Forecast reviews are written mostly for myself to help improve future forecasts. I care deeply about the success and accuracy of my forecasts. I share my observations with you so we can learn together but also to demonstrate this process is more than just regurgitating a model forecast or taking somebody else's idea.
From the beginning I felt this storm was pretty stable in terms of location. I always felt the heaviest amounts would be across northern Illinois, including places like Chicago. This ideology remained strong through the entire forecasting process. I was also more willing to push totals in northern Illinois because it's not Wisconsin. The idea of pinch and swell; I believe that if heavy snow banding occurs (swell) it's likely that the snow trends immediately north or south will be suppressed (pinch). In otherwise, the result of moisture pooling/daming.
Forecast models were showing a pretty clear picture of the main snow band across northern Illinois. For mesoscale and features like snow-banding I prefer to use high resolution short range models. I used GFS/CAN QPF. In northern Illinois I blended NAM QPF only for northern Illinois. I viewed the EC as being an outlier with very conservative QPF amounts. The EC didn't handle the Feb 3 event as well as the other models did so I was sticking with the hot hand.
Nothing went super wrong, nothing went super right. For the most part the storm followed forecast and the wiggles in the snow band still fell within ranges with exception of the winter storm warning in southwest WI where totals came in under. I didn't try anything exotic or bold on this one.
- The entire forecasting process went very smoothly.
- The big one is northeast Iowa, southwest Wisconsin. This was handled better by short range high resolution models. This did not fit the GFS/CAN big picture. It became very clear as the event folded, certainly not before 8:00PM. The 8:00PM update last evening showed everything going exactly as forecast. The northern squiggle was killer. It shoved snow totals to the top of the range south of La Crosse. Then it gave out around midnight and everything relocated south. I would have blended this into the forecast if timing was different; for example if I created the snow map later in the day.
- WI/IL border heavy snow and Winter Storm Warning. The fact there was a warning tells you that I wasn't alone in my predictions for the border area and south. I was close but not close enough, and perhaps had this morning's snow not completely ended like it did, this spot would look better.
- I mentioned snow ratios of 20-1 in the snow bands across northern IL. 20:1 did in fact occur within the main band while it was in southern Wisconsin. The QPF max was not aligned with the best SLR's which I think relates to squiggly forcing.
- Timing of heaviest snows was late PM into Fri AM. Early forecasts focused them into strictly Fri AM. This was my mistake since I did observe the intial snow influx to be realtively heavy. In conclusion it didn't need to be emphasized.
- The NAM was right about the southern trend on Feb 3. The NAM was right again about the northern bend in snow band for this event. My false assumption based on Feb 3 that the NAM would trend south. I'm starting to believe the NAM trend is related to the upper level flow/momentum. I always try to respect it's northern tendencies in west or southwest flow systems. Somewhere between the GFS and NAM would have sufficed for the general location of things. QPF on the other hand more closely resembled the EC, the outlier.
- Storm ending times could have been handled better.
I'm giving this one an B+.
I really liked the forecast flow. Communication was smooth and forecast stable in the last 24 hours. Everything generally went as planned. We were able to generate a bunch of traffic for our facebook page and website. Traffic is not important but it is a positive feedback of the entire process, nice reward at the end of the day.
I'm not sure how I could have predicted most of these failures. I wouldn't break the forecast to fit the NAM location. I wasn't going to break the forecast to fit the EC QPF, especially with all the dynamics I've mentioned and recent performance. I'm worried making major changes based on the results of this forecast would do more harm than good. I still feel communication can be improved. The miss in southwest Wisconsin can be fixed by respecting short range models more next time. Perhaps improving the 24-72HR forecast will help.
Become a member or sign in to unlock this content!