Will it snow? That is the question of the day in Alabama. Everyone wants to know whether we will see snow on Wednesday night or Thursday. Weather forecasts depend heavily on numerical computer models. These models do an excellent job with general trends but they do a poor job pinpointing exact locations and times.
Sometimes models have tendencies for errors. For example, models seem to do a poor job of predicting shallow layers of cold air. Another problem they have is the “wedge”, or cold air damming caused by the Appalachians. In addition, when there is a lot of snow on the ground to our northwest, the air does not heat as rapidly as it would without snow cover. As a result, with winds out of the northwest, our temperatures are sometimes colder than the models project. These local features sometimes make the model predictions a complete bust.
Also models do a poor job sometimes by underestimating the amount of moisture that can feed into the Southeast from the Pacific Ocean as it rides on a strong southern branch of the jet stream. This is where meteorology and weather forecasting become an art almost as much as a science. Forecasters have to account for variables and biases that come with the models.
As I write this on Monday evening we have two models with completely different solutions. The NAM is predicting snow will fall Wednesday night roughly along and north of a line from Hamilton in Marion County to Cullman, to Gadsden in Etowah County. It also shows snow continuing and intensifying during the day Thursday north and west of a line from Hamilton to Scottsboro in Jackson County. The GFS on the latest run is predicting rain across all of North Alabama throughout the period.
Why the radical difference between the model output? I am not certain but I suspect that the NAM is factoring in things that tend to keep north Alabama colder than the GFS. One thing to look for in models is consistency. Models that tend to predict a storm in the same location from runt to run, day to day tend to produce more accurate results. I suggest looking very closely at the Nam and GFS for all four runs during the next 24 hours. They are updated at 0Z, 6Z, 12Z, and 18Z very day. 0Z refers to Greenwich Mean Time which means that it is Midnight in London, or 6 p.m. in Alabama. After the model runs are examined during the next 24 hours we can get an idea as to which is more consistent.
As of now I will say this: Areas north of Hamilton, Cullman, and Gadsden may see snow. I would say that it is a 50/50 proposition at this time for those areas. There is a chance of some significant accumulation but there is an equal chance of nothing but rain. The best chance of significant accumulations would be in Northwest Alabama and the Tennessee Valley. Snowfall in areas to the south of that line is very unlikely. Any snow that falls in those areas will probably be limited to a few flurries as the storm exits the area and cold dry air enters from the north. Perhaps by Tuesday evening, which will be 24 hours from the event, I will have some better idea of what to predict.
Why no tornadoes? April 26, 2016 outflow boundary in northeast Kansas - I haven't done a blog post in a long long time. But several people asked me why last Tuesday (April 26, 2016) wasn't a big tornado day in north & north...
2 hours ago