A Brief Look at Rapid Environmental Changes in a HSLC Case (January 29-30, 2013)

Note: A few figures are provided in the following PowerPoint: Figures

One of the goals of the current CSTAR project is to study the environmental evolution that occurs during HSLC events.  Dr. Parker and I have been primarily focused on the changes in the synoptic-to-mesoscale environment in the ~6 hours leading up to convection. In order to study a handful of HSLC events, we have performed simulations with the Advanced Research Weather Research and Forecasting (WRF-ARW) model using NAM analyses as initial and boundary conditions. These simulations are run at 3 km horizontal grid spacing, and data is output every 5 minutes.  This allows us to examine the evolution of the simulated synoptic and mesoscale environments on relatively small time and spatial scales.

Animations of simulated composite reflectivity and surface equivalent potential temperature (see PowerPoint link) inform us that these environments can change very rapidly, and that the associated boundaries (e.g., cold fronts, outflow) can be quite extreme. In the particular case shown from January 29-30, 2013, an intense outflow boundary races ahead of a surface cold front, triggering significant convection in Tennessee and Kentucky regardless of CAPE values less than 500 J/kg. An interesting thing to note is the increase of surface equivalent potential temperature just ahead of the outflow boundary as it approaches; this occurs from 0500 to 1000 UTC and therefore cannot be attributed to diurnal heating. Our goal is to clarify how the environment is changing in the few hours prior to severe convection.

In order to determine how and why CAPE might be changing, we took 6 hour boundary-relative time series of surface based CAPE, surface potential temperature, surface mixing ratio, and lapse rate at several points ahead of the outflow boundary and took an average (see plots in PowerPoint link). Increases in all variables are evident, though some increase more than others.  This is common among the several cases we have simulated thus far.

The question we are targeting to answer here is how changes in each of these variables individually affect CAPE.  The next step is to determine how much of the increases in surface-based CAPE can be attributed to increases in solely surface temperature and/or surface moisture.  We have made some progress in answering this question and an update will be provided soon.

Posted in CSTAR, High Shear Low Cape Severe Wx | Tagged , | 3 Comments

Impressive GOES-14 Super Rapid Scan Imagery Over the Carolinas on 09 June 2015

Fairly widespread convection developed across the Carolinas on 09 June 23015 resulting in numerous reports of wind damage and flash flooding (link to SPC storm reports). A fantastic animation of GOES-14 Super Rapid Scan imagery provided by NOAA/RAMMB/CIRA shows the evolution of the convection below. The rapid development of the convection and the explosive updrafts are easily seen.

GOES SRSOR visible satellite loop from 1730-2359 UTC on 09 June 2015

GOES SRSOR visible satellite loop from 1730-2359 UTC on 09 June 2015

Posted in Convection, Satellite | Leave a comment

GOES-14 will be in Super Rapid Scan Operations with imagery over the Carolina’s and Virginia’s Today (5/21)

GOES-14 Super Rapid Scan visible satellite imagery from 1336 UTC on 21 May viewed via the CIRA web site.

GOES-14 Super Rapid Scan visible satellite imagery from 1336 UTC on 21 May viewed via the CIRA web site.

GOES-14 Super Rapid Scan Operations for GOES-R (SRSOR) began on 14 May and will continue for through 12 June, 2015. Super Rapid Scan Operations (SRSO) will provide 1-minute imagery to support multiple research and GOES-R/S user readiness activities. The SRSO domain is usually selected a day or two in advance. The domain schedule along with selected imagery from prior days is available at: http://cimss.ssec.wisc.edu/goes/srsor2015/GOES-14_SRSOR.html#sched_and_movies  Additional background information including training and links to online imagery is available at: http://cimss.ssec.wisc.edu/goes/srsor2015/GOES-14_SRSOR.html

This will be a great opportunity to view the data over our region. NWS forecasters will be able to view some of this data in real-time in AWIPS.

Imagery including visible, infrared, and water vapor is available on the web at the links below…

Posted in Satellite | Leave a comment

NWA Journal of Meteorology article examines utility of total lightning data in weak shear Appalachian storms

A collaborative effort between VA Tech and the NWS Blacksburg office, funded by the GOES-R program, studied the potential utility of total lightning data (from Earth Networks Inc) in weak shear storms over the Central Appalachian region.   Recently, a summary of this work was published in the NWA Journal of Operational Meteorology.  VT graduate student Paul Miller, now a PhD candidate at the University of Georgia, is the lead author. The title and abstract are below, and you can find the article here:

http://www.nwas.org/jom/abstracts/2015/2015-JOM8/abstract.php

Essentially, while there is certainly promise in using total flash rates and trends in diagnosing the severe potential for these weak shear (often known as “single cell” or “pulse” storms), a 2-sigma lightning “jump” (as defined in previous research) proved not to be a practical algorithm due mainly to a very high false alarm.  Differences in how storms are defined (i.e., radar vs. lightning clusters), the actual detection network, geographic region, as well as environment and storm mode, may all determine the effectiveness of any future algorithm designed to use flash counts and trends to better anticipate severe potential.

Fig_5

Article title and abstract:

Single-cell Thunderstorm Severity: Examples from the Central Appalachians Region

Paul W. Miller1, Andrew W. Ellis2, and Stephen J. Keighton3

1University of Georgia, Athens, Georgia
2Virginia Tech, Blacksburg, Virginia
3NOAA/NWS, Blacksburg, Virginia

Abstract

The performance of a total lightning jump algorithm for guiding severe thunderstorm warnings within a weakly sheared environment was investigated using data from the Earth Networks Total Lightning Network. Total lightning observations from two summers for a study domain within the central Appalachian Mountains region were clustered into likely thunderstorms using single-linkage clustering. The spatial and temporal characteristics of each flash cluster were evaluated and used to assign a “storm index” (SI) score to each cluster. Small, short-lived, slow-moving, circular clusters—consistent with single-cell thunderstorms—were given large SI scores, and large, long-lived, fast-moving, linear clusters—inconsistent with the single-cell mode—received smaller SI scores. Statistical testing revealed that days with a simple majority of lightning-defined (LD) single-cell storms possessed significantly weaker 0–6-km wind shear than days with a majority of non-single-cell storms. After classifying 470 clusters as either LD single-cell or multicell/supercell, the 2σ lightning jump algorithm was applied to the flashes associated with each cluster. Total lightning jumps identified by the algorithm were aligned with severe weather report data to evaluate the accuracy of the algorithm. Although probability of detection values for both categories compared well to previous studies, false alarm rates were significantly larger than previously documented. The algorithm performed unsatisfactorily among the LD single-cell and multicell/supercell storms studied, and its performance deteriorated further when applied to a subset of storms most clearly defined as single-cell. However, severe LD storms demonstrated greater flash rates, a promising characteristic for future lightning-based warning tools.

 

Posted in Lightning | Tagged , , | 1 Comment

Examining Gust Factors at the Land-Water Interface More Closely

A recent CSTAR project with NC State and over a half dozen WFOs in the Southeast, examined ways to improve inland wind and wind gust forecasts associated with tropical cyclones. The CSTAR project produced three primary improvements including a bias correction of the TCM wind vortex, using collaborated wind reductions over land, and using collaborated wind gust factors for wind gust grids across the domain.

Fig. 1. Example wind gust forecast as viewed in GFE showing wind gusts at land locations in southeast VA are greater than the wind gusts at adjacent marine areas.

Fig. 1. Example wind gust forecast as viewed in GFE showing wind gusts at land locations in southeast VA are greater than the wind gusts at adjacent marine areas.

Feedback from project participants noted that at times, the use of gust factors near the land-water interface can produce undesirable wind gusts in some scenarios. In an example described in this previous blog post (Examples of Gust Factor at Water-Land Interface), wind gusts at land locations in southeast VA are greater than the adjacent marine areas.  It is worth noting that the processes governing wind gusts at the land-water interface can be quite complicated and occur on very small spatial scales, including those smaller than current forecast grid lengths in GFE. Still, project participants desired a more seamless transition between the land and marine wind gust values.

Fig. 2. The location of the 9 oceanfront locations examined and a table noting the approximate location of the observational platform with the closest surf zone.

Fig. 2. The location of the 9 oceanfront locations examined and a table noting the approximate location of the observational platform with the closest surf zone.

compare.table

Fig. 3. A table comparing data from the 5 locations examined including all locations, non-oceanfront (inland) locations, oceanfront locations, Hatteras, NC (KHSE), and all marine locations.

In order to examine this, NC State student Victoria Oliva, examined the sustained winds, wind gusts, and GFs for 15 tropical cyclones that impacted the Carolinas, Virginia and Maryland. Routine hourly METAR observations with sustained wind speeds of 10 kts or more were used to calculate the hourly GF for each location.  For land locations, the METAR locations varied for each storm and were selected to capture the variations in the wind field with a total of 13,121 GFs computed. In order to examine the land-water interface more closely, we examined gust factors at 9 locations in which the METAR is located in close proximity to the coast, specifically within at least 2 miles from the surf zone. We labeled these locations as “oceanfront.” The METARs included in the oceanfront data set include KCRE, KFFA, KHSE, KMQI, KMRH, KMYR, KNBT, KNJM, and KSUT and their locations are shown in Figure 2. A total of 2,289 GFs from the 9 locations for the 15 storms were examined. In addition, the Hatteras, NC observation (KHSE) was singularly examined as it is located on an island just 0.2 miles from the surf zone. A total of 488 GFs were computed from 15 storms that impacted KHSE. Finally, GF were computed for more than two dozen buoys with an anemometer height of 5 meters that were impacted by the 15 storms. Hourly marine observations with wind speeds of 10 knots or more and wave heights less than 5 meters were used to compute 3,026 gust factors.

Fig. 4. Regression curves for the various locations examined including all locations, non-oceanfront (inland) locations, oceanfront locations, Hatteras, NC (KHSE), and all marine locations.

Fig. 4. Regression curves for the various locations examined including all locations, non-oceanfront (inland) locations, oceanfront locations, Hatteras, NC (KHSE), and all marine locations.

A table comparing the gust factors from the various locations is shown above in Figure 3. The locations include: all gust factors, non-oceanfront (inland) locations, oceanfront locations, Hatteras, NC (KHSE), and all marine locations. The results were surprising with the oceanfront gust factors having a slightly larger average and larger mean value than the non-oceanfront land locations. We had expected the oceanfront gust factors to be considerably smaller than the non-oceanfront or inland gust factors.  The KHSE average and mean gust factors were slightly lower than the average and mean oceanfront gust factor. Still, we expected the KHSE gust factor would be much lower.  The KHSE average gust factor of 1.50 was still significantly higher than the average marine gust factor average of 1.23. A plot of regression curves from the various locations is shown in Figure 4. Note the similarity in the curves for all of the land locations, whether inland, oceanfront, or KHSE which is located on an island. The land locations differ considerably with the gusts factors for the marine locations shown in the blue curve.

It was hoped that the gust factor values for the oceanfront locations would show a transition from the non-oceanfront or inland values where gust factors average 1.53 to the much lower marine values that average 1.23. Since the oceanfront observations were unable to capture a gradual change between the land and marine gust factors, some sort of blended approach could provide the transition that forecaster’s desire.
Additional investigation of oceanfront gust factor based on wind direction would be instructive. In addition, examining observations right at the shore and near the dunes might provide the expected transition. It is worth noting however, that KHSE is located on an island and is well offshore from the mainland, and that wind from just about any direction would provide a long fetch marine exposure.

Posted in Uncategorized | Leave a comment

Another Early Season Tropical Cyclone on which to Utilize CSTAR Based TC Wind Tools

Fig. 1. NDFD wind gust forecast valid 8PM EDT Sun 10 May, 2015.

Fig. 1. NDFD wind gust forecast valid 8PM EDT Sun 10 May, 2015.

The early arrival of Subtropical Storm Ana will provide Southeast WFOs with another opportunity to use and evaluate a new technique to forecast tropical cyclone winds and wind gusts. A Collaborative Science, Technology, and Applied Research (CSTAR) project with North Carolina State University and over a half dozen WFOs in the Southeast, examined ways to add science and improve inland wind and wind gust forecasts associated with tropical cyclones. The CSTAR TC wind technique consists of three primary improvements including a bias correction of the TCM wind vortex, using collaborated wind reductions over land, and using collaborated wind gust factors for wind gust grids across the domain. This summer, the foot print of the project will expand to WFOs Wakefield and Sterling ,VA and Columbia, SC.  An example of the Wind Reduction Factor and Wind Gust Factor grids for 8PM EDT Sun 10 May, 2015 are shown below in Figure 2.

Fig. 2. CSTAR TC wind technique Wind Reduction Factor (left) and Wind Gust Factor (right) grids, valid 8PM EDT Sun 10 May, 2015 as viewed in GFE.

Fig. 2. CSTAR TC wind technique Wind Reduction Factor (left) and Wind Gust Factor (right) grids, valid 8PM EDT Sun 10 May, 2015 as viewed in GFE.

Posted in TC Inland and Marine Winds | Leave a comment

Two New Convection Allowing WRF Ensembles Provide Unique and Useful Guidance

Forecasters have recently gained access to two new convection allowing (CAM) WRF ensemble model forecasts. One ensemble is provide by NSSL and is a nine-member, WRF-ARW ensemble initialized daily at 00 UTC  while the other is a ten-member, WRF-ARW ensemble initialized daily at 00 UTC and is provided by NCAR/MMM. These two modeling systems appear to be one of the first, regular and widely available convection allowing (CAM) ensembles.

CAMs have proven to be a great utility for forecasters as they can provide information about features smaller than those resolved by coarser resolution models, and in addition, they can predict convective mode, convective system propagation, diurnal cycle, and other characteristics of convection far better than models run with a convective parametrization. More information on CAMs is available at this reference.

In addition, the web sites in which these two modeling systems are available contain an easy to use graphical interface, the ability to zoom over regions, and access to several unique fields. Some of the fields include hourly max and updraft helicity probabilities, composite convective parameters, simulated satellite imagery, and ensemble max fields.

While these modeling systems were not likely created to provide guidance for tropical or sub-tropical weather systems, the developing storm system near the Bahamas provides an interesting example to view one set of output, the 1-km or composite reflectivity products in figures 1 and 2.

Fig 1. 0000 UTC 06 May 2015 NCAR/MMM WRF-ARW ensemble 36-hour 1-km reflectivity forecast valid 12 UTC 07 May 2015.

Fig 1. 0000 UTC 06 May 2015 NCAR/MMM WRF-ARW ensemble 36-hour 1-km reflectivity forecast valid 12 UTC 07 May 2015.

Fig 2. 0000 UTC 06 May 2015 NSSL WRF-ARW ensemble 36-hour composite reflectivity forecast valid 12 UTC 07 May 2015.

Fig 2. 0000 UTC 06 May 2015 NSSL WRF-ARW ensemble 36-hour composite reflectivity forecast valid 12 UTC 07 May 2015.


Nine-member NSSL WRF-ARW Ensemble initialized daily at 0000 UTC
http://www.nssl.noaa.gov/wrf/newsite/

The 9 ensemble members utilize the WRF-ARW V3.4.1 at 4-km grid length with similar configurations but varied initial conditions. The members are comprised of the regular NSSL-WRF, which uses the 0000 UTC initialized NAM for ICs and LBCs, one member that uses the 0000 UTC initialized GFS for ICs and LBCs, and 7 members that use different members of NCEP’s 2100 UTC initialized SREF system for ICs and LBCs. The SREF system ICs/LBCs include 3 WRF-ARW members (the control member and two perturbed members), 2 NMM members (the control and one perturbed member), and 3 NMMB members (the control and two perturbed members). The domain and physics parametrizations for each NSSL-WRF ensemble member are identical to the regular NSSL-WRF and include MYJ BL/turbulence parametrization, WSM6 microphysics, RRTM longwave radiation, Dudhia shortwave radiation, and
Noah land-surface model.

Ten-member NCAR/MMM WRF-ARW Ensemble initialized daily at 0000 UTC
http://ensemble.ucar.edu

The 10 ensemble members utilize the WRF-ARW V3.6.1 at 3-km grid length with similar configurations but varied initial conditions. Initial conditions provided by down scaled members of 0000 UTC WRF/DART EAKF analyses with perturbed lateral boundary conditions originating from GFS forecasts. The domain and physics parametrizations for each NCAR/MMM WRF ensemble member are the same and include MYJ BL/turbulence parametrization, Thompson microphysics, RRTMG longwave radiation, Dudhia shortwave radiation, and Noah land-surface model.

Posted in Convection, NWP | 2 Comments