Comparing GOES-R IFR Probability Fields and Webcam Observations over Alaska


GOES-R IFR Probability, 1345 UTC on 19 July 2016, and 1400 UTC Observations of surface Visibility. Blue arrows point to, from left to right, Wainwright, Barrow, and Deadhorse Alaska (Click to enlarge)

The FAA website that includes webcams over much of Alaska provides a great opportunity to match IFR Probability fields with surface observations. In the image above, Wainwright AK shows a 3-mile visibility, but Barrow and Deadhorse farther east report 10-mile visibilities. All three stations are very near high IFR Probabilities — the red/orange region has IFR Probabilities exceeding 90%. What to the webcams at the three stations show? Screen captures from the three webcams are below. Webcams from both Barrow and Deadhorse show non-IFR conditions. Wainwright shows a dense fog through which the sun is dimly visible. These webcam observations are in general agreement with the GOES-based IFR Probability fields: only Wainwright is in a region where IFR Probabilities are consistently high.


Wainwright AK webcams at 1413 UTC on 19 July 2016 (Click to enlarge)


Barrow AK webcams at 1413 UTC on 19 July 2016 (Click to enlarge)


Deadhorse AK webcams at 1413 UTC on 19 July 2016 (Click to enlarge)



GOES-R IFR Probabilities, 1645 UTC on 19 July 2016, and 1700 UTC Observations of Ceilings/Visibilities (Click to enlarge)

The image above is of GOES-R IFR Probabilities over the Aleutians, along with surface observations of ceilings and visibilities.  In general, IFR Probabilities are a bit higher on the northern side of the Aleutians (False Pass, Cold Bay, Nelson Lagoon, Port Heiden, Pilot Point) than the southern side (Perryville, King Cove).  That is confirmed with the screen captures of webcam imagery, seen below.  The webcam scenes are aligned from west to east, starting at False Pass AK and ending at Pilot Point.


Webcam imagery from False Pass AK (PAKF) at 1653 UTC, 19 July 2016 (Click to enlarge)


Webcam imagery from Cold Bay, AK (PACD) at 1654 UTC on 19 July 2016 (Click to enlarge)


Webcam imagery from King Cove, AK, (PAVC) at 1655 UTC on 19 July 2016 (Click to enlarge)


Webcam imagery from Nelson Lagoon, (PAOU) 1656 UTC on 19 July 2016 (Click to enlarge)


Webcam Imagery from Perryville, AK, 1659 UTC on 19 July 2016 (Click to enlarge)


Webcam imagery from Port Heiden, AK, (PAPH) 1657 UTC on 19 July 2016 (Click to enlarge)


Webcam Imagery from Pilot Point, AK, (PAPN) at 1658 UTC on 19 July 2016 (Click to enlarge)

Fog Detection at Very High Latitudes


GOES-R IFR Probability Fields, 0930-1230 UTC on 14 July 2016 (Click to enlarge)

The low Sun Angle at very high latitudes during Summer presents a challenge to the GOES-R IFR Probability algorithm. Solar backscatter from clouds in the visible and near-infrared (3.9 µm) channels make traditional fog detection methods (Brightness Temperature Difference between 10.7 µm and 3.9 µm) problematic, but they also are a challenge for GOES-R IFR Probability. At mid-latitudes, when 3.9 µm radiation is changing quickly because of rapid changes associated with a rising (or setting) sun, the ‘satellite’ portion of GOES-R IFR Probability will be frozen for a time step or two. In the GOES-R Algorithm, temporal data (in other words, data from previous times) are used starting at the beginning of the terminator transition periods (that is, when the sun is rising or setting) and it’s used until valid data, either day or night predictors, depending on whether it’s sunrise or sunset, is again available. At mid-latitudes (as seen in the example at the bottom of this post), the transition period is typically less than thirty minutes in length. At high latitudes, however, when the sun lingers near the horizon for hours, the use of temporal data can stretch out for several hours and products will not be updated during that time. This effect happens only in the months surrounding the Summer solstice.

Consider the animation above, of GOES-R IFR Probability north of Alaska, over the Arctic Ocean.  IFR Probabilty varies with each time step over much of the image, especially, for example, near the lower left corner (a region that includes the Brooks Range).  There is a significant region, though, where values do not change over the three hours of the animation — this is because of the low sun angle and the difficulty in relating the 3.9 µm emissivity to a fog property.   (You can detect at the end of the animation changes starting to impinge on the region of a static signal from the west) The three hour animation from 1100-1400 UTC, below, shows changes in the field over much of the domain over the Arctic Ocean.

Barrow Alaska (PABR) develops IFR conditions during this animation as low clouds move south from the ocean.  GOES-R IFR Probability fields suggest that the fog does not penetrate far inland.  Much of northern Alaska had record high temperatures on 13 July (It was 85 in Deadhorse, for example) with light south/southwest winds that would keep the ocean stratus offshore.  Webcam views at Barrow (from this link) confirm the presence of IFR conditions.


GOES-R IFR Probability fields, 1100-1400 UTC on 14 July 2016 (Click to enlarge)

When GOES-R is flying, data the 8.4 µm channel will be incorporated into the GOES-R IFR Probability algorithm. This channel (Band 11 on the GOES-R ABI) is not so adversely affected by reflected solar radiation as the 3.9 µm channel so the unchanging nature of GOES-R IFR probability fields will be mitigated.

This stationarity in the GOES-R IFR Probability is apparent at mid-latitudes as well. Careful inspection of the animation below, from 1300-1400 UTC on 14 July 2016, shows a region of static IFR Probability fields off the coast of northern Oregon between 1315 and 1345 UTC.


GOES-R IFR Probability, 1300-1400 UTC on 14 July 2016 (Click to enlarge)

MODIS and GOES IFR Probabilities over Alaska

GOES-R IFR Probabilities computed using GOES-15 pixels over Alaska suffer from problems inherent in any Geostationary Data Product at high latitudes: Pixel sizes are large. In addition, ‘limb brightening’ — that is, the shift in a brightness temperature towards cooler values because the path length of photon towards the satellite travels through more of the upper (colder) troposphere (a cooling that is also dependent on wavelength being sensed) — affects the brightness temperature difference product that is used to detect water-based clouds. MODIS data from Terra and Aqua has a much higher spatial resolution and a superior view angle. It’s fairly simple to use both MODIS data to get an idea of conditions in and around Alaska, and then use GOES data to approximate the temporal change. Terra and Aqua view Alaska frequently (link) — it’s uncommon to go more than 6 hours without a view.

The toggles below show a series of MODIS IFR Probabilities and corresponding GOES-15 IFR Probabilities from late on 8 July 2016 through mid-day on 9 July 2016. From 2100 UTC to 1400 UTC — 17 hours — there are 7 separate MODIS views of Alaska, and they show similar features. For example, the high terrain of the Brooks Range is apparent: larger values of IFR Probabilities are noted there. The same is true over south central Alaska, over the Alaska Range in between Anchorage and Fairbanks. Interpretation of IFR Probability fields over the State require this background knowledge of Topography.


MODIS-based and GOES-15-based GOES-R IFR Probability fields, 2100 UTC on 8 July 2016 (Click to enlarge)


MODIS-based and GOES-15-based GOES-R IFR Probability fields, 2245 UTC on 8 July 2016 (Click to enlarge)


MODIS-based and GOES-15-based GOES-R IFR Probability fields, 2300 UTC on 8 July 2016 (Click to enlarge)


MODIS-based and GOES-15-based GOES-R IFR Probability fields, 0645 UTC on 9 July 2016 (Click to enlarge)


MODIS-based and GOES-15-based GOES-R IFR Probability fields, 0815 UTC on 9 July 2016 (Click to enlarge)


MODIS-based and GOES-15-based GOES-R IFR Probability fields, 1230 UTC on 9 July 2016 (Click to enlarge)


MODIS-based and GOES-15-based GOES-R IFR Probability fields, 1400 UTC on 9 July 2016 (Click to enlarge)

The scenes above suggest that most IFR Conditions near Alaska are offshore during the early morning of 9 July. On 11 July 2016, some of those regions of reduced visibility crept onshore, as shown in the plot below from this site, where surface stations are color-coded by Flight Rules: Red and Magenta denote IFR and Low IFR conditions.


Surface METARS, 1700 UTC on 11 July 2016 (Click to enlarge)

GOES-15-based IFR Probability fields from near that time show high probabilities along the coastline of Alaska.  Note that the presence of IFR Conditions can also be deduced from this set of webcams! Consider, for example, this webcam site just west of Prudhoe Bay, in a region where GOES-based IFR Probabilities are high.


GOES-15-based GOES-R IFR probability fields, 1700 UTC on 11 July 2016, along with surface observations of ceilings and visibilities (Click to enlarge)

Dense Fog over Missouri and Illinois

Dense Fog Advisories were issued before sunrise on 5 July 2016 by the National Weather Service Offices in St Louis Missouri and Lincoln Illinois (link).  This post compares GOES-R IFR Probability fields with Brightness Temperature Difference Fields for that event.  At 0500 UTC, the Brightness Temperature Difference Field had a very strong signal over central/northern Illinois.  This region of mid-level stratus was de-emphasized by the GOES-R IFR Probability fields because of a lack of low-level saturation in the Rapid Refresh Model Fields.  At 0500 UTC, IFR Probability Fields show an increase in values over the lower Ohio River Valley.


GOES-13 Brightness Temperature Difference Field (3.9 µm – 10.7 µm) and GOES-R IFR Probability Fields, 0500 UTC on 5 July 2016 (Click to enlarge)

At 0700 UTC, below, an axis of higher probabilities has developed northwest to southeast across central Missouri;  in addition, IFR Probabilities are increasing slowly over central Illinois to the south of the mid-level stratus that persists over east-central parts of Illinois.  IFR Conditions are present at stations over central Illinois  (Springfield — KSPI — and Litchfield — K3LF, for example)


GOES-13 Brightness Temperature Difference Field (3.9 µm – 10.7 µm) and GOES-R IFR Probability Fields, 0700 UTC on 5 July 2016 (Click to enlarge)

By 1000 UTC, below, when Dense Fog Advisories have been issued, a strong brightness temperature difference signal is present over much of northern Missouri, but mid-level clouds over southeast Missouri prevent a strong signal from occurring there where fog is occurring.  IFR Probability Fields’ use of Rapid Refresh information mitigates this lack of satellite observations.    IFR Probability fields at 1000 UTC show a signal over much of southern Illinois where IFR Conditions are widespread.


GOES-13 Brightness Temperature Difference Field (3.9 µm – 10.7 µm) and GOES-R IFR Probability Fields, 1000 UTC on 5 July 2016 (Click to enlarge)

Use IFR Probability fields as a tool to become situationally aware to the development of lowered ceilings and reduced visibilities.  There are many times when Brightness Temperature Difference fields cannot tell the entire story — when multiple cloud layers exist, for example, or when mid-level stratus is present. A slow increase in GOES-R IFR Probability will often suggest lowering ceilings/reduced visibilities before Brightness Temperature Difference fields do.

Fog/Low Ceilings over Southwest Georgia


GOES-R IFR Probability fields on 27 June 2016 at 0200, 0400, and then hourly from 0700 through 1300 UTC (Click to enlarge)

Late-day thunderstorms on 26 June 2016 set the scene for the development of fog overnight over southwestern Georgia. The animation above shows the GOES-R IFR Probability fields.  An enhancement in the fields that is initially driven by Rapid Refresh Model data showing near-saturation at low levels is apparent at 0200 UTC.  As clouds associated with the departing convection dissipate, satellite data could also be used as input into the IFR Probability fields.  The toggle below of GOES-13 Brightness Temperature Difference fields (3.9 µm – 10.7 µm), at 0200 and 0400 UTC, shows the appearance of low-level clouds as mid-level and higher clouds (dark in the enhancement used) dissipate.  By 0400 UTC, when satellite pixels finally start to suggest low clouds, fog had already started to develop.  IFR Probability fields gave an early alert to the possibility of fog development on this day that was not possible from satellite data alone.


GOES-R Cloud Thickness Fields can give a hint to when radiation fog, as in this event, will dissipate in accordance with this scatterplot. The image below shows the GOES-R Cloud Thickness at 1030 UTC, the last field computed before twilight conditions (indeed, the boundary showing that boundary is readily apparent over eastern Georgia), with values exceeding 900 feet in some places over southwest Georgia.  Based on the scatterplot, that suggests a dissipation time of just over 2 hours (based on the best fit line, but note the scatter in dissipation times associated with cloud thicknesses of 900 feet:  just over an hour to almost 4 hours!) so clear skies would be expected by 1300 UTC.  The animation of visible imagery, here, shows that fog persisted just a bit longer than that, dissipating shortly after 1400 UTC.  GOES-R Cloud Thickness field is an empirical relationship between 3.9 µm emissivity and cloud thickness that is based on SODAR observations off the west coast.  The scatterplot was created based on past observations limited to the southeast part of the US and parts of the Great Plains.


GOES-13 Cloud Thickness, 1030 UTC on 27 June 2016 (Click to enlarge)

Low Ceilings and reduced visibilities over the Ohio Valley


Surface Observations at 1200 UTC on 24 June 2016 (Click to enlarge)

A screen capture from this site at 1215 UTC on 24 June 2016, above, shows IFR Conditions (Red) and Low IFR Conditions (Purple) over the upper Ohio River Valley and surrounding states.  The IFR Probability field for the same time, below, shows high probabilities in roughly the same regions that have IFR or Low IFR conditions.  The Brightness Temperature Difference field, also displayed in the toggle below, gives little information at this time of day.  A benefit of the GOES-R IFR Probability field is that it contains a coherent signal through sunrise.


GOES-R IFR Probability fields and GOES-13 Brightness Temperature Difference Fields (3.9 µm – 10.7 µm) at 1215 UTC on 24 June 2016 (Click to enlarge)

The toggle at 0915 UTC, below, before sunrise, shows a second benefit of IFR Probability fields: a useful signal in regions with cirrus clouds. High clouds, of course, prevent GOES-13 from viewing the development of fog/low stratus near the surface. The Rapid Refresh model data on low-level saturation that are part of the IFR Probability Field computations give quality information in regions of cirrus. In the example below, developing IFR conditions are depicted (the yellow enhancement that shows IFR Probabilities around 40%) over much of northern Kentucky and southern Ohio.  This is under a region of cirrus (black in the enhancement used for the brightness temperature difference) north of a convective system that sits over southeastern Kentucky and eastern Tennessee.


GOES-R IFR Probability fields and GOES-13 Brightness Temperature Difference Fields (3.9 µm – 10.7 µm) at 0915 UTC on 24 June 2016 (Click to enlarge)

The waning full moon provided ample illumination for the Suomi NPP Day/Night Band Imagery, shown below, from 0736 UTC on 24 June 2016.  The cirrus shield, mid-level clouds and developing valley fogs are all apparent.


Suomi NPP Day/Night band imagery, 0736 UTC on 24 June 2016 (Click to enlarge)

IFR Conditions over Maine


GOES-R IFR Probability, and surface plots of ceilings and visibility, 0500-1215 UTC on 12 June 2016 (Click to enlarge)

IFR Probability fields, above (a slower animation is here), show high probabilities of IFR Conditions over much of Maine, but a definite western edge is also present, moving eastward through New Hampshire and Vermont and reaching western Maine by 1215 UTC. The screen capture below, from this site, shows IFR (station models with red) and Low IFR Conditions (station models with magenta) over much of southern Maine at 1200 UTC on 12 June in advance of a warm front.

Careful inspection of the IFR Probability animation shows a field at 1000 UTC that is very speckled/pixelated. This likely results from cloud shadowing. The combination of a very low sun and multiple cloud layers resulted in many dark regions in the visible imagery that the cloud masking may have interpreted as clear regions. (Click here for a toggle between Visible Imagery and GOES-R IFR Probabilities at 1000 UTC).


Surface plot at 1200 UTC 12 June 2016. See text for details (Click to enlarge)

Low IFR Probability fields are also computed by the GOES-R Algorithms. Values are typically smaller than IFR Probability. Plots of Low IFR and IFR Probabilities at 0700 and 1215 UTC are shown below.


GOES-R Low IFR Probability and GOES-R IFR Probability, 0700 and 1215 UTC (Click to enlarge)

Power Outages

Two power outages 12 hours apart at UW-Madison CIMSS have impacted distribution of GOES-R IFR Probability fields. It’s possible that products may not be smoothly flowing again until Monday 13 June. Data are flowing as of about 0000 UTC on Saturday.

In the interim, users can find the products at the GEOCAT site. If you’re in the southeast US, near Atlanta, an experimental site that compares IFR Probability and Brightness Temperature Difference fields is here.

For over the Northern Plains


GOES-R IFR Probability Fields, computed from GOES-13 and Rapid Refresh output, 0215-1115 UTC on 1 June 2016 (Click to enlarge)

Dense Fog Advisories were issued over parts of Iowa and Minnesota early on 1 June 2016 (see map below). The fog developed over wet ground left in the wake of convection that moved through the region late in the day on 30 May/early on 1 June (Precipitation totals available here).  GOES-R IFR Probability fields, above, show the two areas of dense fog developing.  The region over Minnesota was characterized a lack of high clouds — the satellite could view the developing fog, and satellite parameters were included in the computation of IFR Probability.  Consequently, the IFR Probability values were larger.

Fog over Iowa initially developed under mid-level clouds behind departing convection. IFR Probability fields in that case show a flatter distribution because horizontal variability is controlled mostly by model fields that are smooth; additionally, IFR Probability values are somewhat reduced because satellite predictors cannot be used. By 0815 UTC, however, mid- and high-level clouds have dissipated, and the satellite has a unobstructed view of the fog/stratus. Satellite predictors could then be used and IFR Probabilities increased, and the field itself shows more horizontal variability as might be expected from the use of nominal 4-km resolution satellite pixels.

Screen Capture of website at 1129 UTC on 1 June. Dense Fog Advisories are indicated over eastern Iowa and northeast Minnesota (click to enlarge)

Fog over the Great Lakes


GOES-R IFR Probability fields computed with GOES-13 and Rapid Refresh Data, 1215 UTC on 26 May 2016 along with surface reports of Ceilings and visibilities (Click to enlarge)

High Dewpoint air (upper 50s and low- to mid-60s) has overrun the western Great Lakes, where water temperatures are closer to the mid 40s.  (Water Temperature from Buoy 45007 in southern Lake Michigan).  Advection fog is a result, and that fog can penetrate inland at night, or join up with fog that develops over night.  The image above shows the extent of low visibilities over the upper Midwest and the IFR Probability field early morning on the 26th of May. Lakes Michigan and Superior are diagnosed as socked in with fog. A similar field from 1945 UTC on 25 May similarly shows very high Probabilities over the cold Lakes. Expect high IFR Probabilities to persist over the western Great Lakes until the current weather pattern shifts.

Brightness Temperature Difference Fields can also show stratus over the Great Lakes, of course, but only if multiple cloud layers between the top of the stratus and the satellite do not exist. Convection over the upper Midwest overnight on 25-26 May frequently blocked the satellite’s view of the advection fog. The toggle below, from 0515 UTC on 26 May, shows how model data from the Rapid Refresh is able to supply guidance on IFR probability even in the absence of satellite information about low stratus over the Lakes.


GOES-13 Brightness Temperature Difference Fields and GOES-R IFR Probability fields, 0515 UTC on 26 May 2016 (Click to enlarge)