Climatology and Variability of Warm and Cold Fronts over North America

An objectively derived climatology of warm and cold fronts and their climatological variability for the period 1979-2018 is described. Prior analyses of the climatology of these features have been limited by inconsistencies in detection methods using manual approaches or numerical frontal analysis of thermal or wind gradients, and the available horizontal resolution of reanalyses. In a new approach, frontal regions are identified using a convolution neural network (CNN) trained on Weather Prediction Center (WPC) warm and cold frontal labels to identify patterns with high likelihoods of being fronts in ERA5 reanalysis (0.25 degree) near-surface and lower-atmospheric pressure level fields at 3-hourly intervals. Unlike earlier approaches, developing this objective approach allows an efficient analysis of an enormous volume of data, allowing for assessment of both trends in frequency and analysis of inter annual variability. Frontal labels are then used to define the climatological frequency of fronts, and make comparison to other climatologies.

The question of variability has rarely been addressed in relation to the frequency of fronts over North America. Previously, the authors hypothesized a shift in warm and cold frontal systems in response to the El Niño-Southern Oscillation (ENSO) state over North America, which would be consistent with the notable displacement of ENSO modulated cyclone track. While other authors have presented evidence for frontal variability as a function of climate states globally, these analyses were conducted using a traditional frontal identification and tracking method that did not have the capability to distinguish between warm and cold fronts, nor the resolution sufficient to clearly characterize these features. In this presentation, we will describe the climatological frequency of warm and cold fronts over North America, the long-term trends in their occurrence, and interannual variability of their positioning as a function of ENSO state. A complimentary talk by Ryan Lagerquist will discuss the tuning and application of the CNN and discuss basic findings from the climatology.