Dense Fog over Southern Iowa

GOESRIFRanim_0315_1215

GOES-R IFR Probability Fields, hourly from 0315 through 1215 UTC on 18 May 2016 (Click to animate)

Dense Fog Advisories were issued by the Des Moines Forecast Office on 18 May as fog developed across southern Iowa. The IFR Probability Field animation, above, articulates where the fog is developing — most of the stations reporting IFR (or near-IFR) conditions are within the region of enhanced IFR Probability, and regions where IFR conditions did not develop show persistent low values of IFR Probability.

Compare the 1015 UTC image of GOES-R IFR Probability and the GOES-13 Brightness Temperature Difference (a traditional way of detecting the presence of low clouds/fog at night), below.  Because the IFR Probability fields incorporate information from the Rapid Refresh Model about low-level saturation, IFR Probability screens out the region in central and northern Iowa where the Brightness Temperature Difference field suggests clouds might be forming (but where surface observations show scant evidence of cloud).

GOES_IFR_BTD_toggle_1015

GOES-R IFR Probability and GOES-13 Brightness Temperature Difference (10.7 µm – 3.9 µm) fields, 1015 UTC on 18 May 2016 (Click to enlarge)

The fog that developed over Iowa was fairly thin and should dissipate quickly under a strong May sun. The GOES-R Cloud Thickness field from 1045 UTC, the last complete field before twilight conditions, below, shows values mostly less than 700 feet with a few pockets of 800-900 feet.  The toggle below shows the GOES-R Cloud Thickness at 1045 UTC and the IFR Probability field 90 minutes later.  Regions with the thickest fog at 1045 UTC do show a persistent signal in the IFR Probability field 90 minutes later.

GOESR_CloudT1045_IFRP_1215toggle

Cloud Thickness and Dissipation Time

CloudThickness_0945_05May2016

GOES-R Cloud Thickness, 0945 UTC on 5 May 2016. Note that Cloud Thickness is not computed in the northeast corner where twilight has begun. This is last scene for northern Ontario and Michigan. (Click to enlarge)

GOES-R Cloud Thickness can be used to make a first guess of when fog and low clouds will dissipate. This is done via a look-up table that is derived from this scatterplot. The y-axis on that plot is the last GOES-R Cloud Thickness field produced before sunrise, such as that shown above, and the x-axis is the number of hours after the plot that clearing is expected, GOES-R Cloud Thickness relates 3.9 µm emissivity to dissipation time based on SODAR observations from off the West Coast, and the scatterplot was derived mostly from observations over the southeast US;  the thickness is that of the lowest water-based cloud field and it’s not computed where multiple cloud layers exist — near Geraldton north of Lake Superior, for example, where IFR Conditions are reported — or near sunrise/sunset.  Values are between 1100 and 1200 feet over northern Lower Michigan,  around 1200 feet over eastern upper Michigan, and around 1200 to as much as 1490 feet over northwest Quebec.  The scatterplot suggests a dissipation time of nearly 4 hours, which would be 1345 UTC.

Imagery below shows low clouds persisting just past 1500 UTC.  For this region where the sun angle is not as high as over the southeast US (where most of the observations used for the scatterplot creation were taken), burn-off took a bit longer.  However, note that GOES-R Cloud Thickness did highlight the thickest clouds that took the longest to dissipate;  so, although the scatterplot underestimated the time of dissipation, Cloud Thickness values did identify which regions would clear last.

A final note: GOES-R Cloud Thickness Dissipation times were computed for Radiation Fog Events, such as the one on 5 May 2016. Dissipation of other fogs that create IFR conditions — Advection Fog, Tule Fog — will not be forecast well with this technique.

GOES13VISLoop_1115_1515_5May2016

GOES-13 Visual (0.63 µm) imagery, 1115-1515 UTC on 5 May 2016 (Click to enlarge)

Low Ceilings and Reduced Visibilities over the Upper Midwest

IFRP_0515_1315anim

GOES-R IFR Probability, 0515-1315 UTC on 26 April, with surface observations of ceilings and visibility (Click to enlarge)

GOES-R IFR Probability fields, above, expand southwestward across the upper midwest as ceilings lower and visibilities reduce.  The fields offered quick guidance on where the lowest ceilings were occurring and how the field of low clouds was evolving.  After sunrise (1215 and 1315 UTC imagery), IFR Probability values increased but continued to show a coherent signal over the region of lowest ceilings and smallest visibility.

The Brightness Temperature Difference fields for the same times, below, have structures that have echoes in the IFR Probability fields. The depiction of low ceilings and visibilities associated with the largest brightness temperature difference values (the deepest orange-red in the enhancement) is lost, however, as reflected 3.9 µm radiation alters the brightness temperature difference field. By 1315 UTC, the end of the animation, only the IFR Probability field is giving useful information about the low ceilings and reduced visibilities.

BTD_0515_1315anim

GOES-13 Brightness Temperature Difference (10.7 µm – 3.9 µm) Fields, hourly from 0515 through 1315 UTC, 26 April 2016 (Click the enlarge)

Multiple Cloud Layers and Topography

IFR_1315_22Apr2016

GOES-R IFR Probability Fields, 1315 UTC on 22 April 2016, along with surface observations of ceilings and visibilities (Click to enlarge)

The 1315 UTC image of GOES-R IFR Probabilities, above, shows an axis of higher probabilities aligned with the topography of the Sierra Nevada. Note that Blue Canyon (KBLU) is the sole station reporting IFR Conditions. Did conventional satellite data capture this event? The Water Vapor (6.5µm) and Brightness Temperature Difference fields (10.7µm – 3.9µm), below, do not show evidence of low clouds;  indeed, the cirrus signature in the water vapor must mask any satellite observation of low clouds banked along the Sierra Nevada. Thus a fused product that combines model data and satellite data (such as IFR Probability fields) must be used, and the relatively flat nature of the IFR Probability field above confirms that Rapid Refresh information on low-level saturation is the reason why IFR Probability values are elevated along the mountains.

WV_BTD_1315_22Apr2016

GOES-15 Water Vapor (6.5 µm), left, and GOES-15 Brightness Temperature Difference Field (10.7 µm – 3.9 µm), right, at 1315 UTC 22 April 2016 (Click to enlarge)

IFR Probability Fields earlier in the night did have a satellite component to them. The values at 0300 and 0600, below, show the gradual encroachment of cirrus from the south and west over the low clouds along the Sierra Nevada. After 0600 UTC, only model data were used over the Sierra as high-level cirrus blocked the satellite view.

IFR_BTD_03_06_22Apr2016

Brightness Temperature Difference (10.7µm – 3.9µm) Fields (Left) and GOES-R IFR Probaility Fields (Right) from 0300 (Top) and 0600 (Bottom) on 22 April 2016 (Click to enlarge)

Reduced Visibilities under multiple cloud layers in the Northern Plains

GIFR_P_1115_18Apr2016

GOES-R IFR Probabilities at 1115 UTC and 1100 UTC surface observations of ceiling and visibility (Click to enlarge)

The image above shows GOES-R IFR Probabilities over North/South Dakota and Minnesota shortly before sunrise on Monday April 18 2016.  There is a distinct difference in the field between western Minnesota and eastern North and South Dakota that occurs because Rapid Refresh model fields (low-level saturation) are used as a predictor of IFR Probability.  Reduced visibilities and ceilings are reported where IFR Probabilities exceed 50% (the orange shading).  In contrast, the window channel and water vapor imagery for the same time, below, gives little indication that fog and low ceilings are present over eastern North and South Dakota:  satellite views of the lowest levels are blocked by mid- and upper-level clouds.   Fusing model data and satellite data into one predictor yields a superior product for detection of low ceilings and reduced visibilities.

3PanelToggle_1115anim.IR4IR3IFRP

GOES Infrared imagery (10.7 µm and 6.5 µm) and GOES-R IFR Probability fields, 1115 UTC on 18 April (Click to enlarge)

Fog along the Gulf Coast in Louisiana

GOESBTD_0245_1245anim_05Apr2016

GOES-13 Brightness Temperature Difference Fields, every two hours from 0245 to 1245 UTC, 5 April 2016 (Click to enlarge)

Fog developed along the Gulf Coast of western Louisiana overnight. Brightness temperature difference fields have been used in the past to diagnose the development of fog. This capability exists because water-based clouds, such as stratus, have different emissivity properties at 3.9 µm and 10.7 µm. A water-based clouds does not emit 3.9 µm radiation as a blackbody (but it does emit 10.7 µm radiation as a blackbody). Consequently, the computed temperature of the cloud based on the detected 3.9 radiation is cooler than the temperature computed based on the detected 10.7 radiation. The animation above shows the evolution of the brightness temperature difference field, and the field is characterized by a lot of scattershot signal. Some river valley can be inferred in the signal, but the region of fog over southwestern Louisiana does not stand out.

In contrast, the GOES-R IFR Probability field, below, diagnoses an initially isolated region of enhanced IFR probabilities near stations that are reporting reduced ceilings/visibilities.  The enhanced IFR Probabilities develop over southwestern Louisiana and expand outward from there (a second region develops south of Houston TX).  Regions where IFR conditions do not develop have very low IFR Probabilities that persist with time.

GOESIFR_0445_1230anim_05Apr2016

GOES-R IFR Probability Fields, every 2 hours from 0445 to 1230 UTC, 5 April 2016 (Click to enlarge)

GOES-R Cloud Thickness relates 3.9 µm emissivity to cloud thickness based on historical relationships between that value and sodar observations off the west coast of the USA. The value is not computed during twilight conditions on either side of sunrise and sunset, but  the last observation taken before sunrise, shown below, is related to dissipation time according to this scatterplot. Cloud Thickness on the morning of 5 April 2016 was at most 830 feet, suggesting rapid dissipation after sunrise. This is what occurred.

GOESCLDTH_1145_05Apr2016

GOES-R Cloud Thickness, 1145 UTC on 5 April 2016 (Click to enlarge)

GOES-R IFR Probability screens out mid-level stratus

BTD_1215_29March2016

GOES-13 Brightness Temperature Difference (10.7 µm – 3.9 µm) at 1215 UTC on 29 March 2016 (Click to enlarge)

Consider the brightness temperature difference field above, from 1215 UTC on 29 March 2016. A strong signal that indicates water-based clouds extends from central Oklahoma to west Texas, and also from south Texas to west Texas. Are these water-based clouds obscuring visibilities at the surface? Over much of the region they are not. The Brightness Temperature Difference field gives information about the top of the cloud, but not the cloud base.

The GOES-R IFR Probability field for the same time, below, has screened out much of the region of mid-level stratus over Oklahoma and Texas. This can occur because IFR Probability fields include information from the Rapid Refresh Model. If that model does not indicated low-level saturation, IFR Probabilities will not be large. In the example shown, large values of IFR Probabilities are restricted to regions where IFR or near-IFR conditions are occurring.

IFRP_1215_29March2016

GOES-R IFR Probability Fields at 1215 UTC on 29 March 2016 (Click to enlarge)

Coastal Stratus/Fog along the East Coast

GOES_IFRP_0215_1400anim_17March2016

GOES-R IFR Probabilities, Hourly from 0215-1415 UTC 17 March 2016 (Click to enlarge)

GOES-R IFR Probabilities captured the development of coastal fog over coast of the Atlantic Ocean from Long Island south to North Carolina on the morning of March 17 2016 behind a weak cold front. IFR Conditions penetrated into the Delaware and Lehigh River Valleys over Pennsylvania. In general, GOES-R IFR Probability fields captured the region of IFR conditions as it developed and expanded. Note that the IFR Probabilities remained elevated through 1400 UTC along the eastern shore of Chesapeake Bay. The 1413 UTC webcam image from Tilghman Island, below, (source), shows an offshore fogbank.

TilghmanMDViewOverBay

Widespread IFR conditions over the Upper Midwest

GOESR_IFRP_0215_1215stepanim

GOES-R IFR Probability Fields, hourly from 0215-1215, 14 March 2016 (Click to enlarge)

A rain-dampened boundary layer allowed fog to form over much of the upper midwest on early Monday March 14 2016 (as a baggy low pressure system moved eastward). GOES-R IFR Probability fields, above, captured the slow expansion of the region of IFR conditions.  The character of the IFR Probability Field varies from smooth (over northern Illinois at 0215 UTC, for example (link)) to pixelated (over southern Minnesota at the same time).  This is related to whether the model only is used as a predictor (over northern Illinois) because of high clouds that prevent the satellite from viewing low clouds or whether model and satellite data are both used as predictors (over southern Minnesota).  The toggle below, of IFR Probability Fields and Brightness Temperature Difference fields at 1000 UTC on 14 March, underscores this relationship between IFR Probability and Brightness Temperature Difference fields.

In the animation above, IFR conditions are in general observed where IFR Probability fields suggest their presence.

GOES_IFR_BTD_14March2016_1000ztoggle

GOES-R IFR Probability and GOES-13 Brightness Temperature Difference Field, 1000 UTC on 14 March 2016 (Click to enlarge)

IFR Probability and Topography

GOES_IFR_15z_9March_Topo.toggle

It is not unusual for land to ascend into the clouds in regions where Moutains abut large valleys or gently sloping plains.  This is especially apparent in California on either side of the Central Valley.  The toggle above shows IFR Probability fields at 1645 UTC on 9 March, and a high-resolution topographic image.  High IFR Probabilities are well correlated with high terrain. This is something a user must consider when using the product.