When I arrived here in 1984, the National Weather Service was just coming off a long-used model called the LFM. For its time, it was a model breakthrough, offering more detail in larger regions such as the northeast United States than had been available 10 to 15 years earlier.
Part of its output produced numerical forecasts that indicated how much liquid precipitation would fall in six-hour increments, how much relative humidity was present at three different altitudes, how stable or unstable the atmosphere would be, what the surface pressure would be, what the wind direction would be in 10-degree increments, how fast in knots the wind would blow, and what the temperature would be at a specific location such as KBUF, the Buffalo-Niagara Airport.
Similar numbers are shown at left, from a more modern model used today. Most young meteorologists I know nearly snicker when I tell them I still take a look at this kind of output. Old guys (and gals) have these old habits.
In fact, under the column “DD” is the wind direction. This primitive product was the first quick sign I saw Monday morning that Wednesday’s lake snow might well head all the way up to and past Tonawanda Creek. It showed “23” in that column for 1 p.m. Wednesday. That told me the output from this NAM model was projecting the wind to back all the way around to coming from 230 degrees on the compass. We know from experience that polar air coming from that direction can bring lake snow as far north and west as Grand Island, the Tonawandas and Niagara Falls.
Not all the many other models agreed with that much backing, but this made me sit up and take notice of the possibility this band would head pretty far north —more than “typical.” One experimental, very high resolution National Weather Service model seemed fairly close to this much backing for several hours, which made me take even more notice.
So, let’s go back to that old LFM. The “L” stands for limited and, for the Great Lakes, it stood for limitations.
The Great Lakes in the LFM weren’t lakes at all. They were represented as flat plains that neither transferred heat nor moisture. Those limitations were not due to the presence of dopes in the modeling community. The problem was mainly the lack of computer crunch power and resolution needed to put sets of equations to good use. We simply did not have the ability to represent relatively smaller areas as bodies of water that could emit or absorb heat, or send water vapor into the atmosphere.
But by the '70s, we did have pretty decent satellite imagery, and we had plenty of surface and atmospheric observations to see what was going on upwind of our location. An experienced forecaster could at least look to Michigan and Ontario and, using pattern recognition, make a fairly good forecast of what would be unfolding for us in a broad perspective. We were able to see Jimmy Griffin’s Six-Pack storm coming days in advance, but we could not narrow down which parts of Western New York would get hit the hardest very far in advance.
Back when I arrived here, even an improved model that was coming on line, called the NGM, could still be wildly off the mark in low-level wind direction. My recollection is that numerical output of that model’s wind direction for KBUF was off by 30 or 40 degrees on the compass on Jan. 20, 1985, preceding the Blizzard of '85 on Jan. 21.
We seldom see errors of that magnitude now that we have such detailed, high-resolution models. However, as I discussed in my primer on lake-effect snow, an error of just 5 degrees on the compass in wind direction can make for a significant forecast error for a very small geographical area with a whole lot of people living in it, especially in central and northern Erie County, our most densely populated region.
After a successful forecast like the one just passed midweek for a major lake-effect event, someone like me might tend to feel a little smug. But nature will usually come along and take someone like me down a peg soon enough. I've been off work since Wednesday evening. Still, I'm fairly confident I would not have forecast Friday's lake effect to produce more than 6 inches accumulation in parts of the metro area if I'd been working. Smugness, be gone!
Story topics: By Don Paul