The Artificial Neural Network (ANN) developed by Dr. Andretta is based on supervised learning algorithms in Python.  The input perceptrons for the network include climatology, observational, and forecast databases over decades.  A simulation of the network uses training and target data sets for the output.  In the slideshow figures (below), the precipitation forecast at KPIH (Pocatello Regional Airport) during the Christmas Blizzard of December 26, 2003 is presented from several sources.  The MOS MET (cyan) and MAV (yellow) guidance underpredicted the heavy precipitation (0.83 inches) at KPIH while the neural network (red) verified closer to observations (green).

  • Figure 1
  • Figure 2
  • Figure 3
  • Figure 4