN was trained using MERRA-2 data and human-identified fronts from the NOAA National Weather Service (NWS) Weather Prediction Center (WPC) <a href="https://www.wpc.ncep.noaa.gov/html/sfc2.shtml">Coded Surface Bulletin</a> dataset. The training datasets covered the years 2003-2007.</p> <p>The dataset contains two sets of files. The first set contains&nbsp;the original front probability maps. The second set contains &quot;one hot&quot; versions of the front probability maps. In the one hot version the five front-type probabilities for a spatial grid cell for a given time step are replaced by the value 1 for the largest front-type probability, and by 0 for the others.</p> <p>The front probability files have names that follow the form merra2_merra2-1deg_fronts_&lt;year&gt;.nc. The one hot files have names that follow the form merra2_merra2-1deg_onehot_&lt;year&gt;.nc. Each file contains one year&nbsp;of hourly spatial data grids.</p>