An even stronger cold front and even more blowing dust – 18 March 2014

In a recent post we showed a dust storm that moved through southeast Colorado and into the Texas Panhandle on 11 March with a strong cold front.  Another cold front, even a bit stronger than the one last week, pushed southward through the same area almost exactly one week later as a deepening surface low emerged from the Rockies into the South-Central Plains (see Figure 1 below).  Here we will again take a look at the CIRA Proving Ground Pink and Yellow Dust Products from MODIS.

Figure 1. Visible image combined with radar reflectivity at 1745 UTC on 18 March, overlaid with 1800 UTC METAR observations and a pressure (MSLP) analysis.

A closer look at the area of interest is shown in Figures 2 and 3, where we also get a feel for how well the dust showed up in this case with GOES visible satellite imagery.

Figure 2. Visible satellite image at 1630 UTC with the 1700 UTC METAR plot.

In the image about the cold front has moved into the Texas Panhandle with a distinct wind shift.  Stronger wind gusts exceeding 50 mph are found just to the north in southeastern Colorado, where temperatures have dropped into the 40s.  Even colder air lies just to the north, with Limon (at the northern edge of the image) reporting heavy snow.  One area of blowing dust is in southeastern Colorado with the stronger winds, although it is not obviously dust from the image above.  Another area of dust is seen blowing from west to east ahead of the cold front in the warmer air and dry westerly flow, and this dust is more apparent in the image, as it is in the next image (below) from 30 minutes later.

Figure 3. Visible satellite image at 1700 UTC with the 1700 UTC METAR plot.

The next visible image (Figure 4) is from one hour later, and at this time we also have a MODIS pass so we can compare how the dust looks with the two CIRA dust products.

Figure 4. Visible satellite image at 1800 UTC with the 1800 UTC METAR plot.

Here are the two dust products for near this time; the “pink” dust product (Figure 5) and the “yellow” dust product (Figure 6).

Figure 5. CIRA Proving Ground Pink Dust Product at 1810 UTC.

Figure 6. CIRA Proving Ground Yellow Dust Product at 1810 UTC.

In both CIRA products the dust is clearly distinguished from the background clouds (which are bluish in the Yellow Dust Product and more cyan colored in the Pink Dust Product).  It is especially easy to see the large amount of dust in far southeastern Colorado into the Oklahoma Panhandle with the stronger winds behind the cold front. This particular area of dust continued to consolidate during the afternoon as it pushed southward.  A lot of wave structure is seen in the dust, given the horizontal resolution of 1 km in the MODIS-based imagery.  Figure 7 shows the visible image at 1900 UTC again with a METAR plot at the same time.  The image in Figure 8 is the same for 2000 UTC.  We can compare these images with the CIRA dust products from the next MODIS pass at 1948 UTC, in Figures 9 and 10.

Figure 7. Visible satellite image at 1900 UTC with the 1900 UTC METAR plot.

Figure 8. Visible satellite image at 2000 UTC with the 2000 UTC METAR plot.

Figure 9. CIRA Proving Ground Pink Dust Product at 1948 UTC.

Figure 10. CIRA Proving Ground Yellow Dust Product at 1948 UTC.

The CIRA Dust product images again reveal two areas of blowing dust, a dense area just moving into the Texas Panhandle associated with the stronger northerly winds and sharp cooling behind the cold front.  The southern area of dust is still in the same general area  as before (near Lubbock, TX (LBB), similar to what occurred a week earlier when they experienced a pre-frontal dust event before a second one behind the cold front).  There likely is more blowing dust in southeastern Colorado than is shown in the imagery, but it is obscured by low cloudiness in the colder air well behind the front.

A more recent addition to the Polar orbiting satellites is the Suomi/NPP satellite with the VIIRS instrumentation, which also has the channels needed to make an image similar to the one shown in Figure 10.  The pass on this day was at 1946 UTC, and the CIRA pink dust image is shown in Figure 11 below.

Figure 11. CIRA pink dust image at 1946 UTC from the VIIRS instrument aboard the Suomi/NPP satellite.

The area of dense dust well behind the cold front continued to expand as it moved southward across the Texas Panhandle, becoming a larger scale Haboob that spread through Amarillo (AMA) and LBB, again almost exactly a week after the previous event. Some METAR observations are shown below (Figure 12, from La Junta (LHX) and Figure 13, from Lamar (LAA), both in southeastern Colorado and equally far south, from AMA in Figure 14 (located in the central Texas Panhandle) and then farther south from LBB, in Figure 15).

Figure 12. METAR observations from La Junta (LHX), Colorado from 0853 UTC (bottom) to 2003 UTC (top) on 18 March.

Figure 13. METAR observations from Lamar (LAA), Colorado from 0853 UTC (bottom) to 2003 UTC (top) on 18 March.

Figure 14. METAR observations from Amarillo (AMA), Texas from 1153 UTC/18 March (bottom) to 1153 UTC/19 March (top).

Figure 15. METAR observations from Lubbock (LBB), Texas from 1253 UTC/18 March (bottom) to 0853 UTC/19 March (top).

The various NWS Weather Forecast Offices (WFOs) issued numerous statements, warnings and other graphical products to convey information about the blowing dust, which can be very hazardous to travel when visibilities are suddenly reduced.  Below we show a sample of these from north to south.  First is a look at the “Weather Story” from the Boulder (BOU) WFO issued on the morning of 18 March, in Figure 16, then moving farther south the same product from the Pueblo (PUB) Colorado WFO in Figure 17.

Figure 16. BOU WFO Weather Story graphic conveying information on the blowing dust. Note the snow also in the forecast area.

Figure 17. The Pueblo WFO Weather Story graphic shows the area where both a Dust Storm and a High Wind Warning had been issued.

Social media is also used to help convey warnings and other information from a WFO.  In Figure 18 is an example of such a product from the AMA WFO, highlighting the area of dust near LBB and the “Wall of dust” approaching from the north.  An image posted by NWS personnel from the WFO, along with another image as the dust entered a neighborhood, both from the WFO Facebook page, are shown in Figure 19.

Figure 18. A graphic image posted by WFO AMA on their Facebook page showing the dust areas.

Figure 19. Photos of the dust on the AMA WFO Facebook page from the late afternoon on 18 March.

The Weather Story issued by the LBB WFO late in the afternoon on 18 March also nicely describes in graphical form the two areas of dust (Figure 20).

Figure 20. Weather Story graphic from the LBB WFO.

A picture posted on their Facebook page shows the dusty scene in the afternoon in the area of dust that persisted ahead of the cold front (Figure 21).

Figure 21. Dusty scene in Lubbock, Texas in the dust area ahead of the cold front during the afternoon of 18 March, posted on the WFO LBB Facebook page.

Posted in Blowing Dust (Blue-light absorption technique) | Leave a comment

More blowing dust with strong cold front on 11 March 2014

A strong cold front moved south across the High Plains on Tuesday afternoon and evening, bringing an episode of blowing dust with it.  Here we look at what happened and how it appeared with some GOES-R Proving Ground products that highlight blowing dust.  The sequence of surface maps displayed below in Figure 1 show the southward plunge across the plains east of the Front Range, with very strong northerly winds pushing the cold air rapidly southward.  It appears that the strongest northerly winds (20 to 30 knots sustained with gusts over 40 knots) occur behind the initial frontal surge.

Figure 1. Sequence of surface plots showing the push of cold air southward on 11 March.

One of the problems that operational meteorologists have when it comes to blowing dust is that it can be hard to see in conventional visible satellite imagery.  We demonstrate this with the GOES visible image below taken from AWIPS.

Figure 2. Visible satellite image at 1745 UTC with METAR observations.

A number of stations report visibility restrictions due to haze (in southeastern Colorado) or haze and blowing dust (for example, KGCK in southwestern Kansas).  But it is difficult to see this dust in the image.  The CIRA MODIS dust products use the 11 and 12 micrometer thermal IR bands form the Moderate Resolution Imaging Spectroradiometer (MODIS) instruments which fly on the polar-orbiting NASA Terra and Aqua satellites to highlight dust in either yellow or pink.  Because the satellites are Polar orbiters we only get one pass each during the daytime over the area of interest.  An example of each product is shown below for the pass closest to the time of the image shown in Figure 2.

Figure 3. CIRA yellow dust product for 1805 UTC, with METARs.

Figure 4. Pink dust product for 1805 UTC on 11 March.

Both dust products clearly highlight the most intense areas of blowing dust, which are found in southeastern Colorado at 1805 UTC.  A similar comparison is made below for the time of the next pass at 1945 UTC, in Figures 5-7 below (note that for this time we have a visible image at the same time).

Figure 5. GOES visible image from AWIPS at 1945 UTC with 2000 UTC METARs.

Figure 6. CIRA yellow dust product for 1945 UTC with 2000 UTC METARs.

Figure 7. CIRA pink dust product for 1945 UTC.

Blowing dust can be a significant hazard to all modes of travel, and in an earlier blog we documented traffic accidents that occurred during a dust storm earlier this winter in eastern Colorado.  The National Weather Service (NWS) Weather Forecast Offices (WFOs) at Boulder and Pueblo issued Dust Advisories for this event (shown below in Figures 8 and 9) and included discussion of the dust in their “Weather Stories” (Figures 10 and 11).

Figure 8. Boulder WFO forecast graphic page for Tuesday afternoon showing the location of the Dust Advisory.

Figure 9. Pueblo WFO forecast graphic page for Tuesday afternoon showing the location of the Dust Advisory.

Figure 10. Boulder WFO "Weather Story" graphic page for Tuesday afternoon.

Figure 11. Pueblo WFO forecast graphic page for Tuesday afternoon.

The strong winds and dust moved south into the Texas Panhandle during the late afternoon and evening hours of Tuesday (after 0000 UTC on 12 March).  An extraordinary photo taken from a low flying aircraft was posted on the WFO Amarillo Facebook Page and is shown below.

Figure 12. Great view of the arc of dust (with some small clouds) moving south near Amarillo, Texas, as taken from a low-flying aircraft. Courtesy of WFO Amarillo.

The dust then moved further south into the Lubbock (LBB) WFO forecast area.  WFO LBB made an excellent post on this event (see http://www.srh.noaa.gov/lub/?n=events-2014-20140311-wind) and images below are taken from this summary.  The first one is a view of the dust at sunset (about 8 PM Local Time) approaching Lubbock (Figure 13).  Their graphic summary of the event follows in Figure 14, then in Figure 15 is a summary of wind gusts in the WFO LBB forecast area.

Figure 13. View of the approaching dust from Lubbock, Texas near 8pm on Tuesday, 11 March.

Figure 14. Description of the dust event from WFO LBB.

Figure 15. Collection of wind gust reports (mph) from WFO LBB.

The METAR observations from LBB nicely show a period of blowing dust that continued for several hours, as shown below in the last figure.

Figure 16. METAR observations for KLBB from 1053 UTC on 11 March (bottom) to 1153 UTC on 12 March (top).

Posted in Blowing Dust (Blue-light absorption technique), Blowing Dust Detection (Split-window technique) | Leave a comment

Great Lakes Ice Cover

The exceptionally cold winter over the Great Lakes region has led to relatively high ice coverage across the Great Lakes.  The GLERL analysis of ice cover percentage across the Great Lakes is 91% as of March 4:

To put this into a historical perspective, the ice coverage across the Great Lakes is highest since 1980 according to this image from the Canadian Ice Service (Environment Canada):

The highest Great Lakes ice coverage is just prior to the above graph at 95% set in 1979.

Given the historical level of ice coverage across the Great Lakes, it is of interest to look at satellite imagery of the ice cover on the Great Lakes.  First we will look at a VIIRS image (courtesy CIMSS at University of Wisconsin – Madison):

Click on the image then click “full size is 3055 x 1499″ to see the image at highest resolution.  One of the noteworthy aspects of the image are the cracks evident in the ice evident on the Great Lakes.  These cracks can be seen to move in animated geostationary satellite imagery from the GOES satellite (imagery provided by Dan Lindsey):

The winds at low-levels play a key role in the distribution of cracks along the ice.

Posted in Uncategorized | Leave a comment

A look at the 16 Jan 2014 dust storm in eastern Colorado using VIIRS imagery

Strong winds coupled with dry soil led to widespread blowing dust across the eastern plains of Colorado and areas east on Thursday 16 January 2014.  A short video clip shows the blowing dust obscuring visibilities in Logan County in far northeastern Colorado (see www.youtube.com/watch?v=uxMdMTzlRFI ).  The lowered visibilities in the blowing dust resulted in a multi-car accident that closed Interstate 70 near Burlington Colorado (close to the Kansas border) for several hours from 11:30 AM MST to 4:25 PM (1830 to 2325 UTC).  A story of the dust storm and the accidents that closed Interstate 70,  courtesy of Denver 7 News, can be found here http://www.thedenverchannel.com/news/local-news/dust-storm-closes-eb-i-70-us-24-to-kansas-line-us-34-closed-as-poor-visibilty-causes-accidents.

Below is a surface plot from early afternoon on the 16th showing sustained wind speeds of 30 to 40 knots with gusts as high as 55 knots.

Surface plot at 2043 UTC on 16 January.

In this blog entry we will take a look at how the dust appeared in standard GOES visible imagery and compare this to different imagery from the SUOMI NPP VIIRS instrument Polar satellite.  Some of the images we will show demonstrate the capabilities that will be available when GOES-R is launched.

We start with a look at “standard” GOES visible imagery during the event, shown for two times below.  In the visible images the plains of eastern Colorado and surrounding areas have a collection of clouds at different levels in addition to some areas of old snow cover as well as the blowing dust.  From this imagery alone it is difficult to distinguish between these.

Shown below is similar type of visible satellite imagery but using VIIRS with a natural color background and of course higher spatial resolution (~0.5 km vs 1.o km from the current GOES).  Right away we can see that the improved contrast and resolution allows one to see dust plumes in the first image below from 1848 UTC, but things are not so apparent in the second image from 2029 UTC.

VIIRS True Color visible satellite image at 1848 UTC on 16 January.

VIIRS True Color visible image at 2029 UTC on 16 January.

Using different bands from VIIRS a specialized type of imagery can be created which isolates the dust, making it stand out far more clearly.  CIRA creates two types of dust discrimination imagery, one with the dust highlighted in yellow and the other where the dust has a pinkish color, as shown below.  The purpose of these products is to clearly isolate dust from other surrounding clouds and general background.

Pink dust product using NPP VIIRS imagery at 1848 UTC on 16 January

Pink dust product using NPP VIIRS imagery at 2029 UTC on 16 January

The above imagery is especially useful for the 2029 UTC time, when the earlier visible satellite image showed a more complex mix of dust and clouds.  Similar imagery is also made using MODIS Polar orbiting satellites.  Shown below are the pink and yellow dust products using the MODIS Aqua satellite imagery at 2020 UTC.

Pink dust product using Aqua MODIS imagery at 2020 UTC on 16 January.

Yellow dust product using Aqua MODIS imagery at 2020 UTC on 16 January

We noted earlier that there was also snow in the background in some spots across the Plains, as well of course in the mountains.  The next set of imagery shows the CIRA snow/cloud (discriminates snow from clouds) and snow/cloud layer (further discrimination of the type of clouds, pinkish for generally higher ice clouds and yellow for generally lower water clouds).

Snow/cloud discriminator product using NPP VIIRS imagery at 1848 UTC on 16 January

Snow/cloud layer discriminator product using NPP VIIRS imagery at 1848 UTC on 16 January

Snow/cloud discriminator product using NPP VIIRS imagery at 2029 UTC on 16 January

The snow/cloud layer discriminator product uses the 1.38 micron “cirrus” band to isolate higher level ice clouds (an added step from the binary snow/cloud discriminator product).  An example of the cirrus imagery by itself is shown below for the 2020 UTC time.

Cirrus product using Aqua MODIS imagery at 2020 UTC on 16 January

Posted in Blowing Dust (Blue-light absorption technique), Blowing Dust Detection (Split-window technique), MODIS Snow/Cloud Discriminator | Leave a comment

Fog and Low Clouds in Eastern Colorado as seen with GeoColor and Day Night Band NPP VIIRS imagery on 20 Dec 2013

The Suomi NPP VIIRS Day/Night Band (DNB), which can use moonlight to produce visible-light imagery during the nighttime, offers a unique capability that is slated to continue on the JPSS constellation concurrent to the GOES-R era.  This presents intriguing potential for synergy between the DNB and Advanced Baseline Imager.  One can use the imagery just as visible imagery is currently used in the daytime, so there are many applications.  Here we show how fog and low clouds distinctly appear during the nighttime hours using the DNB imagery, shown below first on a larger scale and then zoomed in over Colorado.

DNB imagery over the western half of the CONUS at 0914 UTC on 20 December 2013.

DNB imagery centered over Colorado at 0914 UTC on 20 December 2013.

Lights from the various towns and cities are seen in the DNB imagery.  City lights are also a part of one of the earlier CIRA Proving Ground demonstration imagery known as GeoColor imagery.  GeoColor is intended to be a survey type imagery that has a seamless transition from night to day by combining different bands and background imagery.  The current version uses a static nightime background of city lights (though it will be possible to update to quasi-real time imagery using Polar satellites), as seen in the image below for the same time as the DNB images displayed above.  An added feature of the GeoColor imagery is the use of the 11-3.9 micron difference (the standard AWIPS fog product) to color low clouds and fog in a pinkish tone to highlight them from other clouds, as seen below.  One characteristic of the imagery that has been noticed during cases of fog is that there may be some information about the thickness and/or density of the fog by how much of the city lights appear through the fog layer, and it will be interesting to examine more DNB imagery in this regard.

GeoColor imagery over the western half of the CONUS at 0945 UTC on 20 December 2013.

Note how in the GeoColor imagery higher clouds appear in white tones, and we can see that high clouds extend across portions of eastern Colorado, covering the lower clouds and fog.  However, going back to the DNB imagery,  the cirrus barely shows in this imagery (as it would with visible imagery in the daytime) and we are therefore able to see the low clouds and fog more clearly.  This example shows how VIIRS and GOES-R could be used in synergy during the GOES-R era.

Posted in GeoColor Imagery, Uncategorized | Leave a comment

Orographic Cirrus of 18 December 2013

Orographic cirrus (i.e., mountain wave) clouds can have a significant influence on temperature forecasts, particularly during the cold season when a reduction in insolation can drastically affect temperatures during the daytime.

On December 18, 2013 the CIRA synthetic 4-km NSSL-WRF ARW and NAM-Nest initialized at 0000 UTC 18 December forecasted orographic cirrus downwind of the Front Range of Colorado during the early morning hours:

http://rammb.cira.colostate.edu/templates/loop_directory.asp?data_folder=training/visit/loops/18dec13_synthetic&image_width=1020&image_height=900

the synthetic NSSL WRF-ARW is shown on the left, while the synthetic NAM-Nest is shown on the right, the loop spans from 0900 UTC 18 December – 0300 UTC 19 December (9 to 27 hr forecast).  During the mid-day time period, the forecasts begin to diverge with the NSSL WRF-ARW showing a thinning out of the orographic cirrus while the NAM-Nest does not show this trend.  Both models indicate redevelopment of orographic cirrus in the evening hours.

The NWS Boulder forecast discussion issued at 4:49 AM MST 18 December highlighted the importance of orographic cirrus potentially limiting the daytime high temperature forecast, note the use of the CIRA synthetic imagery as a forecast tool:

SHORT TERM

MAIN QUESTION TODAY IS THE AMOUNT OF CLOUD COVER AND

ITS EFFECT ON TEMPERATURES. PLENTY OF MOISTURE UPSTREAM THOUGH

THERE IS SOME VARIABILITY. WITH NOT MUCH CHANGE IN THE WIND

PATTERN…EXPECT THE WAVE CLOUD COVERING THE PLAINS TO CHANGE

LITTLE TODAY WHILE THERE MAY BE SOME DECREASE IN THE HIGH CLOUDS

OVER THE MOUNTAINS. CIRA SIMULATED SATELLITE IMAGERY SHOWS A

PRETTY SOLID WAVE CLOUD PERSISTING DOWNSTREAM OF THE FRONT RANGE

THROUGH TONIGHT WITH POOR AGREEMENT ON THE TIMING OF ANY CLEARING

ELSEWHERE. SHOULD BE ENOUGH CLOUDS TO LIMIT TEMPERATURES…BUT THE

AIRMASS HAS ALSO WARMED A LITTLE SINCE YESTERDAY. WE WILL BE OFF

TO A WARM START…ALREADY IN THE 50S IN THE LOWER FOOTHILLS. UNDER

CLEAR SKIES IT WOULD PROBABLY BE 70 IN DENVER TODAY…SO THE

CURRENT MID 60S FORECAST IS PROBABLY STILL ALRIGHT.

An interesting way to view the influence of the mountain wave is a cross section (red line) oriented east-west across the Front Range as shown here:

The cross section of potential temperature from the NSSL WRF-ARW between 1300 UTC 18 December – 0300 UTC 19 December is shown below:

http://rammb.cira.colostate.edu/templates/loop_directory.asp?data_folder=training/visit/loops/18dec13_xsec_theta&image_width=1020&image_height=900

The mountain wave clearly shows up in the vertical through the depth of the troposphere.  Also note the sloped potential temperature lines do not make it all the way to the surface, which is consistent with the lack of high downslope winds for this event.

The verifying GOES IR imagery is shown here:

http://rammb.cira.colostate.edu/templates/loop_directory.asp?data_folder=training/visit/loops/18dec13_goes_ir&image_width=1020&image_height=900

Note the thinning out of orographic cirrus by the early afternoon hours, which seems to be more consistent with the NSSL WRF-ARW forecast.  This kind of monitoring GOES imagery versus synthetic imagery can assess how much confidence to put in one model forecast versus another.

The CIMSS blog has an entry on this event which includes views from polar orbiting satellites:

http://cimss.ssec.wisc.edu/goes/blog/archives/14607

Posted in Synthetic NSSL WRF-ARW Imagery | Leave a comment

Determining clouds from snow – an example from 5 Dec 2013

As winter continues to settle in across the nation and snow cover increases, an issue that arises is trying to see cloud cover over a snow pack during the daytime with visible satellite imagery, since both appear white.  Certainly looping the imagery can help although issues can still remain.  Snow cover can have a significant effect on the local weather, including influencing both maximum and minimum temperatures, so it is important to determine the status of snow on the ground.

There are various methods that can be used to help discriminate clouds from snow cover, requiring information from different channels that are not available on the current GOES satellites.  These channels are found on some of the Polar-orbiting satellites, and similar channels will be available when GOES-R is launched.  Here we show an example of the type of satellite imagery that will be available during the GOES-R era by utilizing channels from the MODIS sensors on board the Terra and Aqua Polar-orbiting satellites, and from the VIIRS sensors on the new Suomi NPP satellite.

The first image shown below is a visible image from the Suomi NPP satellite, in this case True Color Imagery where the background is natural color.  A mix of snow and clouds exists from the Rockies eastward across the Central and Northern Plains, but it is difficult to distinguish one from the other.

Suomi NPP true color visible image at 2016 UTC on 5 Dec 2013.

Next is a CIRA product known as Snow/Cloud Layer Discriminator imagery, in this case from the Suomi NPP satellite utilizing various channels from the VIIRS sensors (details on how this imagery is constructed can be found here.

CIRA snow/cloud layer discriminator image at 2016 UTC on 5 Dec 2013.

In the image above the colors are defined as follows: green = land (clear sky, devoid of snow cover), white/bluish-white = snow cover, yellow = low-level liquid-phase clouds, and orange/magenta = mid/high level ice phase clouds.  There is a significant amount of tuning that is required for the various colors to match the phenomenon, and in the image above in some areas the snow appears with an excessively bluish tone.  The next image shows the same CIRA snow/cloud layer discriminator product but from the MODIS sensors on the Terra Polar-orbiting satellite.

CIRA MODIS snow/cloud layer discriminator image from the Terra satellite at 1805 UTC on 5 Dec 2013.

Although the channels used are similar in the two images, there are some slight variations between the two satellites.  For the snow/cloud layer discriminator product the MODIS imagery has been well tuned so that the colors show the desired contrast.  The slight color differences for the snow between the MODIS and VIIRS images reflect that more fine tuning is needed for the newer VIIRS imagery.  When GOES-R is launched similar tuning may be necessary.

Posted in MODIS Snow/Cloud Discriminator | Leave a comment

Synthetic imagery for the 3 Dec 2013 fog/low cloud case

The previous blog entry discussed CIRA satellite imagery that can be useful in highlighting fog and low clouds.  These images utilize existing satellite imagery to create images that try to replicate those that will be available in the GOES-R era.  Another method to create GOES-R type imagery is to use output from a high-resolution model (in this case NSSL’s 4-km WRF-ARW model run at 0000 UTC) to create “synthetic” satellite imagery.  An advantage to synthetic imagery is the ability to replicate many of the bands that will be on GOES-R.  We also replicate satellite imagery that forecasters currently use, such as IR, so that they can determine how well the model actually does in comparison to real-time available imagery.  A number of forecasters have found synthetic imagery to be a useful way to visualize model output, much in the way model radar reflectivity is now widely used.

One of the synthetic images being generated by CIRA replicates the AWIPS fog product, and uses the 10.35 minus 3.9 µm difference to highlight fog and low clouds as blue. Higher level clouds appear as black.  The image shown below is a 12-h forecast valid at 1200 UTC on 3 Dec, the type highlighted in the earlier blog.

Synthetic "fog product" valid at 1200 UTC on 3 Dec.

Of course this synthetic image, like its real image counterpart available in AWIPS, does not in itself discriminate between low clouds and fog.  But the model provides output that does allow one to better determine areas of fog.  This is shown in the image below, where the blue represents areas in which the cloud liquid water content at the lowest model level is nonzero, a proxy for fog.

Same 12-h forecast but here isolating moisture in the lowest model level to determine fog from low clouds.

Comparison of the two images shows that some of the areas in blue in the first image was not fog but low clouds.  Another advantage to the image above is that areas of fog are shown beneath the higher clouds (for example across Missouri and eastern Iowa) that obscure the fog in the synthetic image that includes higher level cloudiness.

Posted in Synthetic NSSL WRF-ARW Imagery | Leave a comment

Widespread Fog East of the Rockies – 3 Dec 2013

Widespread fog and low clouds covered much of the nation east of the Rockies 0n the morning of 3 Dec in the moist airmass ahead of the developing western storm and strong cold front that has since pushed south.  Dense Fog Advisories were issued by many WFOs, as seen in the graphic below valid at 1324 UTC on 3 Dec.

NWS Watches and Warnings as of 1324 UTC on 3 Dec 2013

The surface weather map at 1300 UTC on 3 Dec is shown below.  A complex storm system is taking shape in the Rockies while an Arctic high pushes cold air southward into the Northern Plains.

Surface analysis and plot at 1300 UTC on 3 Dec 2013.

A closer look at the observations is given in the plot below.  In this plot visibility is given below the station circle, weather to the right, and ceiling height (feet, AGL) to the left. Note the large number of stations reported low visibilities due to fog.  East of the dense fog area was a huge area with more scattered fog reports but widespread low overcast conditions.

METAR station plot of weather, ceiling and visibility at 1200 UTC on 3 Dec.

CIRA has two types of overview imagery that make fog and low clouds easier to see, and both will be available in improved versions in the GOES-R era.  Below we see an example of GeoColor imagery, one version with night lights and the other without, both for 1145 UTC on 3 Dec.  In this imagery low clouds and fog appear in red/pink tones, while clouds or snow cover are white.  Much of North America east of the Rockies is covered by low clouds and fog, interrupted in the imagery in some areas by higher clouds.  GeoColor imagery is intended as a general overview type of satellite imagery.

In the GeoColor imagery all clouds other than low clouds (or fog) appear white.  Another type of imagery (called Low-cloud/fog imagery) is designed to highlight low clouds and fog in a whitish color both during the night and during the day, but also colors higher clouds thereby giving more information about them.  An image also for 1145 UTC is shown below.  Note that thin cirrus appears as black during nighttime imagery, and we see that there are areas of thin cirrus covering some of the regions of low clouds and fog.

CIRA Low-cloud/fog imagery for 1145 UTC on 3 Dec.

Posted in Uncategorized | Leave a comment

October 17, 2012 fog over Wisconsin

Let’s examine the synthetic low cloud / fog product from the 4-km NSSL WRF-ARW model.  This is from the 0000 UTC 17 October run valid between 0900-1600 UTC 17 October:

http://rammb.cira.colostate.edu/templates/loop_directory.asp?data_folder=training/visit/loops/17oct12_syn_fog&image_width=1020&image_height=900

Low cloud / fog is depicted as blue in this color table, with mid- to high level clouds being black / dark grey.  Early in the loop, note the region of blue (low cloud / fog) in northern Wisconsin.  As the loop progresses, this region quickly becomes obscured by mid- to high level clouds that are moving eastward.  It’s unclear when this low cloud / fog is forecast to dissipate in northern Wisconsin due to the higher levels clouds obscuring what is happening at lower levels.

When you come across a situation such as this, one option to get around this issue is to use the NSSL WRF-ARW experimental fog product.  This is simply the model output of cloud liquid water at the lowest vertical level in the model.  Here is the loop from 0900-1800 UTC, with cloud liquid water depicted as blue:

http://rammb.cira.colostate.edu/templates/loop_directory.asp?data_folder=training/visit/loops/17oct12_exp_fog&image_width=1020&image_height=900

Fog can be inferred from the blue regions we observe in this loop.  We avoid the issue of obscuration of higher level clouds, and we may also discriminate between low cloud and fog (as forecast by the model).

Another option for fog identification is the GOES-R Fog / Low Stratus Product from CIMSS.  This product uses a blend of satellite imagery and 00 hour forecast model output (from the RAP, except it uses the GFS over oceans beyond the RAP domain)  to assign a probability of fog.  This differs from the previous experimental fog product above which is solely model output, run out to a longer forecast time.  The following loop is the GOES-R FLS product for IFR fog:

http://rammb.cira.colostate.edu/templates/loop_directory.asp?data_folder=training/visit/loops/17oct12_goes-r_fls&image_width=1020&image_height=900

The color bar is given as a percentage, so that greater percent probability of IFR fog exists in the orange to red colors.  The visibility (bottom number, miles) and ceiling (top number, hundreds of feet AGL) observations are also overlaid so that locations of fog correspond to the regions of low visibilities.  Note the regions of orange and red in northern Wisconsin indicating higher probabilities of fog which would cause IFR conditions, this closely corresponds to the observed fog from the observations.

For more detailed information on either the GOES-R fog / low stratus product, or the synthetic low cloud / fog product, see the training session from VISIT:

http://rammb.cira.colostate.edu/training/visit/training_sessions/

Posted in GOES Low Cloud / Fog Imagery, Synthetic NSSL WRF-ARW Imagery | Leave a comment