I wanted to take a moment, and go back to Oct 21, when I had noticed on our home Alexa that the forecast MaxT for the following Tuesday (Oct 27) was 39!!! Of course, several laughed at that, as we were having MaxTs in the low 80s recently.
Now, the event has come and gone, and sure enough, the Oct 27th Observed MaxT at KDFW was 40 (daytime), and 41 (9pm that evening).
So, looking back, first I wanted to share a glimpse of the GEFS/GFS output. Using EMC’s GEFS plumes page, and looking at the Oct 20th, 12UTC run, the GEFS/GFS appears to have done QUITE well.
The operational GFS is in BLUE, the GEFS mean in BLACK, and the observed T values in yellow dots.
In short, both the GEFS mean and operational GFS advertised well the drop from MaxTs in the low 80s, to the spell of sub-50 MaxTs that we just experienced for a few days. The GFS actually was forecasting temps in the upper 30s in this run (37, as can be seen in NBM views below).
So, to the NBM, though not shown, I recall seeing text bulletins from the extended range NBM that were showing ranges of potential MaxTs of about 40 degrees!!
So, I went back and took a look at the NBM Std Dev, from the Oct 20th 0700UTC run. Using the NOAA Weather and Climate Toolkit (W&CT), pointed at past NBM runs available through the NOAA Big Data Program on AWS, I was able to create this view of the large area of greater than 5 degK (~9degF) stdev spread within the NBM's contributing model members (~40+ degF max/min range) for the Oct 27th MaxT!
As comparison, one can use the newly-updated WSUP viewer to view the same, but in units of degF. Note the coincidence of the above 5degK+ area, with the below 9degF+ area.
Similarly, one can use the two sources to compare the NBM forecasts, using the 2 viewers.
In this first image, from the WSUP viewer, note the MaxT=58 in the vicinity of KDFW, and the sub-50 MaxTs forecast in central OK.
This next image is the same NBM forecast run, for the same valid time, but has been trimmed in the W&CT to only show MaxTs of 50 degF or warmer. [Filtering data/imagery to subsets of the entire grid is useful for exporting/highlighting impacting values, which can be used in downstream GIS interfaces/dashboards.]
For comparison to the above Oct20, 0700UTC NBM run, I also looked at the MaxT forecast in the NDFD, 15 hours later in the 2135UTC issuance available on AWS. The same-color-range image, also only showing MaxT>50degF, shows how several sites near WFO AMA properly identified a forecast Target of Opportunity to adjust down the forecasts from the NBM guidance. This was the correct adjustment, as I believe MaxTs were in the teens in the panhandle that day! Obviously, a 3rd image of the Observed MaxT similarly colorized would be helpful in this review.
This can also be nicely visualized in a short loop of the two images, which can be viewed here.
Lastly, I wanted to make use of the ptvis display options in the WSUP viewer, to review the same.
So, looking again at the Oct 20th, 0700UTC NBM run, and highlighting the MaxT violin pop-up information valid at 06UTC Oct 28th, one can see the 38-degree F spread in the NBM (80-42.5) in its MaxT forecasts, with a mean of 60.
Next, one can use the new dModel/dt feature in the v2.2 version of the WSUP to explore run to run comparisons of NBM runs leading up to the event valid time---in this case, Oct 28, 06UTC. In the image below, you can see the values from the 00UTC Oct 20th runs of the ECMWF and GFS, respectively. The GFS_raw, as similarly seen in the above EMC plumes display, had a MaxT=37.1 in that run, while the ECMWFd_raw had 42.5. One can see the subsequent (06UTC) GFS run warmed, and the 07UTC NBM was in the upper 50s.
[ It would be nice to have the option, in the extended ranges to dModel/dt compare just runs from every 6hrs vs each hour; but perhaps that is already possible? ]
In short, very good forecasts for an anomalous cold event so early in the season at KDFW. I believe the 2nd coldest MaxT for the date.
Though not shown by me, there were MANY other impacting events that occurred along with the cold temps at KDFW, such as heavy snow in the TX panhandle, and a crippling ice storm in OK, not to mention Zeta in the central Gulf. All would be worth looking at to compare and contrast model/guidance/NDFD performance. I encourage others to contribute more details from their areas, about other aspects of this event.
