Author Archives: Scott Lindstrom

Identifying regions of fog under cirrus

GOES-R IFR Probabilities computed from GOES-13 and GOES-East Brightness Temperature Differences (10.7 µm – 3.9 µm) at 0500 UTC on 5 March 2014 (click to enlarge)

Fog developed overnight over Tennessee on March 5th, but Cirrus clouds prevented the traditional brightness temperature difference product from observing low-level water-based clouds. It is for events like this that the IFR Probability fields (that incorporate surface-based information by way of the Rapid Refresh Model) is important. The IFR Probability fields use predictors from the Rapid Refresh model to showcase where low ceilings and reduced visibilities are most likely. Satellite predictors are unavailable where cirrus clouds are present and the probability field shows lower values. So when you see low values, make sure you understand why the values are low: is it because cirrus clouds are present?

A side-by-side animation of the IFR Probabilities and the Brightness Temperature Difference Field is presented below. The effects of cirrus on the Probability field is obvious.

GOES-R IFR Probabilities computed from GOES-East (left) and GOES-East Brightness Temperature Differences (10.7 µm – 3.9 µm) (right), hourly from 0500 UTC through 1515 UTC on 5 March 2014 (click to enlarge)

Fog over northeast Florida and coastal Georgia and South Carolina

GOES-R IFR Probabilities computed from GOES-13 (Upper left); GOES-East Brightness Temperature Differences (10.7 µm – 3.9 µm) (Upper Right); MODIS-based IFR Probabilities or VIIRS-based Brightness Temperature Difference (11.35 µm – 3.74 µm) (Lower Left); GOES-R Cloud Thickness computed from GOES-East (Lower Right) (click to play animation)

Cold air has swept down the east coast into northern Florida, and the leading edge of that cold air, marked by a shift to northeasterly winds and low clouds, shows up well in the GOES-R IFR Probability fields, displayed above, because the airmass with the northeasterly winds also included low clouds/fog. Note in the animation how IFR conditions develop in Jacksonville as the higher IFR probabilities slide southward. Similarly, IFR conditions diminish over Savannah as IFR Probabilities drop.

This is a case for which the heritage method of detecting fog had difficulties because multiple cloud layers existed. For example, a stratus deck over central Florida shows up very well in the brightness temperature difference field from both GOES and VIIRS, but IFR conditions are not initially seen there (and GOES-R IFR Probabilities are small). The GOES-R Cloud Thickness is not computed in regions with multiple cloud layers, typically, because it shows the thickness of the highest water-based cloud layer. If any overlaying cloud layer at high levels contains ice, the field is not computed.

Sea fog over the western Gulf of Mexico

The National Weather Service in Houston/Galveston has issued Dense Fog advisories for Sea Fog in the easterly flow south of a cold front draped across the northern Gulf:

MARINE WEATHER STATEMENT
NATIONAL WEATHER SERVICE HOUSTON/GALVESTON TX
602 PM CST SUN FEB 23 2014

GMZ330-335-350-355-370-375-260000-
MATAGORDA BAY-GALVESTON BAY-
WATERS FROM FREEPORT TO THE MATAGORDA SHIP CHANNEL OUT 20 NM-
WATERS FROM HIGH ISLAND TO FREEPORT OUT 20 NM-
WATERS FROM FREEPORT TO THE MATAGORDA SHIP CHANNEL 20 NM TO 60 NM-
WATERS FROM HIGH ISLAND TO FREEPORT 20 TO 60 NM-
602 PM CST SUN FEB 23 2014

…DENSE SEA FOG POSSIBLE ACROSS THE AREA FOR THE NEXT SEVERAL DAYS…

AREAS OF SEA FOG…SOME DENSE WITH VISIBILITIES OF 1 NM OR LESS…WILL
CONTINUE TO BE POSSIBLE IN AND AROUND THE GALVESTON AND MATAGORDA BAY
AREAS ALONG WITH THE UPPER TEXAS COASTAL WATERS OUT TO APPROXIMATELY
20 NM. DENSE FOG ADVISORIES MIGHT BE NEEDED.

LITTLE CHANGE IN THIS PATTERN IS EXPECTED UNTIL THE PASSAGE OF THE NEXT
COLD FRONT SOME TIME AROUND LATE TUESDAY NIGHT OR EARLY WEDNESDAY
MORNING.

MARINERS SHOULD BE PREPARED FOR SUDDEN CHANGES IN VISIBILITY OVER SHORT
DISTANCES. REDUCE YOUR SPEED AND KEEP A LOOKOUT FOR OTHER VESSELS…BUOYS
AND BREAKWATERS. KEEP YOUR NAVIGATION LIGHTS ON. INEXPERIENCED MARINERS…
ESPECIALLY THOSE OPERATING SMALLER CRAFT OR NOT EQUIPPED WITH RADAR…SHOULD
CONSIDER SEEKING SAFE HARBOR.

How does the GOES-R IFR Probability field handle this event?

GOES-Based GOES-R IFR Probabilities (Upper Left), GOES-East Brightness Temperature Difference Fields (10.7 µm – 3.9 µm) (Upper Right), Suomi/NPP Day/Night band and MODIS-based IFR Probability fields (Lower Left), GOES-East Water Vapor Imagery (6.7 µm)(Lower Right), hourly from 0400 UTC through 1600 UTC 14 February 2014 (click image to enlarge)

IFR Probabilities are correctly limited to coastal regions of east Texas, with high values off shore. The brightness temperature difference field has difficulty identifying regions of low clouds over the Gulf of Mexico because of southwesterly flow aloft that contains mid- and high-level cloudiness. The relatively flat field over the Gulf — large values, but little variability — correspond to regions where high clouds exist. These high clouds prevent satellite predictors from being used in the IFR Probability algorithm because the brightness temperature difference does not observe low clouds, so only the Rapid Refresh model output is used to compute the IFR Probability. Therefore the IFR Probability fields are a bit flatter. Where there are breaks in the high clouds, the brightness temperature difference field can be used in the IFR Probability algorithm, and the computed IFR Probability is larger. In addition, the character of the probability field is more pixelated like a satellite image.

The bottom left image in the 4-panel composite above includes both the Day/Night band from Suomi/NPP (an image that — because of scant lunar illumination — gives little distinct information about the clouds present) and a MODIS-based IFR Probability field. For selected still imagery of ~0830 UTC Suomi/NPP click here; click here for ~0730 UTC MODIS-based IFR probability.

IFR Probability Fields are an early-alert for Developing Fog

GOES-R IFR Probabilities from 0400 through 1400 UTC on 14 February 2014 (click image to enlarge)

Fog and Low clouds resulted in IFR conditions along a long swath of the western Gulf Coast today. IFR Probability fields warned of the development of these conditions long before a strong signal appeared in the traditional brightness temperature difference fields. The animation above, of hourly GOES-R IFR Probability fields (and surface observations of ceiling and visibility). There are indications by 0615 and 0702 that fog/low stratus is developing, and those indications are matched by some observations of IFR conditions. By 0915 UTC, widespread IFR conditions are present from southwest Louisiana southwestward through coastal Texas.

GOES-R IFR Probabilities, MODIS IFR Probabilities, and MODIS Brightness Temperature Difference fields, all from ~0845 UTC 14 February 2014 (click image to enlarge)

MODIS data can be used to generate IFR Probabilities as well, as shown above. The MODIS-based and GOES-based fields both generally overlap regions with developing IFR conditions. The MODIS-based Brightness temperature difference product (called MODIS FOG in the image annotation) shows little signal in central Louisiana/east Texas (near Lufkin, for example) or southwest of Houston, two places where near-IFR conditions are developing (and where the IFR Probability fields have a signal).

GOES-R IFR Probabilities, MODIS IFR Probabilities, and MODIS Brightness Temperature Difference fields, all from ~0845 UTC 14 February 2014 (click image to enlarge)

Suomi/NPP data (Brightness Temperature Difference fields, and the Day/Night band) from the same hour (0822 UTC) as the MODIS data similarly underpredicts the areal extend of the developing IFR conditions.

GOES-Based GOES-R IFR Probabilities (Upper Left), GOES-East Brightness Temperature Difference Fields (Upper Right), GOES-R Cloud Thickness (Lower Left), GOES-East Visible Imagery (Lower Right), hourly from 0400 UTC through 1600 UTC 14 February 2014 (click image to enlarge)

The animation above shows hourly views of GOES-R IFR Probability and GOES-East Brightness Temperature Difference fields. There is little discernible signal in the brightness temperature difference field until about 0915 UTC (several hours after the IFR Probability field has been suggesting fog development). Thus the GOES-R IFR Probability field is giving better lead time is diagnosing where visibility restrictions might occur/be occurring. In addition, the GOES-R Cloud Thickness product shows that the thickest fog/stratus field just before sunrise is just east of Austin/San Antonio, and that is the last region to clear out after sunrise.

The presence of high clouds has an effect on both the IFR Probability fields and the brightness temperature difference field. When high clouds are present (in the brightness temperature difference enhancement used, high clouds are dark), IFR Probabilities drop in value and the field becomes flatter because satellite data cannot be used in the computation of the IFR Probability Field.

IFR Conditions over the southern Plains

GOES-R IFR Probabilities from 0100 through 1500 UTC on 10 February 2014 (click image to enlarge)

Cold air dropping southward through the southern Plains is sometimes shunted westward towards higher elevation on the Equatorward side of the Polar High that anchors the cold air. That upslope flow facilitates the development of fog and low stratus. That was the case today, and the southward and westward movement of low clouds/fog is obvious in the IFR Probability field animation shown above. Much of the High Plains south of Kansas had reduced visibility and lowered ceilings, and IFR Probabilities were high.

The IFR Probability field does a better job of outlining where the low clouds associated with IFR Conditions are present. Compare the half-hourly loop above to the hourly loop of the Brightness Temperature Difference (10.7 µm – 3.9 µm), below. Regions with multiple cloud layers show little signal in the brightness temperature field, and the flip in signal at sunrise — as 3.9 µm radiation from the Sun is reflected off the clouds, overwhelming the emitted signal — is obvious.

GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm) 0300 through 1500 UTC on 10 February 2014 (click image to enlarge)

Polar-orbiting satellites can give high-resolution views of scenes. Suomi/NPP carries the VIIRS instrument, which has a Day/Night band and 11.35 and 3.74 µm channels, shown below in a toggle. Of course, these views are telling you something about the top of the clouds only. Whether of not visibility/ceiling restrictions are happening is unknown. GOES-R IFR Probability algorithms have not yet been configured for Suomi/NPP data (such a configuration is complicated by the lack of a water vapor channel on VIIRS). The characteristic signal of high clouds — dark features in the enhancement used — shows up in the VIIRS Brightness Temperature Difference field.

Suomi/NPP VIIRS Day/Night band and Brightness Temperature Difference (11.35 µm – 3.74 µm) at 0756 UTC on 10 February 2014 (click image to enlarge)

MODIS data includes water vapor imagery, and thus GOES-R IFR Probability fields can be computed using MODIS data, as shown below from 0907 UTC. This toggle includes MODIS-based Brightness Temperature Differences, MODIS-based GOES-R IFR Probabilities, GOES-based GOES-R IFR Probabilities and GOES Brightness Temperature Differences. The shortcoming of the MODIS-based data is obvious (it doesn’t view the entire scene), but its strength (excellent spatial resolution) is also apparent.

MODIS Brightness Temperature Difference (11 µm – 3.9 µm), MODIS-based GOES-R IFR Probabilities, GOES-based GOES-R IFR Probabilities and GOES-East Brightness Temperature Differences (10.7 µm – 3.9 µm) at ~0910 UTC on 10 February 2014 (click image to enlarge)

GOES-R IFR Probability signal because of co-registration errors

GOES-R IFR Probabilities at 0802 UTC, 30 January 2014 (click image to enlarge)

Special Update, 17 November 2014.

GOES-R IFR Probabilities on the morning of 30 January suggested the likelihood of fog along some of the Finger Lakes of upstate New York. These high probabilities arise because the Brightness Temperature Difference (10.7 µm – 3.9 µm) Product, below, shows a signal there. Note, however, that the Brightness Temperature Difference has a shadow; this is the sign that the co-registration error that is present between the 10.7 µm and 3.9 µm channels is producing a fictitious signal of fog over the lake. Such errors have been discussed here and elsewhere in the past.

GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm) at 0801 UTC, 30 January 2014 (click image to enlarge)

Evidence that fog is not present is available in Suomi/NPP data taken at the same time as the GOES data, above. The toggle, below, of Day/Night Band imagery and of the brightness temperature difference (11.35 µm – 3.74 µm) from VIIRS shows scant evidence of fog/low stratus near the Finger Lakes. Because the moon is new, lunar illumination is at a minimum and surface features in the Day/Night band are not distinct, but the dark waters of the lakes are apparent.

Suomi/NPP VIIRS Day/Night band and Brightness Temperature Difference (11.35 µm – 3.74 µm) at 0802 UTC, 30 January 2014 (click image to enlarge)

MODIS data also suggests no fog/low stratus in the region. Both the brightness temperature difference field and the MODIS-based IFR Probabilities, below, support a forecast that does not mention fog around the Finger Lakes.

MODIS Brightness Temperature Difference (11 µm – 3.74 µm) and MODIS-based GOES-R IFR Probabilities at 0746 UTC, 30 January 2014 (click image to enlarge)

=====================================================================

Update, 17 November 2014

NOAA/NESDIS has tested a software fix to align better the longwave infrared (10.7 µm) and shortwave infrared (3.9 µm) channels. The toggle below is of the Brightness Temperature Difference Field with (After co-registration correction) and without (Prior to co-registration correction) the realignment.

GOES-13 Brightness Temperature Difference Fields at 0802 UTC, 30 January 2014, with and without the co-registration correction as indicated (Photo Credit: UW-Madison CIMSS; Click to enlarge)

The correction of the co-registration error translates into more realistic IFR Probabilities in/around the Finger Lakes. In this case, IFR Probabilities are reduced because the false strong signal from the satellite is not present because of more accurate co-registration.

GOES-R IFR Probability fields computed Prior to and After co-registration correction, data from 0802 UTC 30 January 2014. IFR Probability fields with the corrected co-registration data are more accurate. (Photo Credit: UW-Madison CIMSS; Click to enlarge)

One more example of extreme cold

Toggle of GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm) and GOES-R IFR Probabilities at 1100 UTC, 27 January 2014 (click image to enlarge)

One more example, above, showing the effects of extreme cold on the IFR Probability. IFR Probabilities correctly ignore the regions of low stratus in advance of the extreme cold air over Kansas and over the Ohio River Valley and Great Lakes. However, because of how the pseudo-emissivity is computed (See here also), and because the Rapid Refresh model show saturation in lower levels, regions with extreme cold will show a pixelated signal with noise.

Fog in Idaho, Oregon and Washington over three days

Suomi/NPP Day/Night band imagery over the Pacific Northwest, 0912 and 1053 UTC on 21 January 2014(click image to enlarge)

Suomi/NPP viewed eastern Oregon/Washington and western Idaho on two successive scans overnight. The 3/4 full moon provides ample illumination, and fog/low stratus is apparent in the imagery above. A view of the top of the clouds, however, gives little information about the cloud base, that is, whether or not important restrictions in visibility are occurring. For something like that, it is helpful to include surface-based data. Rapid Refresh data are fused with the model data to highlight regions where IFR conditions are most likely. The image below is a toggle of the 1053 UTC Day/Night band image and the 1100 UTC GOES-R IFR Probabilities (computed using GOES-West data). GOES-R IFR Probabilities are correctly highlighting regions where ceilings and visibilities are consistent with IFR conditions. Where the Day/Night band is possibly seeing elevated stratus (between The Dalles (KDLS) and Yakima (KYKM), for example), IFR Probabilities are lower.

GOES-based data cannot resolve very small-scale fog events in river valleys (over northeastern Washington State, for example). The superior spatial resolution of a polar-orbiting satellite like Suomi/NPP (or Terra/Aqua) can really help fine-tune understanding of the horizontal distribution of low clouds.

Suomi/NPP Day/Night band imagery and GOES-R IFR Probabilities, ~1100 UTC on 21 January 2014(click image to enlarge)

Added: 22 January 2014

Suomi/NPP Day/Night band imagery over the Pacific Northwest, 0854 and 1034 UTC on 22 January 2014(click image to enlarge)

Suomi/NPP Brightness Temperature Difference from VIIRS, 11 µm – 3.74 µm imagery over the Pacific Northwest, 0854 and 1034 UTC on 22 January 2014(click image to enlarge)

The stagnant weather pattern under the west coast ridge allowed fog to persist overnight on January 22nd, and once again, the Day/Night band observed the fog-filled Snake River Valley of southern Idaho. The newly-rising moon at 0854 UTC provided less illumination than the higher moon at 1034 UTC, but both show fog/low stratus over the Snake River Valley of Idaho, and over parts of northern Oregon and central Washington. It is difficult to tell where the stratus is close enough to the ground to produce IFR conditions, however. The brightness temperature difference product from VIIRS, above, can distinguish between low clouds (orange enhancement) and higher clouds (dark grey) because of the different emissivity properties of water-based low clouds and ice-based higher clouds.

The toggle below shows how the higher-resolution VIIRS instrument can more accurately portray sharp edges to low clouds. Both instruments show the region of high clouds moving onshore in coastal Oregon (at the very very edge of the Suomi/NPP scan). These high clouds make satellite-detection of low clouds difficult because they mask detection of lower clouds.

Suomi/NPP Brightness Temperature Difference from VIIRS (10.35 µm – 3.74 µm) and GOES-15 Imager Brightness Temperature Difference (10.7 µm – 3.9 µm imagery over the Pacific Northwest, ~0900 UTC on 22 January 2014(click image to enlarge)

GOES IFR Probabilities at 0900 UTC and at 1030 UTC (click image to enlarge)

GOES-based IFR Probabilities show the probability of fog and low ceilings (IFR conditions) even where high clouds are present. In the toggle above, note the regions where the IFR Probability field is uniform (off the coast of Oregon, yellow, and over west-central Washington State, orange and yellow, both at 0900 UTC). These smooth fields are typical of IFR Probabilities that are determined primarily from Rapid Refresh data. Where those smooth fields exist, satellite data does not give a signal of low clouds — usually because of the presence of ice-based clouds at higher levels; therefore, model data are driving the IFR Probability signal, and model data are typically smoother than the more pixelated satellite field. There are places, however, where model data alone does not accurately portray IFR conditions (at KGPI, for example (Glacier Park), where high clouds are present).

IFR Probability algorithms have not yet been extended using data from Suomi/NPP, in large part because the VIIRS instrument does not detect radiation in the so-called water-vapor channel (around 6.7 µm). The MODIS detector on board Terra and Aqua does have a water vapor channel, and IFR Probabilities are routinely produced from MODIS data, as shown below. MODIS, like VIIRS, has a 1-km pixel footprint that excels at detecting very fine small-scale features in clouds, especially small valleys, that are smeared out in the GOES imagery. The toggle below is of MODIS Brightness Temperature Difference, MODIS-based IFR Probabilities, GOES Brightness Temperature Difference, and GOES-based IFR Probabilities, all at ~1015 UTC on 22 January. Two things to note: MODIS has cleaner edges to fields, related to the high spatial resolution. The GOES-based brightness temperature difference highlights many more pixels over central Oregon where fog is not present. These positive hits bleed into the GOES-based IFR Probabilities, and they occur because of emissivity differences in very dry soils (See for example, this post). As drought conditions persist and intensify on the west coast under the longwave ridge, expect this signal to persist. The signals are not apparent in MODIS or VIIRS brightness temperature differences because of the narrower spectrum of those observations.

MODIS Brightness Temperature Difference (11 µm – 3.74 µm), MODIS-based GOES-R IFR Probabilities, GOES-15 Imager Brightness Temperature Difference (10.7 µm – 3.9 µm), GOES-based GOES-R IFR Probabilities and MODIS-based IFR Probabilities (again), all near 1015 UTC 22 January 2014(click image to enlarge)

Added, 23 January:

Fog persists in the Snake River Valley and elsewhere. It has also become more widespread over the high plains of Montana. Note the difference in the Day/Night band imagery below. At 0834 UTC, the rising quarter moon is unable to provide a lot of illumination; by 1015 UTC, however, the moon is illuminating the large areas of fog. Because the moon is waning, however, Day/Night band imagery will become less useful in the next week. A toggle between the 1015 UTC Day/Night band and the GOES-R IFR Probabilities computed using GOES-West (below the Day/Night band imagery) continues to demonstrate how well the field outlines the region of IFR conditions.

Suomi/NPP Day/Night band imagery over the Pacific Northwest, 0834 and 1015 UTC on 23 January 2014(click image to enlarge)

Suomi/NPP Day/Night band imagery and GOES-R IFR Probabilities (from GOES-15 and Rapid Refresh data) over the Pacific Northwest, 1015 UTC on 23 January 2014(click image to enlarge)

IFR Probabilities can identify frontal zones

Hourly images of GOES-East IFR Probabilities and surface plots of visibilities/ceilings as well as 3-hourly HPC depictions of frontal positions, 0701 – 1515 UTC on 21 January 2014(click image to enlarge)

Frontal zones are often accompanies by lower visibilities, and the image above shows how IFR Probabilities march steadily eastward as a front marches eastward across the southern Gulf states. Thus, the IFR Probabilities can, at times, serve as a proxy for a frontal zone.

Dense fog on the East Coast

GOES-East IFR Probabilities and surface plots of visibilities/ceilings at 0615 UTC 15 January (Upper Left), GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm), 0615 UTC 15 January (Upper Right), GOES-R Cloud Thickness, 0615 UTC 15 January (Lower Left), and Suomi/NPP Day/Night Band and Brightness Temperature Difference toggle (11.35 µm – 3.74 µm), 0605 UTC 15 January (Lower Right)(click image to enlarge)

The image above documents the GOES-R IFR Probability field during a fog event over the East Coast. Note how the IFR Probability field shows more horizontal uniformity than the traditional brightness temperature difference field over eastern Pennsylvania (where IFR conditions are reported). For example, both Selinsgrove along the Susquehanna and Reading in south-central Pennsylvania report IFR conditions in regions where the IFR Probability field has a strong return, but where the brightness temperature difference field’s diagnosis is less certain.

The Suomi/NPP field demonstrates the importance of higher resolution from polar orbiting satellites. Both the Day/Night Band and the brightness temperature difference fields suggest the presence of river valley fog over the West Branch of the Susquehanna and its many tributaries in central Pennsylvania. This continues at 0743 UTC, below, when Suomi/NPP’s subsequent overpass also viewed the Susquehanna valley. At both times, the river fog is too small-scale to be detected with GOES-13’s nominal 4-km pixel size.

GOES-East IFR Probabilities and surface plots of visibilities/ceilings at 0745 UTC 15 January (Upper Left), GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm), 0745 UTC 15 January (Upper Right), GOES-R Cloud Thickness, 0745 UTC 15 January (Lower Left), and Suomi/NPP Day/Night Band and Brightness Temperature Difference toggle (11.35 µm – 3.74 µm), 0743 UTC 15 January (Lower Right)(click image to enlarge)

The animation of the fields, below, done to demonstrate the importance of GOES-13’s temporal resolution, shows how the GOES-R IFR Probability field accurately captures the extent of the fog, even as the sun rises and causes the sign of the brightness temperature difference to flip. The traditional brightness temperature difference field has difficulty both in maintaining a signal through sunrise, and it diagnosing the region of fog/low stratus over northcentral Pennsylvania in and around the Poconos and in the Susquehanna River valley. The IFR Probability field has a minimum over/around Mt. Pocono, where IFR conditions are not observed until close to sunrise. IFR probabilities are small over Altoona, where the brightness temperature difference field shows a strong signal developing late at night (and where observations suggest an elevated stratus deck). In this region, although the satellite suggests fog might be present, model conditions do not agree, and IFR Probabilities are correctly minimized.

GOES-R Cloud thickness suggests that the thickest blanket of fog is over New Jersey. This diagnosis continues up through the twilight conditions of sunrise, at which point Cloud thicknesses are no longer diagnosed.

GOES-East IFR Probabilities and surface plots of visibilities/ceilings (Upper Left), GOES-East Brightness Temperature Difference (10.7 µm – 3.9 µm) (Upper Right), GOES-R Cloud Thickness (Lower Left), and GOES-East Water Vapor (6.7 µm), all times as indicated (Lower Right)(click image to enlarge)

GOES-R Fog Product Examples

Fog detection fusing GOES, Terra/Aqua or Suomi/NPP Satellite data with Model output