Well, the 2011 Spring Experiment is now officially over. Ironically I just made some major facelifts to the blog, including a "follow by email" tool on the right-hand side. I would encourage everyone to follow this blog throughout the year as it is likely to be used by the UW-CIMSS local testbed with the Milwaukee/Sullivan WFO. We may also have an informal late summer experiment here in the HWT regarding fire weather that may have some interesting results.
I would like to thank everyone that visited this year and participated in our activities. Without visiting scientists, the feedback gained and interactions with the forecasters would have been greatly reduced. I would also especially like to thank everyone who helped set up and run the Spring Experiment this year. There are too many to list here, but without their efforts the Spring Experiment would not have happened.
Thank you all!
-Chris Siewert
Friday, June 10, 2011
EWP end of week debrief... 10 June
Today marks the last day of the last week of the experiment. As is tradition, we debriefed the EWP visiting forecasters on their experiences throughout the week...
Convective Initiation
- (Thursday event - N.E.) Cloud-top cooling products seemed to work in diagnosing the strength of storms on the southwest edge of the line that were newly developing.
- Even though CI didn't always occur... false hits were useful in identifying clouds trying to break the cap.
- Forecasters not interested in seeing a binary yes/no output.
- "There were instances where similar looking clumps of Cu that one would flag for CI but the other wouldn't... so I wasn't sure how to interpret that other than maybe this area was more conducive to further development."
- There are lots of products that provide you lead time on CI, the real question would be on the consistency of the output.
- "I would definitely look at this in my WFO... especially the cooling-tops product gave me a lot of information on the relative strengths of the storms."
- "If both groups could work together and come up with a probabilistic product that combined the strengths of the UAH and the UWCI products, that would be very useful."
- "I think from a purely public forecast perspective, especially this time of year where we get convection every day, it would help you identify when exactly CI will occur."
- Would be very valuable for nocturnal CI.
Nearcast
- (Thursday event - N.E.) "I didn't see a whole lot of trend in terms of gradients developing, but all the sudden on the back side of the squall-line we lost data rapidly, probably due to cloud cover... compared to the other days, I didn't see any real patterns."
- (Thursday event - KS/OK) Showed an arch of destabilization between 2200-0300 across the eastern halves of OK and KS... storms formed on the western edge of this gradient and forecaster did not expect the storms to diminish anytime soon and thus increased warning confidence... stronger wording regarding hail/wind potential in warning was issued.
- There seemed to be small scale features in the fields, areas of relative maximum that were moving around... would be nice to compare to radar evolution and see how those areas affected the storm structure.
- Helped understand why convection occurred and where it would occur... definitely the 1-6 or 1-9 hour timeframe was the most useful aspect of it.
- Having a 4-panel set up of the individual layers in addition to the difference field to help increase the understanding of the product.
- The color-table in AWIPS was poor... Also, the values were reversed from those in NAWIPS and on the web. The individual layers of PW were also not available in AWIPS.
Would it be useful to extend the Nearcast another 3 hours, even if that meant smoother fields?
- "I like the high resolution out to whenever we can have it... it seemed that there was definitely information within the gradients... of course if you add 3 hours, we will definitely take that."
Would you have used the observations without it being advected forward?
- Wouldn't have been as useful... It helped determine the evolution of the environment... The forecast parts tended to build areas of increased instability that helped provide guidance on what was going to happen later on.
Overshooting-top / Thermal Couplet
- (Thursday event - KS/OK) None were detected when forecasters expected to see detections, so was not used, especially with the rapid updates of radar data.
- Need to remember to turn icon density to "MAX" within AWIPS or some detections will be lost.
- Looked at pretty much every day... for the most part, operator identified OTs occurred before the algorithm did.
Pseudo-GLM
- (Thursday event - OK) Some of the storms to the east had higher flash rates, but this was an artifact of the LMA network's detection efficiencies.
- (Thursday event - OK) Flash rates would pick up a short time before increases in reflectivity.
- Was useful for diagnosing lightning danger... get a lot of calls from the public regarding that within the WFO.
Overall
- Would definitely help to have some pre-configured procedures before forecasters arrived... forecasters used the "ultimate CI" procedure heavily and liked to see what we think they should be combining to help enhance the utility of the products. Forecasters can then adjust the color-scales for their own preferences.
- "I liked the morning/evening shift idea... that was nice... got to experience something different everyday."
- Sometimes the forecasters would get to tied into warning operations mode and forget to look at the experimental products... would be nice to make sure that the forecasters understand that there is no real pressure and to take their time to examine all of the experimental stuff.
- Would have been useful to provide the forecasters with the training via visitview prior to arrival, as well as some cases that the forecasters could use to get familiar with the products and decrease spin-up time... This would have to be done very far in advance.
- Interaction with the EFP CI desk was not done because there were so many participants over in the area.
- EFP/EWP daily briefing was seen to be repetitive between all the groups and not very "brief" as one forecaster stated. Also, the information being provided was more academic than pure weather discussion.
- WE NEED MORE CHAIRS!
Convective Initiation
- (Thursday event - N.E.) Cloud-top cooling products seemed to work in diagnosing the strength of storms on the southwest edge of the line that were newly developing.
- Even though CI didn't always occur... false hits were useful in identifying clouds trying to break the cap.
- Forecasters not interested in seeing a binary yes/no output.
- "There were instances where similar looking clumps of Cu that one would flag for CI but the other wouldn't... so I wasn't sure how to interpret that other than maybe this area was more conducive to further development."
- There are lots of products that provide you lead time on CI, the real question would be on the consistency of the output.
- "I would definitely look at this in my WFO... especially the cooling-tops product gave me a lot of information on the relative strengths of the storms."
- "If both groups could work together and come up with a probabilistic product that combined the strengths of the UAH and the UWCI products, that would be very useful."
- "I think from a purely public forecast perspective, especially this time of year where we get convection every day, it would help you identify when exactly CI will occur."
- Would be very valuable for nocturnal CI.
Nearcast
- (Thursday event - N.E.) "I didn't see a whole lot of trend in terms of gradients developing, but all the sudden on the back side of the squall-line we lost data rapidly, probably due to cloud cover... compared to the other days, I didn't see any real patterns."
- (Thursday event - KS/OK) Showed an arch of destabilization between 2200-0300 across the eastern halves of OK and KS... storms formed on the western edge of this gradient and forecaster did not expect the storms to diminish anytime soon and thus increased warning confidence... stronger wording regarding hail/wind potential in warning was issued.
- There seemed to be small scale features in the fields, areas of relative maximum that were moving around... would be nice to compare to radar evolution and see how those areas affected the storm structure.
- Helped understand why convection occurred and where it would occur... definitely the 1-6 or 1-9 hour timeframe was the most useful aspect of it.
- Having a 4-panel set up of the individual layers in addition to the difference field to help increase the understanding of the product.
- The color-table in AWIPS was poor... Also, the values were reversed from those in NAWIPS and on the web. The individual layers of PW were also not available in AWIPS.
Would it be useful to extend the Nearcast another 3 hours, even if that meant smoother fields?
- "I like the high resolution out to whenever we can have it... it seemed that there was definitely information within the gradients... of course if you add 3 hours, we will definitely take that."
Would you have used the observations without it being advected forward?
- Wouldn't have been as useful... It helped determine the evolution of the environment... The forecast parts tended to build areas of increased instability that helped provide guidance on what was going to happen later on.
Overshooting-top / Thermal Couplet
- (Thursday event - KS/OK) None were detected when forecasters expected to see detections, so was not used, especially with the rapid updates of radar data.
- Need to remember to turn icon density to "MAX" within AWIPS or some detections will be lost.
- Looked at pretty much every day... for the most part, operator identified OTs occurred before the algorithm did.
Pseudo-GLM
- (Thursday event - OK) Some of the storms to the east had higher flash rates, but this was an artifact of the LMA network's detection efficiencies.
- (Thursday event - OK) Flash rates would pick up a short time before increases in reflectivity.
- Was useful for diagnosing lightning danger... get a lot of calls from the public regarding that within the WFO.
Overall
- Would definitely help to have some pre-configured procedures before forecasters arrived... forecasters used the "ultimate CI" procedure heavily and liked to see what we think they should be combining to help enhance the utility of the products. Forecasters can then adjust the color-scales for their own preferences.
- "I liked the morning/evening shift idea... that was nice... got to experience something different everyday."
- Sometimes the forecasters would get to tied into warning operations mode and forget to look at the experimental products... would be nice to make sure that the forecasters understand that there is no real pressure and to take their time to examine all of the experimental stuff.
- Would have been useful to provide the forecasters with the training via visitview prior to arrival, as well as some cases that the forecasters could use to get familiar with the products and decrease spin-up time... This would have to be done very far in advance.
- Interaction with the EFP CI desk was not done because there were so many participants over in the area.
- EFP/EWP daily briefing was seen to be repetitive between all the groups and not very "brief" as one forecaster stated. Also, the information being provided was more academic than pure weather discussion.
- WE NEED MORE CHAIRS!
Thursday, June 9, 2011
Overshooting-top/thermal couplets over Wichita, KS not detected
EWP forecasters are in the act of warning on storms across KS and OK. One of the more notable storms this evening formed west of Wichita, KS and began moving east towards the northern part of the city. The storm had a very nice hook echo and many reports of funnel clouds and large hail. The forecasters were asked to examine the overshooting-top / thermal couplet products while they were making their warnings. Unfortunately no overshooting-tops were detected... which means there was no chance of detecting a thermal couplet because it requires an overshooting-top detection. Starting from top left and moving clockwise the image above shows the standard IR window, visible, overshooting-top magnitude and NLDN lightning detections within the 4-panel AWIPS D2D display. While it is fairly clear in the IR window and visible imagery at 2345 UTC that there is an overshooting-top, as well as a nice wake signature in the visible, no overshooting-top was detected. At the time of this blog post the storm still has not had an overshooting-top detection.
It was hypothesized that initially the minimum IR brightness temperature requirement for the overshooting-top product was not reached early on. Then once the storm continued to grow, the spatial tests to identify isolated areas of relatively cold cloud tops associated with an overshooting-top failed because cold cloud top temperatures were too close together. These problems are likely due to the relatively poor spatial resolution of the GOES-13 imager and the product would be much improved if resolutions were similar to what will be available on ABI.
Some oberservations on my last day of observations
Today was an active day for observing CI across the country today. The initiation over the North East this morning had people scurrying to get AFD's, projected areas of interest, and survey's out in time to observe the storms. Hey, that is the point of testing experimental data, to see if it works.
For SATCAST, we picked up the first of the big convective storms moving across New York this morning with about 15-20 minutes of lead time. The UWCI product picked it up as well, just one satellite image later. Unfortunately, the storms started to fire, so time spent looking at CI products pretty much ended right then. Watching online using the research display, SATCAST looked to be doing pretty well picking out the new convection south of the line as it developed.
Watching again this afternoon over the dryline area over Texas, Oklahoma, and Kansas, SATCAST was putting up sporadic flags of CI in the late afternoon. In the early ones, the environment was not letting much grow, so the early flags didn't verify. SATCAST and UWCI put up flags simultaneously on the storm that became the large system that kept forecasters busy in the Wichita area. Before the forecasters switched their displays, SATCAST was flagging the convection that appeared about 20-30 minutes later in the southwest corner of Texas moving from LUB forecaster area to the OUN area.
Of course, the forecasters are now focused on tracking storms, putting out warnings, and using the other experimental products such as 3DVAR, the PGLM Lighting information, and the UW cloud top cooling rates, overshooting tops, and thermal couplet data to help them do it. This week's experience with the tools presented, how they were used, as well as interpreted, has provided both the forecasters and scientist with new ways to play with the data. SATCAST has performed well this week, maybe not perfect in some people's eyes, but well. It will be interesting to see where all of the tools presented go from here.
Utilizing PGLM and Nearcast in warning ops
Forecasters are once again engaged in warning operations within the EWP's evening activities. Bill Bunting (MIC DFW) is currently using the PGLM instantaneous flash extent density, as well as the Nearcast differential theta-e products while issuing warnings for storms over NW OK (see image above). Like forecasters from previous weeks, the Nearcast differential theta-e field provides information on near-storm convective instability and can provide the forecaster with confidence on whether or not the storms are expected to intensify or dissipate. Currently the PGLM is showing healthy flash densities, while the NLDN is only showing sparse CG activity within the storms.
Integrating CI information into a single situational awareness tool
Discussions with forecasters throughout this year's Spring Experiment have raised an interesting topic regarding how to supply information without overwhelming forecasters during forecast and warning operations. In particular, forecasters have expressed interest in looking at all of the background interest fields from the UAH SATCAST product, which is currently being used as a proxy for the GOES-R CI product. The official GOES-R CI product will have 12 interest fields, all with a unique piece of information about the growth of a cumulus cloud object. As defined within the current GOES-R Algorithm Working Group (AWG) requirements, the CI product will only provide the forecasters with a binary yes/no output. Currently being provided within AWIPS/NAWIPS during the Spring Experiment, SATCAST has 6 interest fields that determine whether or not a cloud object is expected to convectively initiate within the next 0-2 hour time-frame. Forecasters have expressed some displeasure with only being provided a yes/no field to help them determine whether CI will occur. Some ideas have been offered that include providing a probabilistic CI nowcast, similar to what UWCI provides, or providing them with all of the CI interest fields individually in addition to the yes/no nowcast. Providing the forecasters with all of the interest fields and expecting them to look at them within their AWIPS/NAWIPS display is not feasible as it will be extremely distracting and time consuming. The problem with providing a probabilistic approach is that you are removing the detailed information that the forecasters would like to see but currently do not have to time to look at. So what do we do?
An interesting solution to this problem arose during a car ride on the way to pick up dinner this evening during a break in EWP operations. Why not provide a display that looks similar to a binary yes/no product, but be able to (on mouse-over perhaps) interrogate all of the interest field information? Currently within AWIPS you are able to do this for a single field. For instance, you can display radar reflectivity and interrogate the actual dBZ value on mouse-over. What if it were possible to provide information from multiple fields on mouse-over in this similar manner while still providing a simple to understand one-size-fits-all display for the product? Some food for thought.
An interesting solution to this problem arose during a car ride on the way to pick up dinner this evening during a break in EWP operations. Why not provide a display that looks similar to a binary yes/no product, but be able to (on mouse-over perhaps) interrogate all of the interest field information? Currently within AWIPS you are able to do this for a single field. For instance, you can display radar reflectivity and interrogate the actual dBZ value on mouse-over. What if it were possible to provide information from multiple fields on mouse-over in this similar manner while still providing a simple to understand one-size-fits-all display for the product? Some food for thought.
Daily debrief... 9 June
Today we met up with the EWP forecasters following the joint EFP/EWP daily map discussion and talked a little about what happened yesterday. Since there was some interesting weather from yesterday we had a good chance to capture some feedback from the forecasters regarding the experimental products...
Convective Initiation
- UWCI triggered CI occurring right as the appearance of a 50 dBZ echo on radar... it was so unstable that it didn't really have much of a chance.
- "I liked the cooling cloud tops product... gave you some awareness of what storms you should pay attention to during warning ops and which ones were strongest."
- "I found it useful this week and look forward to that era when we can get those increased temporal refresh rates. This would be very useful."
- In a more dynamic situation this could be very useful.
We asked the forecasters if they had a preference to see more or less signal (more vs less FAR)...
- "I like looking at them both side by side... you can come up with a poor man's ensemble for CI because we know the limitations of both."
- "There may actually be a benefit of some false hits because it may be telling you that the cap is strong... it can give you a sense of what's going to happen in the short-term."
Nearcast
- The precipitable water product did give a sense of where the convection would start over SW WI.
- There was a sharp theta-e gradient and that seemed to be where storms focused.
Overshooting-tops / Thermal Couplet
- Thermal couplet showed up over SW WI, but warning was already issued and radar was showing clear signals of severe weather.
- Interesting to see all the thermal couplets over northern Maine... might be a good case to examine.
- There were many overshooting tops over the WI area, however forecasters were too engulfed in the rapidly updating radar data.
Today EWP warning operations are going to stick around New England with the ongoing severe convection. We are hoping that some storms can back-build towards the DCLMA so that we can get some PGLM data, although this is conditional. We may also target over Oklahoma if things get going there, but that too is very conditional.
Convective Initiation
- UWCI triggered CI occurring right as the appearance of a 50 dBZ echo on radar... it was so unstable that it didn't really have much of a chance.
- "I liked the cooling cloud tops product... gave you some awareness of what storms you should pay attention to during warning ops and which ones were strongest."
- "I found it useful this week and look forward to that era when we can get those increased temporal refresh rates. This would be very useful."
- In a more dynamic situation this could be very useful.
We asked the forecasters if they had a preference to see more or less signal (more vs less FAR)...
- "I like looking at them both side by side... you can come up with a poor man's ensemble for CI because we know the limitations of both."
- "There may actually be a benefit of some false hits because it may be telling you that the cap is strong... it can give you a sense of what's going to happen in the short-term."
Nearcast
- The precipitable water product did give a sense of where the convection would start over SW WI.
- There was a sharp theta-e gradient and that seemed to be where storms focused.
Overshooting-tops / Thermal Couplet
- Thermal couplet showed up over SW WI, but warning was already issued and radar was showing clear signals of severe weather.
- Interesting to see all the thermal couplets over northern Maine... might be a good case to examine.
- There were many overshooting tops over the WI area, however forecasters were too engulfed in the rapidly updating radar data.
Today EWP warning operations are going to stick around New England with the ongoing severe convection. We are hoping that some storms can back-build towards the DCLMA so that we can get some PGLM data, although this is conditional. We may also target over Oklahoma if things get going there, but that too is very conditional.
This week in lightning safety...
There were a couple news stories this week that caught my interest regarding the weather that are directly applicable to what we are trying to do here within the HWT with the Psuedo-GLM data. On Wednesday afternoon, 77 Air Force ROTC students were struck by lightning on a military base near Hattiesburg, MS...
http://www.cnn.com/2011/US/06/08/mississippi.lightning/index.html?npt=NP1
Below is another story about a man who was thrown 8 feet by a lightning strike on Tuesday night in Morganton, NC. There is an interesting quote from the story that points out the particular dangers in trying to anticipate lightning threats... "According to Davis, the sun was still out and the storm appeared to still be about ten to 15 miles away."
http://morganton.wbtv.com/news/people/man-struck-lightning-thrown-eight-feet/63485?hpt=us_bn5
I am very grateful that no one was killed in these events, and I would like to repost a blog entry written by NASA SPoRT scientist Geoffrey Stano from a couple weeks ago regarding the use of the GLM and lightning safety... READ HERE.
Also, Geoffrey Stano posted about how total lightning measurements of in-cloud flashes preceded the first occurrence of a cloud-to-ground strike... READ HERE.
It will be interesting to see how the GLM data are used within NWS operations for possible lightning safety forecasts in the future.
http://www.cnn.com/2011/US/06/08/mississippi.lightning/index.html?npt=NP1
Below is another story about a man who was thrown 8 feet by a lightning strike on Tuesday night in Morganton, NC. There is an interesting quote from the story that points out the particular dangers in trying to anticipate lightning threats... "According to Davis, the sun was still out and the storm appeared to still be about ten to 15 miles away."
http://morganton.wbtv.com/news/people/man-struck-lightning-thrown-eight-feet/63485?hpt=us_bn5
I am very grateful that no one was killed in these events, and I would like to repost a blog entry written by NASA SPoRT scientist Geoffrey Stano from a couple weeks ago regarding the use of the GLM and lightning safety... READ HERE.
Also, Geoffrey Stano posted about how total lightning measurements of in-cloud flashes preceded the first occurrence of a cloud-to-ground strike... READ HERE.
It will be interesting to see how the GLM data are used within NWS operations for possible lightning safety forecasts in the future.
Wednesday, June 8, 2011
Killing Time..
Sitting amongst the many weather aficionados, scientists, and forecasters in the HWT, the afternoons convective activity is being actively discussed, weighed, and contemplated. Where will be the area of interest be for the afternoon that all of the groups will place their focus?
While all of this goes on, I have been spending my time listening and watching the satellite and radar displays, checking to see how SATCAST is doing. Early this morning, a short finger of clouds sitting diagonally over eastern Iowa, northern Illinois and southern Wisconsin triggered some convection along the line. An example of new convection just north of the main line is shown below producing radar echoes in about an hour. This was repeated along the line in the 2-3 hour period I was watching.
AWIPS Display at 1332 and 1432 UTC.
In other parts of the country, mostly around the Gulf of Mexico, the last few days have provided lots of examples of individual instances along the coast and across Florida where SATCAST was able to forecast initiation with an average of 30 - 45 minutes of lead-time. Radar images taken from the NWS radar pages...
As the afternoon progressed, the number of storms that were caught and developed into the late afternoon all along the Gulf Coast did a good show showing the utility of the SATCAST algorithm.
SATCAST images at 1732 and 1745 UTC.
Mobile, AL radar at 1744 and 1826 UTC.
Tallahassee, FL radar at 1743 and 1829 UTC.
For today, the SATCAST image at 1715 UTC with corresponding imagery from the NWS’s regional weather radar loops, the CI indicators are starting to verify to the south along the coast, while faint radar echoes are becoming visible to the North and East into Alabama, Georgia, Tennessee and North Carolina.
Lower images, 1758 and 1918 UTC SE Region radar image
While all of this goes on, I have been spending my time listening and watching the satellite and radar displays, checking to see how SATCAST is doing. Early this morning, a short finger of clouds sitting diagonally over eastern Iowa, northern Illinois and southern Wisconsin triggered some convection along the line. An example of new convection just north of the main line is shown below producing radar echoes in about an hour. This was repeated along the line in the 2-3 hour period I was watching.
In other parts of the country, mostly around the Gulf of Mexico, the last few days have provided lots of examples of individual instances along the coast and across Florida where SATCAST was able to forecast initiation with an average of 30 - 45 minutes of lead-time. Radar images taken from the NWS radar pages...
As the afternoon progressed, the number of storms that were caught and developed into the late afternoon all along the Gulf Coast did a good show showing the utility of the SATCAST algorithm.
For today, the SATCAST image at 1715 UTC with corresponding imagery from the NWS’s regional weather radar loops, the CI indicators are starting to verify to the south along the coast, while faint radar echoes are becoming visible to the North and East into Alabama, Georgia, Tennessee and North Carolina.
A rough day all around for CI forecasting
To bring us up to date, yesterdays forecast domain for the afternoon warning program forecasters centered on the area covered by the Grand Forks and Duluth CWAs. These areas were decided upon using the inputs from all of the participants of both the EFP and EWP, based on model input, observations, and the vast experience present in the room. My last entry ended with a dinner break to pass the time waiting for CI to initiate somewhere... anywhere.
Unfortunately, Mother Nature did not cooperate and the area stayed pretty well capped into the forecast period. In the short time I was away, the decision was made to shift the focus area to Bismarck, ND, where we had a better chance of seeing some storms. Shortly after my return, some storms popped, giving the forecasters something to look at, analyze, and even issue some statements and a warning on.
So how did SATCAST do? Well, we got the convection in North Dakota as the system moved east, approaching out original focus area. It did catch the 3 cells that caused the increased activity, with lead-times of about 30 minutes as well as some other areas of convection as the afternoon progressed.
SATCAST image at 1845 UTC
The area of clouds (marked by the green arrow) consistently flagged CI in subsequent satellite images. These storms started to show up on radar at 1958 UTC and went on to be warned on as early as 2330 UTC.
NMAP2 radar reflectivity at 1958 UTC.
NMAP2 radar reflectivity at 2337 UTC.
Unfortunately, SATCAST also indicated CI sporadically across the northern part of Minnesota that never had a chance of getting to verify. The capped environment just couldn't be pushed through by the rapidly growing systems that tended to pop up along the boundary. End result, some False Alarms.
A known problem area of the techniques includes high CAPE areas that are capped. The algorithm exploits the satellite signals to flag rapidly growing clouds. Although there are post-processing techniques in place to help eliminate false alarms (such as masking out areas that have no CAPE) it doesn't pick up as well in areas where significant CIN is also present. We as a research group are still hunting for the "best" way to reduce the false alarms without masking out the good data. Also compounding the problem in this area, was the presence of high level cirrus associated with the low-pressure system. The cloud-typing algorithm does a good job identifying ares that need to be masked out, yet high level cirrus seems to be the bane of both CI algorithms tested during this experiment, as evidenced by the CI flags along the boundaries of the system as the afternoon progressed.
In contrast, the UWCI algorithm did not have as many CI flags in the Grand Forks/Duluth focus areas. The one or two flags they did have, verified shortly thereafter. And yes, the UAH SATCAST algorithm did pick up on those cases as well. To be fair, I did not get a chance to see if UWCI flagged the storms in the Bismarck area later, so I am only commenting on the 2-3 hours over the first focus area. But it does bring up an interesting question, is it better to have a few indicators that verify at the cost of missing many of the other cases, or is it better to have more indicators in/of the environment, that although will catch the definite cases of CI, but will also result in higher false alarms?
I asked each of the forecasters and got a few different answers. For one, watching where SATCAST flagged rapidly growing clouds caused him to focus in on the associated environment to ensure a good understand of what is occurring at that time and check for agreement with previous thinking. If he wasn't looking at it, he was checking to see if the environment had changed since he looked last. In short, a situational awareness tool. For another, watching where the algorithm was flagging CI ahead of the warning area he was responsible for queued him in to watch for radar echoes. There were also some statements indicating that the jury was still out. Overall, a broad spectrum of answers yet not overall negative, especially after playing with both tools for only one day. I imagine that in the next day or two, more definite opinions will be formed and shared...
Then again, isn't that the point?
Unfortunately, Mother Nature did not cooperate and the area stayed pretty well capped into the forecast period. In the short time I was away, the decision was made to shift the focus area to Bismarck, ND, where we had a better chance of seeing some storms. Shortly after my return, some storms popped, giving the forecasters something to look at, analyze, and even issue some statements and a warning on.
So how did SATCAST do? Well, we got the convection in North Dakota as the system moved east, approaching out original focus area. It did catch the 3 cells that caused the increased activity, with lead-times of about 30 minutes as well as some other areas of convection as the afternoon progressed.
The area of clouds (marked by the green arrow) consistently flagged CI in subsequent satellite images. These storms started to show up on radar at 1958 UTC and went on to be warned on as early as 2330 UTC.
Unfortunately, SATCAST also indicated CI sporadically across the northern part of Minnesota that never had a chance of getting to verify. The capped environment just couldn't be pushed through by the rapidly growing systems that tended to pop up along the boundary. End result, some False Alarms.
A known problem area of the techniques includes high CAPE areas that are capped. The algorithm exploits the satellite signals to flag rapidly growing clouds. Although there are post-processing techniques in place to help eliminate false alarms (such as masking out areas that have no CAPE) it doesn't pick up as well in areas where significant CIN is also present. We as a research group are still hunting for the "best" way to reduce the false alarms without masking out the good data. Also compounding the problem in this area, was the presence of high level cirrus associated with the low-pressure system. The cloud-typing algorithm does a good job identifying ares that need to be masked out, yet high level cirrus seems to be the bane of both CI algorithms tested during this experiment, as evidenced by the CI flags along the boundaries of the system as the afternoon progressed.
In contrast, the UWCI algorithm did not have as many CI flags in the Grand Forks/Duluth focus areas. The one or two flags they did have, verified shortly thereafter. And yes, the UAH SATCAST algorithm did pick up on those cases as well. To be fair, I did not get a chance to see if UWCI flagged the storms in the Bismarck area later, so I am only commenting on the 2-3 hours over the first focus area. But it does bring up an interesting question, is it better to have a few indicators that verify at the cost of missing many of the other cases, or is it better to have more indicators in/of the environment, that although will catch the definite cases of CI, but will also result in higher false alarms?
I asked each of the forecasters and got a few different answers. For one, watching where SATCAST flagged rapidly growing clouds caused him to focus in on the associated environment to ensure a good understand of what is occurring at that time and check for agreement with previous thinking. If he wasn't looking at it, he was checking to see if the environment had changed since he looked last. In short, a situational awareness tool. For another, watching where the algorithm was flagging CI ahead of the warning area he was responsible for queued him in to watch for radar echoes. There were also some statements indicating that the jury was still out. Overall, a broad spectrum of answers yet not overall negative, especially after playing with both tools for only one day. I imagine that in the next day or two, more definite opinions will be formed and shared...
Then again, isn't that the point?
Nearcast training for severe
This morning UW-CIMSS scientists Ralph Petersen and Bob Aune sat down with the EFP's severe desk to help train the participants on how to use the Nearcast product during morning forecast operations. UW-CIMSS provides us with 10 fields from the Nearcast product, which includes individual layers of theta-e and preciptable water (PW), which are then differenced to provide the differential theta-e and PW water fields that we typically use for forecasts. The most effective way to help forecasters understand what they are looking at in these differential fields is to start with the individual layer PW and theta-e fields and then the forecaster can mentally calculate the differences and compare to what the product is showing them. Ralph and Bob helped the participants understand how the Nearcast can assist them in making their forecasts by pointing out that the Nearcast fields will show you where relatively convectively stable and unstable areas are. Because no forcing mechanisms are included within the Nearcast's output, it doesn't guarantee where convection will occur, but it can help narrow down where you should be focusing your attention. If there is an area of strongly stable air, you're not likely to have any deep convection, even if there is some sort of light forcing present. This may be especially useful within SPC operations for forecasters issuing mesoscale discussions (MD) to determine where a severe thunderstorm or tornado watch will or will not be needed. It may also be useful during the early afternoon convective outlook updates to help trim areas that will not be expected to have thunderstorms later during that day. Following the group training, Steve Weiss (SPC SOO) asked if I would be willing to work with the forecasters in operations this summer in exposing this product to them. In addition, we expect to provide a training session within the SPC's bi-annual forecaster training this fall to help expose all of the SPC personnel to the product.
Tuesday, June 7, 2011
Impressions of a first day...
Sitting through the various discussions on the likelihood of the development of CI, it quickly became apparent that the low pressure system moving eastward along the Canadian border was garnering plenty of interest from the 3 main desks located in the HWT. Watching it through the morning with the morning shift of EWP forecasters, the area remained on the "it's going to be an interesting day" and "it didn't really produce much" line, due to the favorable conditions in moisture and afternoon heating being held back by a pretty definite cap that may or may not break sometime before 00Z. Moving in the afternoon, warning portion of the EWP program, conditions remain decent, yet the cap is still firmly in place. Interest in the CI products is in the forefront of the forecaster's minds and the handy, four panel, product is up on the main screens of the AWIPS displays of all four forecasters present.
In all, so far, SATCAST has been indicating CI sporadically in and around the North Plains for most of the afternoon. Taking into account the capped environment, the algorithm has been picking up the signal of rapidly growing clouds, as it is trained to do, yet convection has been slow to develop in CWA's of Grand Forks and Duluth. This has resulted in some red objects (in the area of interest) that will likely not verify, due to the capped environment and some interfering cirrus atop the low pressure system as it moves east.
So far, the forecasters have still indicated an interest in the product as the afternoon has progressed. Being able to look at the data in a loop has allowed for an awareness of the environment, allowing for the forecaster to be "queued in" to an area of interest. As an example, the UAH CI algorithm has indicated a consistent CI signal moving across western North Dakota that verified about 35 minutes later on radar (according to the notes of forecasters assigned to the Grand Forks CWA).
It is still early, we will see how things evolve after a quick dinner break..
Nearcast in short-term warning operations
EWP forecasters have come up with some novel ways to utilize the Nearcast differential precipitable water/theta-e forecasts within their warning operations. Generally we think of the Nearcast as a forecast tool, but the EWP forecasters have been using it to issue warnings and severe weather statements (see image above). Forecasters have been analyzing increased near-storm convective instability as depicted in the 0-2 hour forecasts from the Nearcast product to help increase their confidence that the storm will intensify. Conversely, the Nearcast is very useful to show when storms will die if the storm is expected to move into an area with very low convective instability or dryer atmospheric columns. This is another example of how forecasters will find some additional utility of these experimental products that the developers may not have originally intended.
Monday, June 6, 2011
Final week of the experiment...
Today marks the first day of the 5th and final week of this year's Spring Experiment. This week the EWP rejoins the EFP from its week off and resumes normal activities. Today the EWP will be providing forecasters with a load of training material as they prepare to use the experimental products during real-time forecast/warning operations. Following the 4-hour PowerPoint session we will have the forecasters work through a WES case to get familiar with where the products are, what they look like and how to use them within their AWIPS workstations.
Subscribe to:
Posts (Atom)