The perils of forecasting for family
Published 11:09 am Saturday, April 9, 2022
By Skip Rigney
My wife is the gardener in the family. In March she looks at our dead flowerbeds and envisions transforming them into colorful havens for the butterflies, bees, and birds, in a way that eludes whatever small sector of floricultural creativity resides, mostly dormant, in my brain. As the buds of spring burst forth, so does her energy. She weeds, digs, plants, fertilizes, and mulches.
How do I contribute to this extravagant enterprise of creativity and energy? I make occasional trips to buy mundane commodities such as dirt and mulch. I listen as she patiently explains the pros and cons for various options for positioning of the cone flowers and the zinnias.
Whereas I really don’t have any idea where she should put the cone flowers or the zinnias, when she asks me for a weather forecast, I proudly reply that I’ll check the National Weather Service (NWS) website and look at the various computer weather models.
So, on Monday morning when she inquired about the arrival and intensity of rain, after a brief consultation with my computer, I confidently said, “Tuesday morning, dear. We will get between 0.75 and 1.50 inches of rain.”
She responded by shifting into an even higher gear of botanical intensity. By getting the last plants in the ground, fertilizing them, and completing the deployment of bags and bags, and a few more bags, of mulch on Monday, our (even though she does 98 percent of the work, she graciously calls them “our”) flowerbeds would be in an optimum position to take advantage of Tuesday’s rain.
Tuesday morning dawned breezy with gray, foreboding skies. My wife innocently asked, “When will the rain be here?” I checked the radar. Vague feelings of disquiet and unease began to percolate within me as I stared at the blotches of green, yellow, and red on the screen. Why were they so far north? Why were they more fragmented, less consolidated, than what Monday’s global and medium-scale models had predicted for Tuesday morning?
I quickly switched from the reality of the actual radar to the latest set of predictions from the NWS’s premier short-range computer model, the High Resolution Rapid Refresh (HRRR) model. The NWS re-runs the HRRR every hour for the continental United States, which is much more frequent than the six-hourly runs of the NWS’s global models. The HRRR is also set apart by its sophisticated, highly detailed assimilation of the latest radar data. The result is that the HRRR is often the best tool available for forecasting the weather for the next one to twelve hours.
By Tuesday morning the HRRR predicted that the scattered thunderstorms in south-central Mississippi would indeed consolidate into a heavy line that would sag southward, eventually dumping one to three inches of rain on the northern half of Pearl River County.
But my wife and I, and our flowerbeds, live in southern Pearl River County east of Picayune. The HRRR was predicting that the line would weaken by the time it got to us, and that we would be lucky to get even one-tenth of an inch.
Thankfully the HRRR was too pessimistic. We ended up with four-tenths of an inch. Drier than my original forecast, but wet enough to keep my job as family flowerbed forecaster.