I guess I have an ingredients-based approach to forecasting. Not because I have conducted any sort of comparison of various methodologies, but because the way Charles Doswell and Tim Vasqeuz write and talk about operational meteorology makes sense to me as a hard-working but nevertheless amateur forecaster. There's nothing either of them have written about forecasting I don't agree with, so none of my ideas are new or original.
Up until 36 hours out, there's little else to do but watch models. You can check the GOM buoys if you're truly bored, but until recovery begins (given applicable situations), there isn't much to watch outside numerical progs. Because we love it and even though we know better, most of my chaser pals and I spend hours scrutinizing output, trying to apply our mesoscale analysis tools to low resolution model data. Of course it's not productive, but it's fun. It may not be good for mental health however.
What I love about the heart of chase season, on the other hand, is that your time to peruse even short-range model data is limited. If things are active, you have more than enough to worry about on Day 1 without getting too deeply into anything beyond. For a long period last season, I never looked more than 48 hours out. Many days last year, people would ask my thoughts about the next day and I had none. I had not even looked at the Day 2 in the morning because I wanted to devote every second to current, real data. When that happens, you know it's an active year.
On Day 1, it's a whole new approach. Over the years I have shaped my process to rely less and less on model output. I don't mean to criticize modelers whose work I respect, but because there is SO much current data to examine and parse, from surface observations, upper air charts, channel imagery, Doppler (even in the morning you're looking for those boundaries in clear air mode), profilers both permanent and VAD, and so on and so forth. At some point you realize that if you can ordinate what data are most applicable in a given scenario, you can run a model in your own head, incorporating what you know about current conditions as well as any pattern recognition skill you bring, climatology, geographic influences--the whole nine yards. If you only have to move the atmosphere forward six to ten hours, then it is vastly more efficient to ingest the data for your own diagnosis and prognosis than to spend critical hours watching the RUC without understanding the actual state of things. I know chasers who pretty much only use model data. They are not altogether unsuccessful. But I have witnessed what Vasquez writes about: that when the setup goes awry or things change unexpectedly, those folks are less prepared to adjust..
With our ever growing access and bandwidth, we can step ever more deeply into the raging data stream, for better or worse. It's hard to know when to stop. It's hard to tell yourself you've examined all the pertinent data, since there is always something you have neglected. If you wake up and discover the target is three hundred miles away, your forecast is so quick and dirty that blundering into a supercell is almost your only hope.
Lastly I examine the RUC and maybe a little ETA (NAM) on Day 1. And of course I read all SPC products, local office discussions, and thunderstorm outlook products. If there's time, I glance at emails from friends or Stormtrack's Target Area.
I try to imagine storms moving through space and time. Where will they initiate? Where will they mature? Where will these boundaries (if any) be at that time? There are many more questions than these. Dozens of questions to ask yourself. When I settle on an area (I try to draw a vaguely oval shape a la Tim Marshall), then I position myself more or less in the middle of that space. Here, I will prefer favorable road networks even if I think some intersection is not the absolute ideal spot. I like intersections a great deal. My final targets are almost always the intersection of major highways or interstates.
On drylines, I like to stay back as far as my curiosity will allow. There is some evidence that the maximum density of pre-init cumulus east of drylines occurs somewhere between 15 and 20 miles from the boundary (Ziegler and Rasmussen WAF, 1998). So I remain east of where I think point convergence might develop or where I can observe vigorous cu forming, perhaps due east or slightly northeast of a dryline bulge. I like room to maneuver. There is almost never a need to close on a storm immediately after initiation. In fact this is often a critical error as a young storm can leave you in the dust if you're parked directly underneath.. Predicting initiation along warm fronts isn't easy either. I suppose I like to stay south and east of where I imagine storms will fire, though in 2004 I repeatedly erred to the north. What we say on the internet in March is often not what we do in May.
The first rule is there are no rules, right?
