Machine learning-based detection of weather fronts with DL-FRONT in CESM1.3

These data are the results of detecting weather fronts with the machine learning algorithm DL-FRONT (see Biard and Kunkel 2019) in simulations with the Community Earth System Model, version 1.3 (CESM1.3, see Meehl et al. 2019). The specific CESM1.3 simulations used here include: a historical climate simulation from 2000 to 2005, a simulation with Representative Concentration Pathway 2.6 (RCP2.6) forcing from 2006 to 2015, and a simulation with Representative Concentration Pathway 8.5 (RCP8.5) forcing from 2086–2100. See Dagon et al. 2022 for a publication analyzing these machine learning detected weather fronts and associated extreme precipitation in historical and future climates.

At each 3-hourly time step of simulation output over a North American spatial domain (10-77ºN, 171-31°W), DL-FRONT produced a set of spatial grids at 1° spatial resolution, for each of the five categories: cold front, warm front, stationary front, occluded front, and no front. 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 the "no front" category, the probability that the cell is not in any weather front boundary region).

The dataset contains three sets of files. The first set contains the original front probability maps. The second set contains "one hot" 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. The third set contains front crossing rates (monthly, seasonal, annual) and climatologies/anomalies/standard deviations (monthly, seasonal) for each front type.

The front probability files have names that follow the form cesm_fronts__.nc for each simulation period. The one hot files have names that follow the form cesm_fronts____MaskedNetCDF_customgrid.nc.  The rates files have names that follow the form cesm_fronts___frontRates_viaPolylines_customgrid.nc. For the one-hot and rates files, the historical and RCP2.6 simulation output have been combined into a single file.