Monthly Archives: February 2013

IFR conditions in Maine

GOES-R IFR Probabilities computed from GOES-East (Upper Left), GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm) (Upper Right), GOES-R Cloud Thickness (Lower Left), Toggle between Suomi/NPP Brightness Temperature Difference (10.8 µm – 3.74 µm) and Day/Night Band “Nighttime Visible” imagery (Lower Right), all around 0815 UTC on 28 February

Weak Low pressure in the Gulf of Maine helped generate IFR conditions over the northeastern United States early in the morning on 28 February 2013.  The brightness temperature difference fields over New England from Suomi/NPP include very sharp cloud edges (also present in the Day/Night band imagery).  Because the GOES-R IFR Probability field also includes information from the Rapid Refresh, it is better able to distinguish fog and low stratus, as present over most of Maine, from elevated stratus, present over western New Hampshire and Quebec.

GOES-R IFR Probabilities computed from GOES-East (Upper Left), GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm) (Upper Right), GOES-R Cloud Thickness (Lower Left), Suomi/NPP Brightness Temperature Difference (10.8 µm – 3.74 µm) (Lower Right), times as indicated

The animation of the imagery, above, demonstrates how the GOES-R IFR probability product can be used to monitor the evolving nature of a low cloud field.  As the low pressure system in the Northeast starts to move away, the fog/low clouds rotate eastward.   Two noteworthy events in the loop are present.  The 0515 UTC imagery (mislabeled as 0510 UTC), contains stray light in the 3.9 µm field, and the traditional GOES brightness temperature difference field is therefore changed significantly, but the GOES-R IFR probability field is not.  Note also that multiple cloud layers exist over coastal Maine and New Hampshire at the end of the animation, but GOES-R IFR probabilities correctly maintain high probabilities in a region where IFR conditions are present and where the traditional brightness temperature difference field does not show a signal consistent with low clouds.

IFR conditions over the upper Midwest

GOES-R IFR Probabilities from GOES-East (Upper Left) at 0832 UTC, along with 0900 UTC observations, Traditional GOES-East Brightness Temperature Difference (10.7  µm – 3.9  µm) at 0832 UTC (Upper Right), GOES-R Cloud Thickness computed from GOES-East (Lower Left), Suomi/NPP Day/Night Band Nighttime visible imagery, 0838 UTC (Lower Left)

Stratus and low clouds persisted over western Minnesota and the eastern Dakotas overnight on 25 to 26 February, and the GOES-R IFR Probability field ably captured the region of lowest visibility.  Note that the IFR probability field extends into northwest Iowa (albeit with relatively lower probabilities).  This is a region where high-level cirrus prevents the traditional brightness temperature difference product from giving useful information about the low levels.  In this region, Rapid Refresh data are used to fill in information and more accurately capture the region of IFR conditions.

As above, but for 0802 UTC for GOES-East Products, and with the MODIS-based IFR probability field at 0801 UTC in the lower left

GOES-R IFR Probabilities can be used with MODIS data as well, and the better resolution (1 km at nadir vs. 4 km at GOES nadir) means the MODIS fields have better small-scale detail.    Note, for exanple, the sharper edge to the IFR probability field in east-central Minnesota.

As at the beginning of the post, except for 0415 UTC (top), 0432 UTC (middle) and 0445 UTC(bottom)

Stray-light issues can influence the 3.9 µm imagery, and therefore the brightness temperature difference field, and therefore the GOES-R IFR Probability field.  In the three images above, Stray Light is noteable in the 3.9 µm at 0432 UTC, but that erroneous information can be de-emphasized in the GOES-R IFR probability field because the Rapid Refresh Data in regions where Stray Light is present may show dryer low levels.

GOES-R IFR Probabilities Refine the Area of IFR Conditions over New Mexico

GOES-R IFR Probabilities from GOES-West (Upper Left), GOES-R IFR Probabilities from GOES-East (Lower Left), GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm) (Upper Right), GOES-R Cloud Thickness of Highest Liquid Layer computed from GOES-West (Lower Right), all from 1100 UTC on 15 February 2015

Because GOES-R IFR Probabilities include information about the near-surface atmosphere in the Rapid Refresh Model (and therefore, through assimilation into that model of surface data, the actual atmosphere), IFR probabilities do a better job of distinguishing elevated stratus from fog/low stratus.  In the image above from 1100 UTC on 15 February, the GOES brightness temperature difference product, the traditional method of alerting forecasters to the possibility of fog, shows a signal over the high plains of eastern New Mexico in regions where observations show high ceilings. The GOES-R Fog/Low Stratus product computed from either GOES-West data (top left) or GOES-East data (bottom left) correctly restricts the possibilities of IFR conditions to regions between the Sangre de Cristo mountains in the north to the Sacramento mountains in the south.

Changing IFR probabilities over Gainesville FL as fog develops

GOES-R IFR Probabilities (Upper Left), GOES-R Cloud Thickness (Lower Left), GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm) Product (Upper Right), GOES-East Visible Imagery (Lower Right), 0500 UTC on 12 February.

This example shows how the IFR Probability Values changed hourly as IFR conditions developed over the north-central Florida peninsula.  IFR Probabilities around Gainesville — KGNV — are below 10%.  These small values persist through about 0545 UTC, but by 0615 UTC, below, they have increased to around 20% although brightness temperature difference returns have not substantively changed.

As above, but at 0615 UTC

By 0702 UTC, below, values continue to increase, to around 30%, but Gainesville continues to report MVFR conditions.  There is a location nearby, however, over eastern Alachua County, where Probabilities exceed 70%.  Abundant surface waters in eastern Alachua County — Newnans Lake, Santa Fe Lake, Ledwith Lake, Orange Lake and Lochloosa Lake — may have contributed to the initial fog development there.  Note also that the traditional fog detection product, the brightness temperature difference, is also starting to show enhanced returns.

As above, but at 0702 UTC.
As above, but at 0802 UTC

Between 0700 and 0800 UTC (above), probabilities at KGNV increase from 31% to 80%.  The increase largely results from increases in the satellite predictors of low clouds/fog, as shown in the rapid increase in the brightness temperature difference signal around Gainesville.  Nevertheless, ceilings and visibilities at KGNV have not deteriorated to IFR conditions at 0800 UTC.  IFR probabilities remain between 70 and 80% from 0800 to 0915 UTC (below), at which time IFR conditions are reported at the Gainesville airport.  From 0915 UTC to 1215 UTC, IFR conditions continue at the airport, and IFR probabilities are from 85% to 95%.

As above, but at 0915 UTC

As above, but at 1002 UTC.

As above, but at 1102 UTC

As above, but at 1215 UTC

Fog and low clouds in the Southeast US

GOES-R IFR Probabilities computed from GOES-East and Rapid Refresh Data, and surface observations of ceilings and visibility, 0645 UTC-0700 UTC on 11 February (Upper Left), GOES-13 Brightness Temperature Difference, 0645 UTC (10.7 µm – 3.9 µm) (Upper Right), Suomi/NPP VIIRS Brightness Temperature Difference (10.8 µm – 3.74 µm) at 0638 UTC (Lower Left), Suomi/NPP Near IR Imagery (3.74 µm) at 0638 UTC (Lower Right)

The image above is a good example of the importance of fused data in many fog/low stratus events.  The near IR imagery, bottom right, shows many different cloud layers.  A strong storm moving towards the East Coast on Monday morning 11 February generated many cloud layers that make the traditional method of fog detection, the brightness temperature difference between 10.7 µm and 3.9 µm, problematic.  Adding information from the model, however, allows the GOES-R product to identify the region of IFR conditions that extends northeastward from central Georgia to central Virginia.

Fatal Crash on I-16 in Georgia: Was it fog-related?

Map of Crash Location, courtesy WMAZ TV in Macon, Georgia

A multi-vehicle accident with fatalities occurred in extreme western Laurens County in central Georgia early Wednesday Morning, 6 February 2013.  (According to this news report, the first crash was around 1310 UTC)  Was fog a factor in this accident?  Fog was reported in and around the scene as first responders arrived.

Laurens County is in the Peachtree City CWA, and the relevant part of the Forecast Discussion at 0000 UTC is as follows: 

000
FXUS62 KFFC 060007
AFDFFC

AREA FORECAST DISCUSSION...UPDATED FOR AVIATION
NATIONAL WEATHER SERVICE PEACHTREE CITY GA
ISSUED BY NATIONAL WEATHER SERVICE BIRMINGHAM AL
705 PM EST TUE FEB 5 2013

.SHORT TERM /TONIGHT THROUGH WEDNESDAY NIGHT/...

FAIRLY QUIET WEATHER THROUGH THE SHORT TERM PERIOD. A WEAK
SHORTWAVE HAS BROUGHT A BAND OF CLOUDS TO THE AREA AND A FEW
SPRINKLES SHOW UP EVERY NOW AND THEN ON RADAR. CLOUDS ARE EXPECTED
TO THIN OUT OVERNIGHT AND FOG COULD BECOME AN ISSUE IN THE FEW
HOURS BEFORE SUNRISE. THIS IS SOMETHING THAT WILL HAVE TO BE
MONITORED THROUGH THE EVENING. MILD TEMPERATURES THROUGH THE
PERIOD.

.LONG TERM /THURSDAY THROUGH TUESDAY/... 
[...snipped...]
&&

.AVIATION... 00Z TAF DISCUSSION.

VFR CIGS WILL CONTINUE ACROSS THE TAF SITES THIS EVENING.
EXPECT MVFR CONDITIONS LATE TONIGHT AS FOG AND LOW CLOUDS DEVELOP
BEGINNING AFTER 06Z NEAR CSG SPREADING NORTH AND EAST. THE ATLANTA
TAF SITES SHOULD EXPECT TO BE IMPACTED BY 08-09Z. IMPROVING
CONDITIONS BY 15Z. LIGHT WESTERLY WINDS EXPECTED. 
 
 So, overnight fog was considered a possibility at 0000 UTC.
And updated AFD was issued shortly before 0600 UTC.  In that update the Aviation discussion was tweaked:
FXUS62 KFFC 060553
AFDFFC

AREA FORECAST DISCUSSION...UPDATED FOR AVIATION
NATIONAL WEATHER SERVICE PEACHTREE CITY GA
ISSUED BY NATIONAL WEATHER SERVICE BIRMINGHAM AL
1245 AM EST WED FEB 6 2013

.SHORT TERM /TONIGHT THROUGH WEDNESDAY NIGHT/... 
 
[...no changes...]
 
.LONG TERM /THURSDAY THROUGH TUESDAY/... 
 
[...snipped for brevity...]
&&

.AVIATION... 06Z TAF DISCUSSION.

PATCHY CEILINGS BKN030-040 LINGERING THROUGH THE NIGHT. POTENTIAL
FOR LIFR CONDITIONS WHERE SKIES ARE CLEAR...BUT EXPECT ALL TAF SITES
TO HAVE AT LEAST MVFR CONDITIONS 08Z THROUGH 14Z. IMPROVING AFTER
THAT TO VFR. WINDS NORTHWEST 5 TO 10KT AFTER 15Z.


//ATL CONFIDENCE...06Z UPDATE...
MEDIUM CONFIDENCE ON ALL ELEMENTS.
 Note that a mention of LIFR conditions has appeared in the Aviation Discussion.  What did the GOES-R IFR Probability show between 0000 and 0600 UTC?
GOES-R IFR Probabilities computed from GOES-East and Rapid Refresh Data, 0115 through 0515 UTC 6 February, along with hourly observations of ceilings/visibility
Especially in the last two hours — 0415 and 0515 UTC — there is a general expansion in high probabilities from west to east across southern Georgia, with probabilities increasing north of I-16.  By 0815 UTC, below, IFR conditions are being reported at many locations in northwest Georgia, and IFR probabilities continue to increase over east-central Georgia.
As in the loop above, but for 0815 UTC only.  Interstate Highways are denoted in Blue.
The AFD issued at 0930 coincided with the issuance of a dense fog advisory (highlighted in red) for most of the CWA:

000
FXUS62 KFFC 060930
AFDFFC

AREA FORECAST DISCUSSION
NATIONAL WEATHER SERVICE PEACHTREE CITY GA
430 AM EST WED FEB 6 2013

.SHORT TERM /TODAY THROUGH THURSDAY NIGHT/...
SOME LOW LEVEL MOISTURE REMAINING THIS MORNING AND THIS HAS PRODUCED
FOG ACROSS MANY AREAS. WILL GO AHEAD AND ISSUE A DENSE FOG ADVISORY
SINCE MOST PLACES SEEM TO BE HEADED FOR VERY LOW
VISIBILITY.
OTHERWISE DRY HIGH PRESSURE BUILDING OVER THE SOUTHEAST TODAY AND
MOVING OFF THE ATLANTIC COAST TONIGHT. NEXT SHORT WAVE COMING ACROSS
THE NORTHERN GULF SHOULD BEGIN TO AFFECT OUR CWA MAINLY 06Z THURSDAY
AND BEYOND. THIS SOUTHERN STREAM SYSTEM MOVES ACROSS THE STATE
THURSDAY AND THURSDAY NIGHT WITH SOME DIFFERENCES BETWEEN MODELS
WITH LOCATION AND STRENGTH OF SURFACE LOW. IN GENERAL...LOOKS LIKE A
WEDGE SETTING UP LATE TONIGHT INTO THURSDAY WITH MUCH OF THE CWA IN
EASTERLY FLOW. THERE SHOULD BE A SHARP GRADIENT OF TEMPERATURES
BETWEEN NORTH AND CENTRAL ZONES AND FOR NOW HAVE KEPT THE WARMER AIR
ON THURSDAY CONFINED TO COLUMBUS TO MACON AND SOUTH. CATEGORICAL
POPS FOR THURSDAY TAPERING OFF WEST TO EAST THURSDAY NIGHT AS THE
SURFACE LOW MOVES EAST. NO MENTION OF THUNDER AS WEDGE SHOULD BE
STABLE. HOWEVER SOME INSTABILITY COULD PUSH INTO THE CENTRAL ZONES
THURSDAY AFTERNOON ALONG THE BOUNDARY.

41


.LONG TERM /FRIDAY THROUGH TUESDAY/... 
[...snipped...] 
01

&&

.AVIATION...
06Z TAF DISCUSSION.
PATCHY CEILINGS BKN030-040 LINGERING THROUGH THE NIGHT. POTENTIAL
FOR LIFR CONDITIONS WHERE SKIES ARE CLEAR...BUT EXPECT ALL TAF
SITES TO HAVE AT LEAST MVFR CONDITIONS 08Z THROUGH 14Z. IMPROVING
AFTER THAT TO VFR. WINDS NORTHWEST 5 TO 10KT AFTER 15Z.

//ATL CONFIDENCE...06Z UPDATE...
MEDIUM CONFIDENCE ON ALL ELEMENTS.

41

&&

.PRELIMINARY POINT TEMPS/POPS... 
[...snipped...]
 &&

.FFC WATCHES/WARNINGS/ADVISORIES...
DENSE FOG ADVISORY UNTIL 9 AM EST THIS MORNING FOR THE FOLLOWING
ZONES
: BALDWIN...BANKS...BARROW...BARTOW...BIBB...BLECKLEY...
BUTTS...CARROLL...CATOOSA...CHATTAHOOCHEE...CHATTOOGA...
CHEROKEE...CLARKE...CLAYTON...COBB...COWETA...CRAWFORD...CRISP...
DADE...DAWSON...DEKALB...DODGE...DOOLY...DOUGLAS...EMANUEL...
FANNIN...FAYETTE...FLOYD...FORSYTH...GILMER...GLASCOCK...
GORDON...GREENE...GWINNETT...HALL...HANCOCK...HARALSON...
HARRIS...HEARD...HENRY...HOUSTON...JACKSON...JASPER...
JEFFERSON...JOHNSON...JONES...LAMAR...LAURENS...LUMPKIN...
MACON...MADISON...MARION...MERIWETHER...MONROE...MONTGOMERY...
MORGAN...MURRAY...MUSCOGEE...NEWTON...NORTH FULTON...OCONEE...
OGLETHORPE...PAULDING...PEACH...PICKENS...PIKE...POLK...
PULASKI...PUTNAM...ROCKDALE...SCHLEY...SOUTH FULTON...SPALDING...
STEWART...SUMTER...TALBOT...TALIAFERRO...TAYLOR...TELFAIR...
TOOMBS...TOWNS...TREUTLEN...TROUP...TWIGGS...UNION...UPSON...
WALKER...WALTON...WARREN...WASHINGTON...WEBSTER...WHEELER...
WHITE...WHITFIELD...WILCOX...WILKES...WILKINSON.

&&

$$

SHORT TERM...41
LONG TERM....01
AVIATION...41

The animation from 0815 UTC through 1315 UTC shows a continued increase in the probabilities over the crash site on I-16.  In addition, observations decrease to IFR conditions.  Note that at 1315 UTC the switch between nighttime and daytime predictors is present in the image as a southwest to northeast boundary over extreme eastern Georgia.  Probabilities are high over the crash site, but by 1415 UTC, they have all but vanished as the fog quickly dissipated at sunrise.  This suggests a fog that is not thick, and the GOES-R Cloud Thickness corroborates this assumption (see below)

As in the loop above, but from 0815 UTC through 1315 UTC.  Crash occurred at 1310 UTC
GOES-R Cloud Thickess just before Twilight Conditions, 1215 UTC 6 February.  Values near the Crash Site are 600-700 feet.
The Cloud Thickness field, above, at 1215 UTC, just before twilight conditions, shows values around 600 to 700 feet.  This chart suggests a rapid dissipation time of around 1 hour.  Indeed, visible imagery around  sunrise shows scant evidence of widespread fog.
Zoomed-in GOES-13 Visible Imagery over east-central Georgia, times as indicated.  Montrose (Yellow square) is indicated within Laurens County (outlined in green)
Polar orbiters also provide information — at higher spatial resolution — about the evolving situation overnight.  Suomi/NPP made two passes over Georgia between 0600 and 0900 UTC, and the brightness temperature difference product that was produced is shown below.    There is a noticeable increase in water clouds over Georgia between 0632 UTC and 0811 UTC, the times of the images.

Suomi-NPP Brigthness Temperature Difference (10.80 µm- 3.74 µm) at 0632 UTC and 0811 UTC.  Interstate 16 crosses Laurens County near the center of these images.

MODIS data from Aqua were used to produce IFR probabilities at 0645 UTC, below.  This was before the most fog had developed, but it does confirm the picture painted with GOES data’s broader brush:  IFR probabilities are increasing over Georgia.

MODIS-based GOES-R IFR Probabilities over Georgia, 0645 UTC on 6 February 2013.

Finally, AVHRR data from NOAA-15 from 1015 UTC show widespread stratus extending eastnortheast from southwest Georgia to central Georgia.

Brightness Temperature Difference (10.8 µm – 3.74 µm) from AVHRR data, 1028 UTC 6 Feb 2013

Fog Dissipation Example over California

GOES-R Cloud Thickness over California, 1445 UTC on 5 Feb 2013

GOES-R Cloud Thickness can be used to esimate when radiation fog will dissipate.   In this example from central California, radiation fog has developed to a depth of around 1000 feet near Hanford, 1000 feet just southwest of Fresno, and 1100 feet near Merced.  This chart shows the relationship between Cloud Thickness and burn-off time.  1000 feet correlates well with a 3-hour burn-off time;  1100 feet correlates well with a 4-hour burnoff time.  The animation below, with imagery at 1800, 1830, 1900 and 1930 UTC shows that the fog around Hanford was slow to burn off — by about an hour.  Fog near Merced was also slow to burn off, but high clouds moving in may have been responsible for that delay.

GOES-15 Visible Imagery over Central California.  Times as indicated.

GOES-R vs. Heritage GOES Fog Products in Arkansas

GOES-R IFR Probabilities, from GOES-East (0732 UTC) and 0800 UTC Surface observations of visibility and ceilings (Upper Left), GOES-E Brightness Temperature Difference (10.8 µm – 3.9 µm) (Upper Right), GOES-R Cloud Thickness (Lower Left), MODIS-based Brightness Temperature Difference (11µm – 3.74 µm) GOES-R IFR Probabilities computed using MODIS data (Lower Left, 0739 UTC)

IFR conditions developed over Arkansas and surrounding states overnight from 4 into 5 February.  Compare the brightness temperature difference (the traditional fog-detection product) over southeast Arkansas (where IFR conditions are not occurring) and over southwest Arkansas (where IFR conditions are present).  Although the satellite signal is very similar over the region, surface observations are very different.  The GOES-R algorithm distinguishes between the region with IFR conditions (east Texas, western Arkansas, northwest Louisiana) and the region without IFR conditions (southeast Arkansas, northeast Louisiana).

On the flip side, in regions over northeast Arkansas, where the brightness temperature difference product is not showing low clouds, IFR conditions are present, and the GOES-R IFR probability is elevated.

GOES-R Cloud Thickness over Arkansas just before Dawn — note that dawn has arrived over Tennessee and Mississippi

Cloud Thickness just before twilight conditions can be used to predict when radiation fog will burn off, using this scatterplot as a guide.  The maximum thickness over south-central Arkansas is 1350 feet, and that thickness corresponds to 5 hours after sunrise, or sometime after 1800 UTC.  The animation of visible imagery, below, shows that fog/low clouds are lingering over parts of southern Arkansas.

GOES-13 Visible Imagery over Arkansas, times as indicated.

California Fog

GOES-R IFR Probability computed from GOES-West (Upper Left), GOES-R Cloud Thickness computed from GOES-West (Lower Left), GOES-West Brightness Temperature Difference (10.7 µm – 3.9 µm) (Upper Right), GOES-West 10.7 µm imagery (Lower Left) at 0400 UTC 4 February

Fog develped over the San Joaquin and Salinas Valleys of California early on 4 February.  At 0400 UTC, the Brightness Temperature Difference signal shows a noisy signal over the east part of the San Joaquin Valley, with a more coherent signal off the coast and over Kern and Kings County.

As above but at 0900 UTC.

By 0900 UTC, five hours later, although the seemingly noisy signal continued over the California in the brightness temperature difference product, the GOES-R IFR probability field is starting to show higher values aligned through the San Joaquin and Salinas Valleys, where IFR conditions have developed.

As above, but at 1200 UTC

At 1200 UTC, above, high IFR probabilities extend through the Salinas Valley and in the San Joaquin valley where IFR conditions are noted.  IFR probability is also high over San Francisco Bay where marine stratus has moved inland.

As above, but at 1500 UTC

The regions of reduced visibility continue at 1500 UTC, the last image before twilight conditions disallows computation of Cloud Thickness (indeed, the terminator is apparent in the image).  The cloud thickness of 1100 feet suggests, based on this scatterplot, a dissipation time of around 4 hours.  The animation below shows visible imagery at 1730, 1830 and 1930 UTC that aligns with the predictions.

GOES-15 0.62 µm Visible Imagery, times as indicated.

Suomi/NPP VIIRS data overflew this region twice during the night, and provided brightness temperature difference information at high spatial resolution.  The GOES-R IFR algorithm is not yet applied to Suomi/NPP data (like it is to MODIS data) however.

As at top, but with Suomi/NPP Brightness Temperature Difference (10.8 µm – 3.74 µm) in the lower right, at ~0900 UTC.
As above, but at 1030 UTC.

Fog in California’s Central Valley

GOES-R IFR Probabilities (Upper Left), GOES-West Brightness Temperature Difference (10.7 µm – 3.9 µm) (Upper Right), GOES-R Cloud Thickness (Lower Left), GOES-West Water Vapor imagery (6.7 µm)

Fog developed in the early morning of February 1, 2013, in California’s Central Valley, the combination of the San Joaquin valley to the south and the Sacramento Valley to the north.  The imagery above shows the GOES-R Fog/Low Stratus product and the traditional fog product, the brightness temperature difference between 10.7 µm and 3.9 µm.  The GOES-R Product (IFR Probability) is first in highlighting the development of fog near the San Joaquin River and reductions in visibility occur in sync with the increase of IFR probabilities. The traditional GOES-West brightness temperature difference product displays considerable signal in the first hours of the animation above, but there is no organization to the signal.  Eventually, however, the brightness temperature difference signal does include the fog and low stratus in the Valley.