EWP Week 1 Day 4 Summary and Weekly Summary

On Thursday, the final day of the first week of the 2017 GOES-R/JPSS Spring Experiment, one of our pairs operated in the Boulder, CO CWA while the other group monitored activity associated with Cindy in the Jackson, MS CWA.

End-of-Week Comments:

• GOES-16 ABI Imagery and products
• Imagery
• Character of Cu in Mississippi signaled to me that convection would not initiate
• High temporal and spatial resolution was huge for us all week
• In Reno, high spatial and temporal resolution is very valuable because radar coverage is bad. Good surveillance in bad radar areas, especially for haboobs.
• Good SA to have vis/ir two panel on radar awips, monitoring new development, strengthening.
• Derived products (CAPE, LI, winds, dsi, tpw)
• CAPE was always underdone compared to spc mesoanalysis, even compared to GFS
• Gradients did seem reasonable
• LAP products were useful for tracking gradients and min/max and trends in instability/moisture
• It is valuable if the gradient locations are corrected in space by GOES LAP
• Derived winds were not realistic, and therefore not helpful, all week.
• RGB composites
• Definitely need more RGB training. Cheat sheets for all. Just how we could use them, scientific details not needed. What do colors mean in each rgb, etc.
• GLM
• Training on lightning and applications needs to be available. Forecasters are confused by all the different types of lightning, and when they should use what. Lightning is a very helpful tool, but forecasters need to know how to use it and what to use.
• GLM quick guide like what is available for 16 ABI bands would be helpful. One was made for HWT which was great, but could be improved - suggestions were made
• Interpolated imagery was odd in AWIPS (small gaps, etc)
• I liked interpolating the lightning data when I overlaid it on satellite imagery. It was nice to see OTs where we had a max in event density for one case.
• A lightning jump product will be helpful down the line (sigma jumps).
• Cases where lightning was offset from lightning/satellite, but other cases where it was right on.
• I liked overlaying it on visible, it was good for SA. Need to determine a good color table for vis overlay.
• I never view lightning by itself, usually view it as an overlay on satellite imagery or radar imagery
• Web based training cases would be preferred by me over WES cases.
• Lightning menu’s need to be made simple
• ProbSevere
• Good for situational awareness, but base data will usually beat it.
• If you have a lot of storms going on, its a good way to keep an eye on things. If color changes, ill look at it
• Helps me to prioritize storms
• As for lead time, i don’t think it gives you lead time over the base data
• It would be cool to see trends, and be able to mouse over the storm and get a graph of how the probabilities have changed over time. A lot of the good info is trends and not just the values.
• Could include  a sigma change in probsevere values like lightning jump
• Warning decision was usually made before probsevere makes the jump
• If on edge to warn it could add confidence
• Probsevere would clue me in on storms that I should investigate further, then issue warning based on base data.
• All good with display
• Didn’t have preference between lite and long. Once information is explained, there is usefulness in seeing all of the input
• No great tornado cases this week, but with Cindy, increases in probabilities were delayed. MRMS delay leads to latency in probsevere
• Fields to add to improve probs: radial change in velocity (but no mrms product for this). This would help with straight line winds.
• Way to incorporate environmental data to get probs for outflow winds - DCAPE? Max radial wind, actual wind speed should be in there somehow.
• NUCAPS
• We liked idea of having the automated modification.
• In every case we saw, the experimental modified sounding looked much better than the operational sounding, and indices appeared more on par with other datasets
• It is great that the surface is modified in the experimental sounding, but unsure of the accuracy of the mixed layer that is developed above the surface. In real life, soundings will not be perfectly well mixed all the time. Challenge is how to best blend modified surface with rest of profile.
• Operational nucaps pretty much always had cape way too low. Modified always had closer to convective environment.
• Id love this in new york. You can have completely different environment east of sea breeze as west. Nucaps allows us to look at env around sea breeze. And to see how environment has changed.
• In jackson yesterday, operational and modified nucaps had no cape. We had no convection. This was a great null case.
• Id like to see a null dryline case. See if modified nucaps would mix out cap.
• For future weeks, would be good to operate in a capped atmosphere.
• What about using a modified nucaps in winter?
• Out west, monitoring mixed layer heights is very important, especially for fire weather
• I will look at operational nucaps in my office if available prior to convection. At the very least, for finding 0C and -20C level.
• Id want to see both unmodified and modified nucaps soundings in awips. And maybe instead of two menu items, allow for swapping between the two profiles.
• It would be nice to be able to compare the soundings in NSHARP directly.
• It was helpful to be able to see a slice through the sounding swath with the plan view display.
• If there was a nucaps page online outside of awips, I would look at it if it had helpful information.
• It would be ok to overlay model wind data on nucaps sounding
• Web based option might be to have unmodified, surface modified with obs, and model influenced.
• Useful to have multiple satellites, more passes, to see trends

GOES-16 DMW Display

NWS forecasters have commented on the present display of the Derived Motion Winds (DMW's) in AWIPS-II. They do not like the display, and would prefer it be made similar to the current GOES-13/15 operational satellite-derived winds.

To start, we review the display of the operational GOES-13/15 satellite-derived winds below. Displayed are all of the winds by pressure level. One can load winds by pressure level, and wind barbs and labels are color coded by pressure level. When sampling the winds in the AWIPS-II display, the output includes: time, wind derived from IR or VIS, temperature at pressure level of wind, pressure level of wind, wind direction and wind speed. Forecasters think this display is great. They like the color coding by height, and they like all of the output that is shown when a wind is sampled. An improvement would be to get more specific about the channel from which the wind was derived (WV, IRW, SWIR, VIS). Ideally, this is what the display would look like for GOES-16 DMW's in AWIPS-II. Unfortunately, it is not...

... Below is a display of of the GOES-16 DMW's in AWIPS. Similar to what we did for the GOES-13 winds above, we loaded all of the GOES-16 winds by pressure level. Instead of by height, the winds are color coded by speed for each level. For each level, winds are grouped into a >=50 kts bin, <50 kts bin, <30 kts bin (3 bins for each level). This creates a lot of separate products to be listed. Instead, all winds for a pressure level should be loaded together in one label. When sampling the wind data, the only output is time, direction, and speed. No indication of wind level, temperature, or channel from which the wind was derived from, is given. Forecasters do not like this display, and find it makes the winds less usable. Forecasters would much prefer a display like that from the GOES-13/15 satellite-derived winds, shown above.

- Bill Line, NWS

Derived winds

Today was my first real chance to glance at the Derived Winds.  They don't seem to be doing very good.

I plotted 1000 mb winds.

Looking along the LA coast SE of Cindy's center, we get a "red" wind that is from the 87 deg, 20 knots.  Winds should be from S/SW.  Also, "red" is supposed to be >= 50 knots.  Flag shows 20 knots.

"Yellow" flag is 58 deg, 42 knots.  Again direction should be S/SW.  Color is right this time, though.

Davis Nolan
WKRN Meteorologist

Derived Motion Winds from GOES-16

This was the first time I had a chance to look at the derived motion winds from GOES-16. I loaded them on several different pressure levels and it became very difficult to figure out what pressure level a specific wind barb is coming from. I turned on sampling and did not see a pressure level indicated.

I recommend adding the height of the specific wind barb when sampling is turned on.

-Ironman

Lightning Menus are out of control

Data Overload:

A big point of contention for field forecasters is the overwhelming number of lightning data options that are available in AWIPS.  The image directly below shows example of the all menu and data options that are present. The secondary image below shows the still large amount of options in just the GLM dataset.

All options available for NLDN, ENTLN, & GLM

It would be much more beneficial to refine this list to operationally useful datasets that forecasters can quickly and efficiently use. Particularly when forecasters are already inundated with other data sources (obs, models, satellite, statistically output, gridded info, etc). This again emphasized training priorities in use and application of lightning datasets. Particularly the high impact lightning has on the public and also considering the promising use lightning has for predicting CG strikes and storm intensification.

After reviewing several of the GLM data sets, I see little use in grid spacing of 20 & 40 km which expands lightning across a far too vast area.

1 & 3 km plotting is also has strange data plotting artifacts that appear to point source like. A bit hard to see below.

I think the sweet spot seems to be in the 5 to 8 km grid which is closer to the convective storm scale which can plot storm cores and lightning moving in the environment surrounding the updraft.

I would recommend removing these other scales and maintain 5 & 8 km to reduce confusion to forecasters when this is shipped to the field.

Forecaster --Tahoe

RAH Case: GOES-16 preference for intial detection, Prob-Severe evident to not warn.

RAH: Developed a secondary line of convection seen below in radar 0.5Z.

• GOES-16 1-min VIS/IR proved the easiest way to target the strongest cell within the line via cooling IR cloud tops and shadowed towering cu with a lower sun angle.

This signature of this stronger cell came a few minutes ahead of signals from elevated cores in reflectivity and relative increases in lightning activity which really didn't show up until about 10 to 15 minutes later below.

Unlike earlier in the day,  ProbWind chances stayed relatively low (<40%) through the next 15 minutes. Based on the lack of wind reported through the day opted to not issue a warning. (Decision mainly based on these tools and not on extensive radar interrogation...aside from monitoring core aloft). ProbWind gave confidence also to not issue a warning. No reports thus far. --Tahoe

The Utility of Specific ProbSevere Hazards

In the example below, ProbWind increased from 1% to 73% between, which indicated an increasing damaging wind threat with the thunderstorms (in addition to hail). There is evidence in the radial velocity of accelerating outbound winds coincident with the time period that ProbWind increased.

Based on other environmental conditions, I decided that the wind threat had increased from 60 mph to 70 mph.

This example shows the utility of the specific ProbSevere hazards.

-Lost Met

18Z NUCAPS (Op vs Experiment), Viewing, and Latency

The 18z SNPP NUCAPS Soundings

As noted from a previous post, the RGB Simple Water Vapor image depicted some increase in water vapor in the mid/upper levels. The 18z pass may have passed a bit too early to capture this moistening, but it served as a good opportunity to compare NUCAPS with the 12z Denver ROAB.  

Overall comparisons of moisture using mixing ratio values were nearly identical when inspecting mandatory pressure levels with the 12z sounding.

  

SNPP NUCAPS Operational

 

The experimental soundings with the adjusted surface values however may have moistened the low levels too much again when assuming a perfectly mixed layer.  This more than doubled the mixing ratio at 700mb from about 4 g/kg to 9 g/kg which inflated the CAPE by about +600 j/kg. 

SNPP NUCAPS Experimental

Some final thoughts on the use of NUCAPS during the Testbed:

1.  Overall use: I find the availability of NUCAPS soundings very beneficial in the field particularly for offices in the western U.S. which are often subject to data voids. Although I think much work can be done to increase its utility in field (better temporal resolution, tailored fields of interest for the a forecast, better gridded data viewing...little ragged at times, and training for forecast operations).

2.  Temporal Resolution: It would very useful in the field to be able to get multiple passes during the day to be able to monitor evolution of fields of interest. Use of MetOpA to supplement temporal resolution is useful and would be good to have these soundings on future polar orbiters. Ideally an hourly availability would be quite useful and even game changing for potential research items such as:

• Monitoring mixed layer depth evolution which are quite important in regards to fire weather/explosive wildfire growth (Haines Index) and smoke dispersion. 
• Monitoring inversion (stable layers) which are critical for downslope wind storms, air quality, convective suppression or cap breaking. 
• Stability evolution (assuming better temporal resolution) for aviation, particularly glider pilots. 

3. Experimental NUCAPS: I do see the high utility in having both the operational NUCAPS soundings available alongside the experimental surface obs adjusted values. Even though I think the correction may be oversimplified in assuming a perfectly well mixed layer of moisture, it is still useful to see.

4. Data visualization: I think both soundings and gridded data are necessary to get the full potential of the NUCAPS data.

• The soundings are useful in analyzing all the traditional features of interests that we use with upper air. (Stability, moisture layers, lapse rates, etc). 

• The gridded NUCAPS I found were a little clunky at first but currently there are no easy menu options to view data and would have to build everything using the volume and product browsers. This would need to be streamlined greatly.  The high vertical resolution is also valuable in that it allows you to evaluate plan views of temperature, moisture, possibly even heights of temperature surfaces (example freezing levels, -20, inversion heights) and whatever pressure level is available.  This evaluation of gradients would be easy to do this way.

5. Data Latency: Simply, the quicker we can get the data the better. Particularity getting both the operational and experimental profiles at near the same time would be ideal to be to do analysis on both at the same time.

Forecaster: Tahoe

Extracting data from NUCAPS Soundings and Potentially Useful Gridded Data...

Yesterday, one of the forecasters came up with a novel use of the NUCAPS soundings data.  That is, extracting information about the altitude of the 0C, -10C and -20C levels for use in radar evaluation of thunderstorms.  In this sample NUCAPS sounding from 18 UTC today in SW Kansas, notice that the freezing (FZL), -20C and -30C levels are noted on the soundings.

Using the sounding in this way, a forecaster can perhaps better determine the altitude of say the -20C threshold.  Corresponding that with the altitude of dBZ returns in radar data, might suggest something about the size of hail aloft.  Forecasters commonly use model soundings to obtain the freezing or -20C levels, but this represented a great use of the NUCAPS data, and I again wanted to mention this.

So, another point to this post is to mention that this process can be a little laborious...the issue of clicking on several soundings in an area and then trying to assess the altitude of the -20C temperature level.  It would thus be fantastic to see this type of analysis available via the NUCAPS gridded data.  That is, the altitude of a constant freezing level surface or a -20C surface.  However, the data are available only in millibar levels and not in terms of MSL or AGL.  Speaking with Nadia Smith, this may be possible in future iterations of the data, which would offer more utility for these types of applications...in addition to the NUCAPS Cold Air Aloft project with Alaska offices (who desired the data in terms of Flight Levels).

Kris W

Original NUCAPS vs Experimental NUCAPS in Dodge City, KS Area

A NUCAPS pass over the Dodge City, KS area occurred around 18 UTC on June 22, 2017. Here is a sounding from this region.

The experimental NUCAPS came into AWIPS about 30 minutes after the original data. Here is the sounding from the same location.

The original profile had limited CAPE, around 500 J/KG.  The experimental profile had much more when the surface observations were included. The CAPE value was around 3,300 J/KG. This value was much more comparable to what was seen on the SPC mesoanalysis in the Dodge City region.

One other quick note: The Baseline Derived CAPE at the same time as the NUCAPS pass was also much lower than the CAPE derived from the experimental sounding.

It appears the reason why the original NUCAPS sounding was more stable was because the surface dew point was in the mid 50s where as actual obs and the experimental sounding had dew points in the mid 60s.

Lastly, I noted there was a gap in the experimental data points with this pass, that was not seen in the original NUCAPS.

Experimental 

Original

-Ironman

NUCAPS Experimental soundings showed why storm diffrences in our CWA; also de-bunked earlier tornado worry

Again, we were working the Jackson, MS CWA where our strong storms were in the western sections of our CWA, and much weaker in our eastern counties.

Also as mentioned before, I have observed that the Experimental (modified) NUCAPS is essential.

The difference between the west NUCAPS Experimental between the western sections of our CWA and the eastern sections matched with what was happening on radar.  Storms were much weaker to the east.

West counties NUCAPS Experimental:

East Counties Experimental:

Meanwhile, these soundings de-bunked my worry of tornadoes due to dry air over-riding the tropical moisture I thought I was seeing in 4 Panel WV.

NUCAPS showed that instead of dry air mid level upward, it extended all the way down to 850mb.

Davis Nolan
WKRN Meteorologist

Gridded NUCAPS Issue .... followup...

I haven't noticed it until now, but when the ~18/19 UTC NUCAPS gridded data swath came into AWIPS today...it came in two different chunks.  This first image shows the northern part of the swath that came into AWIPS for the central/eastern US in the early afternoon.  Part of this swath was labeled 18 UTC and part was labeled 19 UTC (Image 1). 

Image 1.  Sample swath (925 MB Mixing Ratio, 19 UTC 22 Jun 2017)

The next swath along/near the West Coast came in at about 20 UTC.

Image 2.  Sample swath (925 MB Mixing Ratio, 20 UTC 22 Jun 2017)

However, once this last swath came in...part of the previous swath was no longer avaialble...the swath that was labeled 18 UTC and covered much of the remaining portions of the central/eastern CONUS.

So, this is a followup to the previous post where the forecaster noted the absence of available gridded data over the southern/central High Plains.  I'm not sure what is happening, as the purge rules are set lengthy enough to retain more than two swaths (in fact, it's set to 24).  It may have something to do with the domain matching.  Anyway, this is worth further investigation.

Kris W

Huge difference in modified NUCAPS (Experimental) from unmodified

We were working the Jackson, MS CWA east of Trop. Depression Cindy.  Our strongest storms were near the Mississippi River.  We checked out soundings there.  Here's one of the unmodified NUCAPS soundings.

It does not look unstable and has a sfc temp of 74F and DP of 63F.  All CAPE values zero.

However the Experimental modified sounding looked like what we expect to see after using a sfc temp of 85F, DP 74F.

Now much more unstable, and SB CAPE 2625, MUCAPE 2630.

The modification is essential.

Davis Nolan
WKRN Meteorologist

Gridded NUCAPS Issue

I was going to look at gridded NUCAPS data over the Dodge City, Kansas area. However, when I went to load the gridded fields it was only showing a pass to the north earlier and then jumping to a new pass off to the west.

-Ironman

GLM Selection

Throughout my analysis at the HWT, I have found no use out of loading any of the sources within GLM and GLM Grid. The GLM combined is the one I have gravitated to and I have chosen 5 minute Flash Point + Event Density (1-minute update).  I did not find any of the longer accumulation times useful during warning operations or during pre-storm environment. The 5 minute Flash Point + Event Density (1-minute update) gave me the most important information as storms were developing and as storms were maturing.

-Ironman

NUCAPS Soundings and a Tale of Two Air Masses...Part II

Ok, so now for some observations of gridded parameters in this same region.  I decided to take a look at mid-level lapse rates and see if the gridded data showed differences across these two areas on the eastern flank of the broad low and the western flank of the upper ridge.

Image 1.  NUCAPS gridded 700-500 MB lapse rates over the SE CONUS, ~18 UTC 22 June 2017.

First, let me note that the gridded data came into AWIPS here at HWT about 30 mins or so before the NUCAPS sounding data.  Notice that lapse rates were generally higher in eastern portions of the area over Georgia than they were over Mississippi.  Values were near 6 C/km over some areas.  All else being equal, this would suggest stronger convection shouuld be expected over Georgia than Mississippi...and indeed, this is what has occurred.

This next image shows mixing ratios over this scene at this same time...

Notice here too, that values of mixing ratio (g/kg) were also generally higher over Georgia than Mississippi, again suggesting better chances for convection over Georgia.

NUCAPS Soundings and a Tale of Two Air Masses

Sometimes, things don't work out the way you think they will.  The world of meteorology?  Well, do I REALLY need to state how different things can turn out?  Probably not.

So, today we initialized to the Jackson MS office with the through that better instability might develop in the swath of drier mid/upper level air entering portions of southern Mississippi. Here is a scene showing GOES-16 0.64 um imagery overlaid with partially transparent 10.34 um imagery and NUCAPS soundings (green, yellow, red dots).

An earlier blog post showed some sample soundings from this environment, and I'll show another here briefly for quick reference.  This NUCAPS sounding was taken from near Jackson, MS at ~1830 UTC, and was fairly representative of soundings in the area.  

 Meanwhile, here is another generally representative sounding from near Atlanta, GA at the same time...

Notice the much higher instability that was present generally in Georgia early this afternoon.  What has happened?  Well, there has been much more convection and impactful convective cells in central and northern Georgia so far today, with multiple tornado warnings.  the same goes for Alabama, but unfortunately, soundings were not really useful over much of Alabama due to the thick, more uniform cloud cover.

Next will be a post about some observations of gridded data in this region...stay tuned!

Kris W

GLM in Tropical Environment

There have not been many deep storms in the Jackson, Mississippi region today. However, for situational awareness I loaded the local radar with GLM flashes. I was able to quickly see where the stronger storms were based on where there were more GLM flashes.

I have found it easier to visualize the GLM data without the events in this type of display and it may be better to have the GLM event grid data on a separate panel.

-Ironman

ProbSevere Assists in Warning Decision Making in BOU

For a thunderstorm in east central Colorado, ProbSevere rapidly increased from 35% to 87% between 2056 to 2108 UTC:

Based on this trend, I went ahead and took a look at all-tilts to determine the thunderstorm was severe and issued a warning.

In this example, ProbSevere was an excellent situational awareness tool.

-Lost Met