The Satellite Proving Ground at the Hazardous Weather Testbed

Tuesday, May 7, 2013

EWP daily debriefing 5/7

This afternoon we had our first EWP daily debrief from the previous day's activities. Yesterday the focus was on the mid-Atlantic coast, specifically the Blacksburg, VA CWA, and while forecasters did issue a handful of warnings, it was an overall quiet day. However, the more docile weather allowed them a chance to work through each set of new data, learning which products they have at their disposal and how to use them. Here are a few comments picked up throughout the day:

Simulated Imagery
- 'We used the simulated imagery to forecast the dissipation of storms in the evening.' This along with the Nearcasting indicated the dissipation of storms in the area of interest.
- 'The WRF simulated imagery seems to be more spotty; it doesn't often seem to forecast the broader cloud shields associated with storm systems.'

Nearcasting Model
- 'The theta-e difference showed stable air moving into the area' and along with the simulated imagery, was used in the forecasting of the dissipation of the storms.
- 'I liked the two and three hour forecasting'

GOES-R CI and Cloud Top Cooling
- Because of the more docile weather 'there were no 80/90% CI values so it made the product more difficult to use'

RGB Airmass
- 'It takes some time to intuitively understand exactly what you're looking at with this product and how it can be integrated into your forecasting techniques, but it has potential'

Training
- Forecasters completed a training shift previous to their arrival in Norman. Monday was then used as a low key forecasting day, giving them a chance to get familiar with the products in an operational setting
- 'Having a chance to interact and use the products on a less busy day was very helpful'

Stay tuned to the blog for further feedback and updates on daily activity!

Monday, May 6, 2013

Slow day, but good ending for UWCTC

The first day of the HWT was slow weather-wise for the UWCTC product. Most of the convection in the area of focus was not producing any captured CTC signal due to the primarily thicker layers of cirrus over the top of the ongoing convection. However, toward the end of the shift when focus was shifted elsewhere, some promising CTC signal showed itself. At 2302 UTC, two moderate cooling rates were detected over north central North Carolina (See Figure 1).

Figure 1. UWCTC overlaid on visible satellite imagery at 2303 UTC, with a max UWCTC value of ~-8K/15 min.

Cooling rate detection was ongoing for these two areas through 2325 UTC (See Figure 2).

Figure 2. UWCTC overlaid on visible satellite imagery at 2325 UTC, with a max UWCTC value of ~-13K/15 min.

As seen on the Hi-Res VIL product on radar (Figure 3), the left detection of CTC went on to produce VILs in excess of 40-45 kg/m-2.


Figure 3 - High Resolution VIL valid at 0035 UTC.

The first hail report for this particular storm was received at 0040 UTC, lending a one hour and thirty-eight minute lead time ahead of the observed hail. See Figure 4 below for the LSR text.

Figure 4. LSR of estimated 1" hail received at 0040 UTC.

EWP: 00Z outlook and the Nearcasting model

A cluster of storms continues ahead of a diffuse surface boundary located across west-central North Carolina into southwest Virginia. Convergence continues to be the primary convective driver in association with an upper-level low continuing to rotate across eastern Tennessee.

Current radar imagery overlaid with surface analysis

Cold temperatures aloft, as shown by H50 temperatures near -20C, and moderately steep lapse rates will continue to support a marginal threat for large hail through 01Z. Near-surface instability will decrease in the next couple of hours as diurnal heating comes to an end and low-level theta-e values decrease as depicted by the CIMSS Nearcast tool. Tend to side with simulated satellite imagery showing storms coming to an end by 03Z across the Blacksburg County Warning Area.

Nearcasting showing decreasing instability this evening

Simulated Imagery at 03Z as storms come to an end

Hampshire/Guseman

EWP underway

This afternoon the EWP has been focused on the ongoing convective activity over the Eastern U.S., specifically in the Blacksburg, VA CWA. Forecasters have been utilizing the convective initiation and cloud top cooling products to identify areas of developing convection within the broader system, a challenge given that conditions aren't very conducive for robust cell development.

EWP: Simulated Imagery capturing convective activity

Convection is ongoing across North-Central North Carolina westward into the Appalachians. The upper low is currently centered in eastern Tennessee. The most robust thunderstorms of the afternoon have developed across northwest Carolina along a a boundary extending from NW North Carolina into eastern Kentucky.

Current radar imagery overlaid with surface dewpoint

The environment is not overly robust to sustain severe convection. Surface-based CAPE values across Northern North Carolina are roughly around 1000 J/Kg with 0-6 km shear 40-50 knots. Low-level lapse rates will be approaching 6-7.5 C/Km by 21z with mid-level lapse rates around 6-7 C/km. The main threat will be marginally severe hail, but 0-3 km helicity values between 150-200 m2/s2 could support a low-end tornado threat. However, as storms move away from the boundary, this threat will decrease.

The EFP has placed a 5% probability of severe storms for the previously mentioned area, while a slight risk of severe storms is forecasted by SPC in small area shown in the image below.

As a lobe of energy wraps around the upper-low and moves northward into northern NC, the storms which have already initiated should be able to continue northward into the Blacksburg, VA CWA while some continuing activity is possible across the northwestern extent of the Raleigh, NC CWA. This activity is noted in the 22Z simulated satellite imagery as shown below.

22Z WRF simulated IR imagery

Hampshire/Guseman

Week 1: 2013 Spring Experiment

The 2013 Spring Experiment has begun! This year's experiment will run three weeks, ending on May 24th, and though shorter than previous years, we still look forward to a productive demonstration. Participants include 24 NWS forecasters and a variety of visiting scientists, and as with previous years, the experiment contains both Experimental Forecast Program (EFP) and Experimental Warning Program (EWP) activities with a focus on the forecast and warning of convective initiation and severe weather.

The GOES-R portion of this demonstration will for the most part occur within the EWP. Products will include synthetic satellite imagery from the WRF-ARW, 0-6 hour GOES sounder-based 'Nearcasts', sounder-based RGB Airmass, Pseudo Geostationary Lightning Mapper total lightning and other applications, convective initiation nowcasting, and cloud top cooling rates.

Activities will be ongoing from roughly 8am to 10pm, with most emphasis on severe weather warning occurring in the later portion of the day. Feedback will once again be forecaster driven, stemming from real-time interaction and exploration of GOES-R products in this pre-operational environment, and will be visible via this blog. Additionally visiting scientists and PIs will also be asked to share their experiences. Finally, each week will end with the the typical Tales from the Testbed webinar, where forecasters will have a chance to share their overall thoughts from each shift in a summary presentation. The focus of this year's webinar topics have been pre-selected with week 1 concentrated on GOES-R, particularly the uses of the PGLM in severe weather forecasting.

Stay tuned to the blog for real-time updates of our activities and forecaster feedback!

~AMT

Sunday, November 25, 2012

Day/Night Band Provides Better Detail of Low Clouds

During the overnight hours of 23-24 November, 2012, persistent stratus and flurries lingered over much of Wisconsin in the wake of a strong cold front which had swept through the previous day. As indicated by the below Green Bay 00z sounding, the stratus was shallow and only several hundred feet thick. Cloud bases were 2 to 3 thousand feet.

The short-term forecast concern was timing of the cloud erosion. The below CONUS wide view 11-3.9 micron imagery from 08z 24 Nov indicated widespread clouds blanketing most of Wisconsin and the western Great Lakes.

0758z 24 NOV 2012 11-3.9 micron imagery

At about the same time, the below NPP VIIRS Day/Night band CONUS view indicated thin spots and breaks in the cloud cover across Wisconsin and the western Great Lakes.

0758z 24 Nov 2012 Day/Night Band

Taking a closer view of the 11-3.9 micon imagery below, the clouds appear to shroud all of Wisconsin except for the far northwest corner.

11-3.9 micron imagery - closeup

But a closer view of the Day/Night band showed more breaks in the cloud cover across Wisconsin as well as thin spots. The clearing over northern Lake Michigan and Door County was much more discernable using the D/N band.

Also, the illumination (reflectance) from the metropolitan areas highlighted the thinness of the clouds as well as helped make the breaks more discernable.

Day/Night Band Closeup

By using the Day/Night band, a clearer picture of the cloud cover covering Wisconsin was determined. This better understanding would allow the short-term forecaster to make a more reliable forecast of clouds and timing of low cloud erosion.

The moon was in waxing gibbous and was 85.3% lit on 24 Nov 2012.

Marc Kavinsky
National Weather Service - Milwaukee/Sullivan

Tuesday, November 6, 2012

Using GOES-R IFR/LIFR Probabilities for TAFs

Tracking areas of stratus and fog can be somewhat difficult using traditional satellite images along with METAR observations...particularly on a day with a variety of low and high level clouds. Sky conditions across the Great Lakes during the afternoon of November 6th is a good example. Below are infrared and visible GOES images along with plotted METAR ceilings and visibilities:

GOES Infrared

GOES Visible

Note how conditions are evident at point locations, but it is difficult to discriminate areal coverage across the region. For the purpose of issuing and updating TAFs, it would be very useful to better visually display areas of lower ceilings and visibilities, increasing situational awareness and improving the ability to track and forecast improving or deteriorating conditions. Fortunately, GOES-R fog and low clouds products are proving quite useful in filling this need. Below are images of GOES-R IFR probabilitly and LIFR probability:

GOES-R IFR Probability

GOES-R LIFR Probability

The IFR product does a very good job of showing the area of ceilings that are less than 1000 feet, particularly for values above 60 percent. Though LIFR percentages are not very high, the highest percentages do coincide pretty well with some of the lower ceilings and visibilities in the region. Something that was immediately useful was observing a loop of the IFR image, which showed that the eastern edge of the lower ceilings had temporarily stalled across eastern Wisconsin. This might lead to delaying the arrival of IFR ceilings in eastern TAF sites.

Another likely use of the GOES-R fog and low cloud products would be tracking stratus or fog advecting into Wisconsin off of Lake Michigan. Using an example off the California coast, note that GOES infrared and fog images will give you an idea of where there is cloud cover, but will not give and immediate idea of what type of conditions are moving into the area:

GOES Infrared

GOES IR Fog

While the traditional fog product above does give a clue that you're looking at either fog or stratus...the GOES-R IFR and LIFR products confirm this and also give an indication of how low the ceilings and/or visibilites are:

GOES-R IFR Probability

GOES-R LIFR Probability

It is immediately clear the conditions are at or below IFR for much of the area of fog/stratus...with a pretty good probability of LIFR conditions over a large portion of the area. This is quite useful...as ceiling and visibility observations are very limited over bodies of water. Hopefully the use of this product will help in troublesome situations where poor conditions can sneak onshore... particularly when somewhat obstructed by other cloud decks.

In summary, GOES-R fog and low cloud products can be very useful in tracking areas of fog and stratus. In contrast to using a combination of surface observations and traditional visible/infrared products, the GOES-R products can give a very quick and fairly accurate representation of where IFR or worse conditions are located. Being able to track these areas will hopefully lead to improvements in TAFs.

Denny VanCleve
NWS Milwaukee/Sullivan

Tuesday, October 23, 2012

GOES-R AWG Fog/Low Cloud Products

Conditions across the middle of the U.S. were ideal for assessing this product. Strong low pressure was forming over the High Plains while high pressure was situated along the Atlantic Coast of the southeast U.S. Warm air and copious moisture were flowing north up the Mississippi Valley. This morning we focused on the southern coast of Louisiana. The Louisiana case looked almost completely radiation induced.

When I was introduced to this product during the GOES-R proving ground activities of 2011, I was, to be honest, rather underwhelmed. Back then, this product used satellite interrogation coupled with boundary layer RH from the GFS. No observations were included in the assessment and therefore it didn't seem to perform any better than my simply using available satellite imagery and surface obs (and experience.) This year, it appears there has been significant improvement to the inputs of this product. The boundary layer RH now comes from the Rapid Refresh Model. Not only is the RH data resolution improved both spatially and temporally, but the Rapid Refresh incorporates the latest surface observations into its output fields. From the few examples I saw, it is apparent this product will sometimes highlight an area, showing increased probabilities of IFR/LIFR, before these conditions appear in the observations. This appears to be the strength of using the high resolution Rapid Refresh output.

Anyway, here is the example down along the southern coast of Louisana:

The forecaster working on the 06z TAFs would have used the GFS and NAM MOS (MET/MAV) as part of the decision aid on ceiling and visibility. In this example, we'll look at the observation in far southeast TX, MTRJAS. See location in image below:

Here are the METJAS and the MAVJAS products respectively at 00z Oct 23rd, 2012:

_________________________________________________________________________________

Note the ceiling prediction shows VFR conditions dropping to MVFR from 06z-12z Oct 23rd. No real hint of IFR, though the temp/dewpoint spread does get pretty tight in the morning.

Below are a series of 4-panel images running from 04:02Z-09:15Z Oct 23, 2012 showing a combination of GOES-R and heritage products. See caption on first image for details. Note the GOES-R IFR probability is beginning to highlight far southeast TX with increasing probabilities of IFR at 04:02Z. MTRJAS at this point is indicating clear skies and 10 mile visibility.

GOES-R IFR Probability upper left, LIFR probability upper right, heritage brightness temperature difference product lower left, visibible image lower right

The following image is an hour later at 05:02Z: MTRJAS is still clear and 10 miles, but note the increasing probabilities of IFR.

At 06:15Z: MTRJAS is now indicating Clear and 5 miles in fog...probabilities continue to show a trend toward IFR

At 07:15Z: MTRJAS is Clear and 7 miles visibility.

At 08:15z: MTRJAS has now gone down to a 200ft broken ceiling and 7 miles visibility.

And finally at 09:15Z: MTRJAS dropped to as low as 200ft overcast and 2 miles, before seeing improving conditions after about 09:30Z.

In summary, it is clear that, in this case, the GOES-R IFR probability products can give the forecaster an early heads-up that conditions are on their way down to IFR conditions, despite the traditional MOS products from the GFS and the NAM indicating no IFR conditions for that METAR site.

Steve Davis
CIMSS-MKX

Tuesday, September 4, 2012

Cloud Top Cooling Skillful before Severe Storms

Today's severe thunderstorm warnings were preceded by cloud top cooling rates of around 30 degrees Celsuius per 15 minutes -- around 30 minutes before warnings were actually issued!

Here is the time series of cells that developed over Winona County, Minnesota. They produced golf-ball size hail. The panels, from upper left to upper right and lower left to lower right, are: cloud top cooling with visible satellite image background, convective initiation categorical classification with visible satellite image background, IR window satellite image, and radar base reflecitivity.