Problem with Precipitation Type

Forums (Forecast Builder)

RSS (Opens New Window)

Phil Schumacher, modified 8 Years ago.

Problem with Precipitation Type

Youngling Posts: 11 Join Date: 9/24/12 Recent Posts

Hi there,

Over the last week, WFO FSD has had several mixed precipitation events. In the process of doing our grids in both the short term and extended forecast we have noticed some problems with the precipitation type being created. There are two issues:

1.) On the mid shift, Tuesday, November 22, a mix of freezing rain, rain, and sleet was forecast for 1st period in southwestern Minnesota and northwestern Iowa includiing the FSD, MPX, and DMX CWAs. Despite having well coordinated surface T grids, WFO MPX was forecasting FZRA in southwest Minnesota while in an adjacent county WFO FSD was forecasting sleet. The forecaster (me) attempted to raise the MaxTxAloft approximately 1C warmer along the common border with WFO MPX and WFO DMX and sleet was still forecast for the FSD CWA in southwest Minnesota. Once the forecast was completed, WFO MPX ran the ptype process for forecast builder using FSD MaxTxAloft and got freezing rain in the FSD CWA. When WFO FSD ran the FB ptype process using the MPX temperatures, they received sleet in the MPX CWA. WFO FSD did not populate the ProbIcePresent for the event. Talking to the MPX and DMX SOO, it is not common practice for their offices to populate that grid either although neither were sure if the grid was populated for that shift. So it is possible that was the issue although we are not certain it was the issue and thought we should share. Note that FSD has not installed the latest version of FB due on December 16.

2.) This morning we noticed that weather grids at Day 5 were reverting back to an old grid resulting in an inconsistent ptype. This example was outside our CWA so did not impact the actual NDFD grids but has been observed by forecasters previously. In this case, the MaxTwAloft and surface Temperature are both below freezing and now ProbIcePresent grid was initialized. Running the ptype process of FB results in a 100% of snow as would be expected. However, when running the process to populate the Wx grid, it initially loads the correct ptype (snow). However, when it finishes the process and loads in the entire Wx Element group, the Wx grid reverts back to an older version. The result is a few hours of rain in the Wx grid.

Examples of both problems are documented in the attached PowerPoint. Please feel free to call if you have any questions.

Phil

Andy Just, modified 8 Years ago.

RE: Problem with Precipitation Type

Youngling Posts: 89 Join Date: 6/2/15 Recent Posts

Phil,

First, definitely have your office install the CR16-034 tech order to get the ForecastBuilder updates in.

On Tue Nov 22, I'll discuss about this event on today's webinar. Surface wet-bulb was playing a big factor, requiring both T and Td to be consistent. Definitely ProbIcePresent can throw a monkey wrench into the equation, as ProbIcePresent values < 80 will really begin to switch your Ptype over to freezing rain. Also, within the logic, if MaxTwAloft > 1C (which it was) and T > 32, then the assumption is rain (with freezing rain occurring if Td < 32 and 32 < T < 35). I also wonder if you had any ProbRefreezeSleet grids present that differed from office to office.

For the Wx grid revert - first we'll talk on the webinar about population of days 4-7 top down grids. Back to the revert question, unfortunately I don't have an answer for it. I've never seen rain produced in areas where T < 32 unless there was a RoadTemp grid present with values > 32. Let me know if the problem continues after your office installs CR16-034.

Phil Schumacher, modified 8 Years ago.

RE: Problem with Precipitation Type

Youngling Posts: 11 Join Date: 9/24/12 Recent Posts

Hi Andy,

Thanks for the quick feedback. We are installing CR16-034 today so, hopefully, that will fix the Wx grid revert issue. Once we finish the install we will try to run the same time period and see if the Wx grid comes out correct.

Regarding, November 22, we (FSD) did not have the ProbRefreezeSleet grid present since the thermal profile was essentially isothermal to the surface. So our sleet was the result of MaxTwAloft being between 2 and 2.5 C. With Tw(sfc) < 32 and the other two grid elements not filled, the ptype will go to sleet based upon the curve that is in FB as I understand it. I am thinking (hoping) the other offices did populate the ProbIcePresent grid and the result was ZR for MaxTxAloft < 2.5 C because then the difference (sleet vs freezing rain) is completely understandable.

There was another issue with ptype on November 22 that occurred with the ice accumulation grid. During that event there was a persistent southeast wind that kept advecting dew points in that were in the upper 20s to around 30. We were forecasting temperatures from 32 to 34 degrees so the result was a Tw below freezing most of the day. Because of that, our IceAccum grid gave almost 0.25" of ice accumulation in parts of northwestern Iowa. I understand how the Tw impacts ice accumulation but given air temperatures above freezing and from talking to a couple of customers on the mid shift, we know that if there was ice accumulation was very isolated and inefficient (just a few places with a glaze of ice despite light rain much of the night). Given observed meteorological conditions, the expectations that the wet bulb temperature would remain below zero seemed reasonable (here a link to the observations at Jackson, MN - dew points reached 32 F around noon but Tw remained below 34 F all day so even a perfect forecast would have had R/ZR I think). My thought was that any ice accumulation would be low because the efficiency of ice accretion was small due to the air temperature remaining above freezing. After spending more time than I like to admit trying to modify the temperature and dew point to lower the IceAccum, I ended up just using the pencil tool to decrease the IceAccum grid to a few hundredths of an inch. Then we just messaged that there would be a glaze of ice primarily on branches and exposed surfaces like bridges. As far as I know, ice accumulation was very limited in southwestern Minnesota and northwestern Iowa.

I think the FRAM is a huge improvement on ice accumulation so I hope this does not sound critical of that model. However, this case resulted in a significant overforecast of ice accumulation that would have warranted an advisory or possibly an ice storm warning. Do you think there is a way to decrease the efficiency of ice accumulation in those cases where the air temperature is expected to be above freezing even with a colder wet bulb temperature?

Thanks!

Phil