Sometimes, CIMSS Blog writers get emailed letter, like the one below (anonymized) forwarded along to CIMSS with a request

A long-dormant idea has resurrected itself in my mind, which is to look into the “biomass and carbon impact” (e.g. biomass and carbon knocked down) from some MCS and derecho type events.  We may have talked about this briefly some years ago.  Since we share an interest in NW PA, I thought I would ask if you know of any maps or diagrams that would show the affected area from the July 21, 2003 wind event…..of course it produced the tornado that brought down the Kinzua Bridge, but the event itself was much more widespread....

.....Is there an archived database of satellite imagery that would show the damage in NW Pennsylvania from the 21 July 2003 storm?

 

Your blogger first thought of the MODIS Today website — but those images “only” go back to late 2007. There are data sources that allow a reconstruction of those images, however. The big challenge in accessing MODIS data from long ago is knowing when Terra (or Aqua) overflew northwest Pennsylvania, where the Kinzua Viaduct once stood. The CIMSS Polar Orbit Tracking website (https://www.ssec.wisc.edu/datacenter/polar_orbit_tracks/) includes an archive for satellites, so it’s easy to find where Terra was flying on 19 July, 20 July, 21 July, … , 27 July (and other dates too, of course), allowing someone to identify the times of data needed. I’m looking at series of days because I don’t recall cloud conditions over northwest Pennsylvania from back in 2003! The NASA LAADS website (https://ladsweb.modaps.eosdis.nasa.gov/) contains the level-1b hdf files that can be used to create imagery. (This is also discussed in this blog post for imagery over the Himalayas). For this blog post, I searched for Level-1b files ( https://ladsweb.modaps.eosdis.nasa.gov/missions-and-measurements/science-domain/modis-L0L1/ ) and downloaded hdf files containing data to be displayed with polar2grid software. The data needed to construct true-color imagery include quarter-kilometer data (MOD02QKM), half-kilometer data (MOD02HKM) and georeferencing data (MOD03), as shown at the website below (accessible from this link). Each of those folders will have annual and then daily folders to enter; the daily files are 5-minute granules that are matched to times in the polar orbit tracks at the CIMSS website.

Day 207 files (that is, 26 July 2003) that include Kinzua are shown below. Here are the MODIS paths on that day; note the location of 1630 and 1635 UTC on that map!

MOD02HKM.A2003207.1630.061.2017194224505.hdf
MOD02HKM.A2003207.1635.061.2017194225024.hdf
MOD02QKM.A2003207.1630.061.2017194224505.hdf
MOD02QKM.A2003207.1635.061.2017194225024.hdf
MOD03.A2003207.1630.061.2017194181034.hdf
MOD03.A2003207.1635.061.2017194180729.hdf

I downloaded the data from LAADS for multiple days. To display the imagery, polar2grid (free unix-based software from CIMSS available here) is used. First, a geographic location is established for polar2grid:

./p2g_grid_helper.sh Kinzua -78.6 41.8 250 -250 1120 840 > Kinzua.yaml

That polar2grid command creates gridding information for an 1120×840 grid centered at 41.8^oN, 78.6^oW with 250-m resolution in the n-s and e-w directions. The polar2grid command to create a true_color image and then to place borders, coastlines, and a lat/lon grid are shown below.

./polar2grid.sh -r modis -w geotiff -p true_color -g Kinzua --grid-configs ./Kinzua.yaml -f /path/to/downloadedfiles/Kinzua/206/*.hdf ./add_coastlines.sh --add-borders --borders-resolution f --add-coastlines --coastlines-resolution f --add-grid --grid-D 1.0 1.0 --grid-d 1.0 1.0 --grid-text-size 12 terra_modis_true_color_20030726_??????_Kinzua.tif

The imagery created is below.

An animation from 19-27 July 2003 is shown below. An abundance of clouds is obvious. Perhaps days later in July, or in August, will show clearer skies! Still, a difference field between 19 and 25 July (that is, before and after the Kinzua destruction) might show something. A toggle between those two dates is at the bottom of this blog post.

---

Thanks to Kathy Strabala, SSEC/CIMSS, for reminding me about the LAADS data source. A next step for anyone wanting to look at that region would be to access Landsat data and hope for clearer skies.

View only this post Read Less

Very cold air over the Great Lakes on 17 February helped lead to the development of one long Lake-Effect Snow (LES) band that stretched from the Upper Peninsula of Michigan east-southeastward to Lake Huron then over the Bruce Peninsula and into eastern Lake Ontario and on to the Tug Hill Plateau over upstate New York. Surface observations show surface winds converging into the line: west-southwest winds over southwestern Ontario just south of the band and west-northwest winds just to the north of the band. In the RGB, the LES band clouds are orange/yellow in color. The cold snow-covered land surface is green, as are ice-covered regions on the Great Lakes (for example: Green Bay and the North Channel of Lake Huron). Low clouds (that probably aren’t supporting significant precipitation) are cyan or periwinkle in color.

Microwave sounder data (from, for example, ATMS on Suomi-NPP, NOAA-20 and NOAA-21 or AMSU-A/-B on NOAA-19) can be used to estimate snow fall rates, and this information is available online here. (Click here for a Quick Guide) Snowfall rates for this event are shown below. The LES snow band is obvious from Lake Superior east-southeastward into upstate New York, and strongest snowfall rates are just downwind of Lake Ontario. As the cold air sweeps from Canada southeastward across Lakes Superior, Huron and Ontario, it is warmed and moistened at the lowest levels (that is, it is destabilized) by the open waters of the Great Lakes. Once the band reaches the Tug Hill plateau, it has been prepped by the lakes upstream to unleash a torrent of snow.

Merged snowfall rate, shown below, combines MRMS radar observations with satellite-based microwave estimates of snowfall rate. As you see, the radar observations show a considerably thinner snow band. Even though the microwave footprint has a larger horizontal scale than the LES snow band, the microwave signal is large enough to show where snow is falling. A user must combine the microwave estimate of snow fall rate with their knowledge of the size of the LES — based on infrared satellite imagery, or on surface observations — to use the microwave snowfall rate estimates most effectively.

View only this post Read Less

GOES-18 infrared imagery, above, shows a cyclonic circulation over the northeastern Gulf of Alaska. Shower development is suggested by the evolution of brightness temperatures changing from cyan to blue to green (showing cooling); showers then dissipate as the brightness temperatures warm. Do you think these are showers of rain or of snow? Note in particular the motion of the clouds over Prince of Wales island, the southernmost large island of the southeast Alaskan archipelago, just north of Haida Gwaii (formerly called Queen Charlotte island) that is at the southern edge of the domain. Cloud motion is parallel to the development of colder cloud tops.

Microwave sounder data are used to estimate snowfall rates (SFR), and those estimates are available at this website; additionally SFR is inserted into Alaska Region AWIPS. Because microwave sounder data are available from many different polar orbiting satellites, coverage over Alaska can be quite solid. In particular, a cluster of observations comes in from NOAA-19 and Metop satellites; then a second cluster of observations occurs thanks to NOAA-20/NOAA-21 observations.

For this event, training convective snow showers spread over Prince of Wales Island allowing regions to receive around 8 inches of snow at sea level and over 15 inches of snow at higher elevations (as predicted well by forecast models). Satellite SFR products showed the convective snow showers over the southern Gulf of Alaska before they moved over the island and produced 1”/hour snowfall rates with visibilities of less than a quarter mile. The animation above shows the estimates during the day on 30 January; the animation below focuses on times after 1900 UTC. This case demonstrates how model information can raise situational awareness for incoming adverse weather (in the near term), and satellite information can confirm its existence in the short term as well as aid forecasters in making operational decisions to notify the public. Note how the SFR persistently identifies a snowband over Prince of Wales Island, something that continues through at least 0500 UTC on 31 January!

---

Many thanks to Spencer Fielding, WFO Juneau, for alerting me to this event. A comparison between Microwave snowfall rates and radar observations over CONUS is here. Note in particular how the width of a snowband might be overestimated by the large microwave footprint!

View only this post Read Less

It has been an active period of impactful weather in and near Australia, and today is no exception. Severe Tropical Cyclone Alfred continues to spin northeast of Queensland. The latest update from the Australian Bureau of Meteorology indicates a central pressure of 956 mb. Over the last day, it was downgraded from a Catergory 4 to a Category 3, but sustained winds are still at 85 knots (155 km/h) and gusts are reaching 120 knots (220 km/h). Recent Band 14 (infrared window) and true color imagery from AHI aboard Himawari-9 as displayed on RealEarth confirms Alfred’s weakening, with a filled-in eye that is well over 100 km across.

The current forecasted track for Alfred is to continue along a largely southerly path, and it is  expected to pass between mainland Australia and the French island of New Caledonia with no landfall.  The most significant impacts expected to be storm surges and waves. However, outer rain bands may prove to be challenging to the Queensland coast, which has already been struggling with the aftermath of recent significant rainfall.

MetOp-C passed directly over the center of Alfred on 26 February enabling ASCAT analysis of the near-surface winds. While most meteorologists already know that the strongest hurricane winds are not directly in the center of the rotation but near the eyewall, this image helps to reinforce that idea. Due to the relatively narrow swath of ASCAT and the geometry of low earth orbits, it can be challenging to get more frequent wind updates.

Alfred was one of three named tropical cyclones that were concurrently active in the tropical western Pacific Ocean this week, alongside Seru and Rae. While these latter storms were not as intense as Alfred at its peak, the impacts were more significant due to passing much more closely to the Pacific island nations of Vanuatu and Fiji. This true color AHI loop shows the rare sight of three Category 1+ tropical storms in such close proximity. Since then, the later two storms have both weakened and moved away from the populated islands.

View only this post Read Less