SuPEr Cell Thunderstorm Research and Analysis (SPECTRA)
SEVERE WEATHER PARAMETERS INFLUENCED BY THE INFLOW OF
SUPERCELL THUNDERSTORMS: A RESEARCH PROPOSAL TO STUDY HOW THE ENVIRONMENT IS INFLUENCED AND
CAN HELP TORNADOGENESIS
Abstract
Mesoscale analysis models used in forecasting have come a long way, but there are many limitations to its accuracy, mainly due to its temporal and spatial resolution within the models. Combined with a sparse network of METARs used for surface observations and an even greater lack of upper air observations, severe weather parameters cannot be accurately measured on the mesoscale where storms are happening. Using custom built instrumentation, Operation Supercell Thunderstorm Research and Analysis (SPECTRA) closely examines the environment in the inflow region of supercells to see how the environment is influenced and how the severe weather parameters are enhanced. It is expected that moisture pooling and strongly veered winds below the mesocyclone create an environment very suitable for tornadogenesis and tornado maintenance far more extreme than what can be seen on the Numerical Weather Prediction (NWP) models. This project examines how much the environments can be enhanced, and to help find and fix biases in the mesoscale analysis.
Introduction
Operation Supercell Thunderstorm Research and Analysis, or SPECTRA, is a joint effort between multiple research teams with the goal of studying supercell thunderstorms and tornadogenesis. SPECTRA is organized by Nicholas Stewart of Midwest Weather Center. The data collected in this experiment include participation by a number of teams with varying instruments and objectives. These separate teams include Jared Stevenson of Project Recon, Chris Heater from Midwest Storm Chasers and Researchers (MIDSCAR) and Warren Clark Causey from The Sirens Project. Each team, while independent, will collaborate and share data collected during the research period. Using different instruments, each team will collect a dataset which will help in the research of tornadoes.
SPECTRA’s mission is to look at the environment around supercells and see how it is influenced by the storm. The goal is to look at, in high temporal and spatial resolution, severe weather parameters right around the mesocyclone of the supercell and see how the parameters are influenced and how they change. Some of these parameters include the lifted condensation level (LCL) heights, the convective available potential energy (CAPE), 0-1km storm relative helicity (SRH) and 0-6km bulk shear just to name a few.
There are many challenges with the project including the weather actually cooperating and providing supercells and tornadoes to study. As this is an in-situ experiment, there is no guarantee of a complete dataset. This is dependent on a storm firing, dropping a tornado in an area with good terrain and a good road network, and the instruments functioning properly and being successfully deployed in the path of the tornado.
Literature Review
Field experiments have been taking place in tornado alley for the last decade, with major experiments, including the second Verification of the Origins in Tornadoes Experiment (VORTEX 2) experiment in 2009 and 2010, happening more frequently. Yet, certain aspects of tornadogenesis are still not known and are still being researched. The VORTEX 2 mission included 24 portable weather stations, 11 mobile mesonets and 16 tornado pods which were used to collect temperature, relative humidity, wind and pressure data at the surface in order to record the thermodynamic and kinematic properties of the environment around supercells (Wurman et al. 2012). On a smaller scale, teams like TWISTEX, which have multiple mobile mesonets, are on a similar mission to collect data and use it to better understand the development and maintenance of tornadoes by sampling the hook echo and rear flank downdraft (RFD) of supercells (Lee and Finley 2012).
SPECTRA, will use aspects from these previous studies to look at the near-storm environment and see how it is influenced by supercells by studying the inflow region of supercells. The main goal is to look at how severe weather parameters in the tornado region of the storm are intensified on a scale not able to be seen by models. For this research proposal, using in-situ instruments, the mechanics for tornado maintenance and near-tornado environment influences will be studied.
Using in-situ weather stations will dramatically improve the spatial resolution of the data collected, as well as improved accuracy of the data collected. The May 22, 2010 Bowdle, South Dakota tornado was in an environment that had very tight gradients of wind and the thermodynamics of the atmosphere were small enough that it would not be seen in the Storm Prediction Center’s (SPC) Mesoscale Analysis (Lee and Finley 2012). Collecting data will also help with model bias, because the CAPE and shear values can be corrected. Model physics may mis-forecast these values which may change the areas of maximum tornado probability (Hamill and Church 2000).
The mesoscale analysis error may also increase as the radiosonde network launched by the National Weather Service (NWS) has large spacing upwards or 350 km, and the timing of the launches is only twice per day, which are not in sync with convective time frames to further complicate the problem (Orlanski 1975). Surface observation sites are spread out with an average gap of 100 km between stations, and most can only record with hourly frequency. The Numerical Weather Prediction (NWP) models are used to fill in this large time gap and to depict the atmosphere at these times (Coniglio 2011). One of the many side projects of the VORTEX 2 mission was to compare the performance of the NWP models and the actual observed information from the VORTEX 2 instruments, specifically the Rapid Update Cycle (RUC) model (Coniglio 2011). The VORTEX 2 project showed a low bias in the thermodynamic environment of the Planetary Boundary Layer (PBL), and a moist bias in the upper atmosphere. There was also a cool, “significantly higher” moist bias compared to the RUC analysis. Finally, the wind was analyzed to be too fast in the lower 1 km, and too slow in the 2-4 km region of the atmosphere. This leads to a CAPE error, with the RUC tending to overdo the values which can have big impacts on short-term forecasts. Overall in terms of severe weather forecasting, the errors with Convective Inhibition (CIN) and Level of Free Convection (LFC) values were quite significant in the RUC analysis showing model bias (Coniglio 2011).
Having in-situ instruments to collect this data will not only help correct model bias, but it will also provide a better representation of the overall environment far better than mesoscale analysis with the help of faster time resolution of measurements and a much better spatial resolution. This, in return, will help better the understanding of a near-tornado environment and seeing how severe weather parameters are intensified right in the storm on a finer scale than the RUC analysis.
Like in the TWISTEX experiments, the surface data collected in SPECTRA can be applied to the RUC analysis (the upgraded RUC, now called the Rapid Refresh, or RAP) to help get a more accurate idea of the environment by correcting the temperature and dewpoint values at the surface. This will correct the CAPE and CIN by redrawing the parcel ascent with the data collected with SPECTRA instruments (Lee and Finley 2012).
Large values of potential buoyancy were in place in the near-tornado environment on May 22, 2010. TWISTEX mobile mesonets, using RUC analysis for the upper levels, measured CAPE values exceeding 3000 j/kg surrounding the tornado and values over 4000 j/kg in the path of the tornado (Lee and Finley 2012). Another big finding was the CIN values being the lowest right in the path of the tornado below 100 j/kg. These numbers were influenced by the inflow of the supercell and not picked up on fully by mesoscale analysis as the gradients were too tight and on a small scale (Lee and Finley 2012).
The actual data collection for this project stems from research conducted by the ROTATE program, which was also used by VORTEX 2 in 2009-10. According to ROTATE, the most likely time for tornado occurrence is in the month of May and early June (Wurman et al. 2012). ROTATE, during their 12 years of research, observed tornadoes on only 4.9 days per season – only 1.3 of these days had significant tornadoes. For the project however, 33% of the years (4 out of 12) had no tornado collection (Wurman et al. 2012).
VORTEX 2 instrumentation had a high spatial resolution from an array of instruments geared for various jobs. The Texas Tech University StickNet platforms were deployed with spacing of 1 km to 5 km apart to measure the full lifespan of the supercell (Wurman et al. 2012). The StickNets gathered temperature, humidity and 2 m wind. The tornado pods, primarily owned the Center for Severe Weather Research (CSWR), were more hardened with the goal to record in the tornado itself. These also measured temperature, humidity and 1 m wind.
All of this instrumentation was meant to record and map out what helped tornado maintenance. During the initial VORTEX mission, it was discovered that storm-induced wind profiles, especially in the low level vertical wind shear, determined the strength of the mesocyclone in the lower levels (Thompson 1998). A balance must be met between the storm-relative winds. If they are too weak, precipitation can wrap around the mesocylone cutting off the inflow and undercutting the storm. If the winds are too strong, most of the precipitation might be removed from the mid-levels preventing the Rear Flank Downdraft (RFD) from developing which is likely a trigger for vertical vorticity near the ground to assist the low-level mesocyclone (Thompson 1998).
A delicate balance must be established between the RFD and supercell inflow to favor the repeated development of tornadoes. Being able to take measurements inside the inflow of the supercell will help measure this balance, and see how strong the inflow must be, and how this would impact the other severe weather parameters in the near-storm environment. Seeing how the vertical velocity is impacted by the baroclinicity generated by the RFD and inflow is important to better understanding tornado maintenance.
The low pressure that forms in the mesocyclone and the tornado will help sustain the storm by increasing the vertical vorticity and the horizontal convergence. The faster inflow winds get unbalanced by the friction of the surface of the earth increasing the imbalance of the pressure gradient and the centrifugal forces (Dowell and Bluestein 2002). This convergence would also help in cloud formation flowing into the supercell. Feeder clouds forming due to the convergence were shown to be a sign of rapid supercell growth (Mazur, Weaver and Haar 2009).
Tornado dissipation begins when the inflow of warm, moist air into the supercell gets interrupted or the inflow begins to weaken. Cooler and drier air flowing into the supercell lessens the buoyancy, thus causing tornado dissipation. Cooler and precipitation loaded air (more dense) would also impact the tornado maintenance by increasing the LFC beyond what can be reached by the surface temperature (Dowell and Bluestein 2002). These small mechanisms would be too small to see in the NWP models.
Gathering in-situ measurements through this SPECTRA project could help forecast storms that can develop tornadoes. Using a variety of instrumentation placed in the supercell inflow, this could greatly benefit the understanding of a supercell and how it impacts the near-storm environment. Knowing the true impacts on the environment and seeing real-world measurements and seeing how they compare to the NWP models and mesoscale analysis would help to fine-tune the bias and possibly correct models. This project may also be able to pick up on smaller features that would intensify severe weather parameters on a very small scale.
Based on previous research, supercells are expected to have a large impact on the surrounding environment that the storm will move into. This includes enhancing severe weather parameters like CAPE, LCL heights and 0-1 km SRH. This research proposal will cover the background research for the study, the study area for this experiment, the description of the data that will be collected, a detailed methodology for this experiment, expected results, a proposed budget for this experiment and a timeline for the research project.
Study Area
The SPECTRA project is dependent on good terrain, good road networks and areas where severe weather and tornadoes are a likely occurrence. The study area picked for this project is the central United States from the Canadian Border south to the Texas panhandle. It is in this area where all the criteria is met for a successful deployment.
The study area for this project will be in the central Midwest, commonly known as tornado alley. Using a similar area that VORTEX 2 used in 2009-10 as stated by Wurman et al. (2012), this area covers the Great Plains from the Dakotas down to southwestern Texas (Figure 1).
Figure 1: Study area for SPECTRA, which includes the central Midwest where the terrain and road network is best for experiment deployment.
It is in this area of the central United States where the tornado frequency is the highest and the terrain is the best for conducting research. The road network, especially in the northern half of this domain, is also an important factor for this research to happen easily.
In the study area shown, and in areas in close proximity to the shaded region, the terrain is mostly flat allowing easy viewing. The flatter terrain, with a few exceptions typically in river valleys, also allows the road network to be much more cooperative for this experiment. The total landmass across the Midwest that could be utilized for this project includes 1.3 million km2. The small size of this project, only four vehicles at maximum, makes it easy to travel large distances in small amounts of time.
The time period for this research, using previous research by Wurman et al. (2012) mentioned in the literature review, puts the best time for researching tornadic supercells from mid-May through mid-June. During this time, ROTATE averaged 4.9 days per season with tornadoes, and 1.3 days with significant tornadoes (Wurman et al. 2012). According to the SPC, tornado probabilities across the southern Midwest are especially higher mid-May through the mid-June timeframe (Figure 2).
Figure 2: Using the SPC climatology data, this is the tornado probabilities during the SPECTRA research timeframe in late May and early June.
Overall, SPECTRA will deploy across the central United States where the terrain and road networks are suitable for deployments. It is across this region where there is also the best likelihood for severe weather based on climatology and previous research by the CSWR. Using this climatology and past research, the timeframe for this project is set for late May into early June where the potential for tornadic thunderstorms is the highest.
Data Description
The data collected for this experiment requires a lot of hands-on work out in the field to actually collect the meteorological parameters that may impact tornadogenesis and tornado maintenance. This will be discussed more in the methodology section of this proposal. However, the custom built instrumentation will collect surface data that we need to assess the state of the atmosphere, and this would be applied to the RAP analysis to complete the full atmosphere. Again, more of this will be discussed in the methodology. This data would give insight into the environment that supercells are moving into, and be able see how the atmosphere is actually being modified by the supercell.
The Arduino based weather instrumentation SPECTRA currently operates includes three stations; Helen Hunt, Bill Paxton and Phillip Seymour Hoffman as an homage to the movie “Twister.” These small instruments measure the temperature, relative humidity and pressure. These instruments have been recently improved to also collect wind speed and direction with attached anemometers and wind vanes, as well as a rain gauge. These weather parameters will be collected and analyzed to generate more accurate information than the RAP analysis at a much higher spatial and temporal resolution. The specifications of these weather instruments can be found in Table 1.
Table 1: Instrument specification for the sensors on the Arduino Weather Shield.
These Arduino based instruments will be installed on a custom built platform called the Rapid Deployment Weather Instrument (RADWIN) as showed in Figure 3. Powered by a 12,000 mAh battery and housed inside a Davis radiation shield, the RADWIN will collect 1 m temperature, humidity and pressure, and 1.5 m wind speed and direction at a temporal resolution of 1 reading per second.
Figure 3: The Rapid Deployable Weather Instrument (RADWIN) diagram. The instruments are capable of measuring 1m temperature, humidity and pressure, as well as 1.5m wind speed and direction. An attach rain gauge will also allow measurement of precipitation.
SPECTRA will also operate two mobile mesonet stations capable of measuring the temperature, humidity, dewpoint, pressure, wind speed and direction and the total rainfall. The instruments are mounted on the roof of the vehicles with radiation shields at a height of .25m above the vehicle to help eliminate any heat that may be added to the air by the vehicle exhaust. This should be negligible, especially if the vehicle is in motion as the wind generated by the movement will help mix the air in the radiation shield.
The mobile mesonet weather instrumentation will also have GPS logging information. The latitude, longitude, elevation and speed are all logged. This helps for two reasons; for one the latitude and longitude will help plot the data collected on a map. Also, elevation data will help calculate the mean sea-level pressure (MSLP). Finally the speed and direction of travel collected by the GPS can be subtracted from the wind measured by the mobile anemometer to get a true wind in the environment. This will be tested more in depth to see if this is accurate enough. Otherwise, the wind measurements will only be taken when stationary.
Data can be collected with fairly high spatial and temporal resolution. The two mobile mesonets can cover large distances in short periods of time. The mobile mesonets will allow constant data collection at one measurement per minute. Three RADWINs can be placed in the field for a long duration. These RADWINs will sample the pre-storm environment, the inflow to the supercell, the outflow of the supercell and finally the post-storm environment. The RADWINs will be collecting data once per second for time periods up to four hours limited by the battery. Two tornado pods are also going to be utilized for most research days which will be used in similar ways as the RADWINs, but with the goal of getting inside the tornado itself to measure the wind speed and direction, temperature, pressure and humidity there. The instruments are built to measure at once per second, but the goal is to upgrade them to measure at a tenth of a second to better measure the rapidly changing conditions inside the tornado. These measurements will deliver much higher temporal resolution over the NWS METAR network which records data either every 20 minutes. With all this instrumentation in the field, it is estimated to have a spatial resolution close to 4 km, much better than the 13 km RAP analysis and 100 km METAR network operated by the NWS.
The RADWINs and tornado pods, as mentioned, can collect humidity and pressure data. From this, dewpoint and MSLP can be calculated during data analysis. The MSLP will be calculated with the help of the GPS device on each instrument.
From the many instruments used in this experiment, this will help even further create a more accurate surface representation of the near-storm environment. In particular, the dewpoint will help find the LCL height, a critical parameter which will be examined in this experiment. The LCL heights may be heavily influenced by the supercell inflow as moisture pooling may occur right in the inflow notch. The moisture pooling would help lessen the temperature/dewpoint spread thus allowing LCL heights to fall. This will be discussed more in the expected results section of this proposal.
Having the temperature and dewpoint information, this can be input into RAP analysis data to calculate the true CAPE values at a high resolution. This will be talked about more in the methodology section of this proposal. CAPE values are another parameter that will be closely examined seeing how this might be heavily influenced by the supercell inflow, again which will be discussed more in the expected results section of this proposal.
With the anemometers attached to the weather stations, this should help calculate true wind speed and direction at the surface on a scale too small for RAP analysis to see. Wind shifts inside the inflow region of the supercell enhanced by the low pressure under the mesocyclone would intensify speed and directional shear values. This would change bulk shear and helicity values, which are likely enhanced by supercell inflow, which will be discussed more in the methodology and expected results section of this proposal.
The data collected in this experiment will be done through multiple instruments, all of which have varying roles in the grand scheme of the project. These instruments will be able to measure data to calculate severe weather parameters on a much higher spatial and temporal scale than the METAR network and the NWP models. This project has the potential to measure certain parameters around the supercell at one measurement per second. The multitude of instruments will allow for a spatial resolution close to 4 km when the RADWINs and tornado pods are deployed. In addition to the mobile mesonets, the resolution can be much greater.
Methodology
The SPECTRA project requires an assortment of instruments and strategies to fully complete the research goal. As covered in the data description section, mobile mesonets, RADWINs and tornado pods make up the heart of the data collection instruments. Preparations began in April 2015 to develop and test the instrumentation for this experiment. During the field experiment which will take place over a two week timeframe, the data will be collected across the central plains. Once the field experiment stage comes to a close, the data analysis portion will begin which will look through the collected data and begin making findings about the severe weather parameters in the inflow notch of the supercell. The goal is to see how these parameters are influenced by the supercell.
I. Pre-project planning
Starting in the early spring, the dates for the research experiment was set in order to allow everyone to clear their schedule, the equipment to be finished and vehicles ready for deployment. Using research by the CSWR as mentioned in the literature review, as well as climatology data from the SPC, the dates May 23 through June 3 were selected for the project. It’s during this time that tornado activity should be at its peak across the central plains.
With a date set, the focus shifts to the instrumentation preparation. This experiment involves an array of instruments measuring a variety of meteorological parameters. At the heart of the experiment are the RADWIN instruments. Each instrument is assembled using Arduino components. The instrument, as shown in Figure 3, has two main parts; the base and the weather instrument package.
The construction of the RADWIN is one of the first steps of the SPECTRA project. A detailed look at the RADWIN was covered in the Data Description portion of this proposal.
Building each instrument requires soldering the SpearkFun Weather Shield to the Arduino Uno board. Two RJ11 jacks also need to be soldered to the Weather Shield board. These are the inputs for the rain gauge and the anemometer. Then, the MicroSD datalogger is wired to the Weather Shield. The GP-635T GPS receiver is then plugged into the serial interface on the Weather Shield. Finally, a power source is plugged in and the RADWIN instrument is active.
Using the Arduino software, the code written for this instrument is imported to the Arduino instrument. This code allows for one second data collection of the temperature, humidity, pressure, wind speed and precipitation accumulation. Real-time data can also be monitored on a computer using Tera Term. This gives the readout of the data stored on the microSD card.
Testing of the RADWINs take place during the months of April and early May to make sure they function as they are supposed to. It’s also during April when practice deployments take place fine-tuning the deployment process. Early season setups are the perfect time to make sure everything works the way it is supposed to when the situations are not as stressful.
In addition to the testing of the experiments, they will also be calibrated correctly making sure they have little to no error during the experiment stage. Calibration would involve taking one instrument and measuring known variables. Submerging the instrument in ice water will find the drift away from 32° Fahrenheit, and then taking it and putting it in boiling water to find the error from 220° Fahrenheit. A calibration would be applied to the Arduino code, either an addition function for a linear error, or a more complicated function for non-linear error.
Alongside the RADWIN instruments, Davis Vantage Vue weather instrument are also installed to both vehicles involved in the experiment. Attached using a custom made PCV pipe frame, each instrument is mounted to the roof rack of the vehicles. The data collected via these instruments is wirelessly transmitted to a console inside each vehicle. This data can be downloaded to a computer via a USB connection to the console. This data will be used in conjunction with another Arduino instrument inside each vehicle. The Arduino instrument will record the pressure and GPS data, while the Vantage Vue records the wind, temperature, humidity and dewpoint data. The Arduino instrument GPS data recorded will allow for latitude and longitude coordinates of the measured variables. Also, the vehicle’s movement, recorded via the Arduino board, will be subtracted from the wind speed measured by the Vantage Vue which will remove the vehicle’s movement from the recorded wind using vector subtraction. This will give a true wind of what is happening on the surface.
II. The Field Experiment
The field experiment is a week and a half experiment through the central plains. On May 23, the SPECTRA team will depart from Macomb, Illinois toward Hays, Kansas, the base of operations in the plains. May 23 is scheduled to be a travel day. Once in Hays, the separate teams will meet. The two vehicles, Scout 1 from Macomb, and Probe 1 from Marietta, Georgia, will then start forecasting for the experiment. Probe 2 operated by MIDSCAR will join on days they are out in the field. The Sirens Project will also join if we’re on the same storm. As mentioned in the introduction, MIDSCAR and The Sirens Project are independent teams only collaborating if on the same storm as the SPECTRA project.
May 24 will start like every other day of the experiment. A morning meeting will be conducted in the hotel conference/business room if offered, otherwise one of the team’s rooms will suffice. The team will go over the forecast models including the Global Forecast System (GFS), North American Model (NAM), Rapid Refresh (RAP) and the High Resolution Rapid Refresh (HRRR). A higher confidence will be placed on the HRRR and the NAM given past experiences with each.
Using the models, a target city will be plotted. Out of the severe weather parameters, the 500mb heights and winds, SBCAPE/MLCAPE, dewpoint values, LCL heights, 0-1km SRH, 850mb wind/direction and the 0-6km bulk shear will be closely looked at. These parameters are very good indicators if supercells are forecast to develop.
Surface observations will also be looked at in the morning. Surface wind speed/direction and dewpoints will be analyzed to look for more suitable areas. Veered surface winds will greatly intensify the hodographs thus increasing the SRH.
Visible satellite imagery will also be analyzed. Two main things will be looked at in the satellite information; clearing over target areas and potential mesoscale boundaries. Clouds clearing out in the morning hours will allow the instability to build. Small mesoscale boundaries are also very important features when forecasting tornadoes. These boundaries, perhaps outflow boundaries from nocturnal thunderstorms the night prior, will help SRH values which will greatly enhance tornado potential if storms track along these boundaries.
Once a target is chosen, the forecast will then be compared to the Storm Prediction Center and to the local Weather Forecast Offices to fine tune the forecast target. After a target is set, the journey begins to the target city.
Once on a storm, and the storm is showing a good chance of producing a tornado, the storm deployment plan goes into effect. As shown in Figure 4, instrumentation deployment is done through two vehicles; Probe 1 and Scout 1. The Scout 1 vehicle will deploy the RADWINs ahead of the supercell’s track. RADWIN 1 is to be placed south of the expected tornado track where it will sample the inflow into the supercell and the RFD of the supercell after it passes. RADWIN 2 is to be placed just north of the expected track again measuring the inflow into the supercell and to experience the environment near the core as the storm passes. RADWIN 3 is to be placed south of the expected track well ahead of the supercell’s position in order to get the inflow into the supercell and experience the RFD of the supercell.
Figure 4: The SPECTRA deployment plan. This is based on ideal conditions. The RADWINs, labeled as Sticknets, are to be deployed around the path of the tornado to measure the pre-storm, storm and post storm environment. Tornado pods will be placed ahead of the tornado.
The network of the RADWINs should provide a nice spatial resolution of the parameters around the storm which will provide a great pre-storm, storm and post-storm environment.
Meanwhile, the Probe 1 vehicle is to deploy once a tornado has fully condensed and it looks to maintain for several minutes and several miles. By judging the track of the tornado, the Probe team will place a tornado pod in front of the advancing tornado. Once deployed, the team will back off of the tornado and begin heading south away from the circulation.
During each team’s route, the roof-mounted instruments will be sampling the environment. Once the instruments are deployed, both the Probe and Scout team moves back toward the tornado. The Probe team would stay in the inflow region of the tornado while the scout team sits on the downdraft side of the tornado as seen in Figure 5. Having two instruments measuring this data simultaneously in two different regions around the tornado will provide a great dataset for the difference in weather parameters on the inflow and outflow side of the storm. Safety is of absolute importance and trumps everything else for this project. This will only be conducted in a safe manner.
Figure 5: Once the RADWINs and tornado pods are deployed, the Scout and Probe teams will move around the tornado in order to measure the environment around the tornado.
During this deployment process, both teams will be in constant contact via radio. Each deployment site will be relayed to the other vehicle so that there is redundancy in the location where each team dropped each instrument.
Once the storm dissipates, the Probe and Scout team will collect the instruments dropped during the deployment. This could be a challenging process if dirt roads are turned to mud, and if the instruments are a long distance from where the storm dissipates. There might be a chance instruments will be left overnight if they cannot be safely retrieved.
Once the instruments are collected, the two teams will meet and discuss the plan for the night. A hotel will be chosen depending on where the following day’s setup lies. A trip to the hotel may take a few minutes if the following day’s setup is close, or it may be a several hour ride to get close to the following day’s target area.
Once at the hotel, the data from all the instruments is collected from the microSD cards and stored on a central laptop. The data is also backup up on a cloud server for redundancy.
III. Post-Experiment Data Analysis
Following the experiment, the data analysis portion of the experiment begins. Archived RAP Analysis data will be downloaded. This RAP analysis will build the upper atmosphere where this experiment cannot measure. This data will be downloaded from the Penn State Bufkit Data Distribution System. Analysis sounding will be collected from areas closest to the storm that was sampled by the instruments. The Bufkit site must be on the inflowside of the storm just prior to initiation to eliminate and convective feedback that may be create.
In addition to the sounding information from the Bufkit data, RAP analysis maps will also be downloaded from the Storm Prediction Center Mesoscale Analysis Archive. Finally, archived radar data will be downloaded from the National Climatic Data Center.
The radar data will be formatted into a GIS file that can be opened in ARCMAP. Once in ARCMAP, the data collected during the experiment will be overlaid on top of the radar data. This includes the RADWIN, mobile mesonets and tornado pod data. The temperature, dewpoint (derived from pressure/temperature/humidity) and pressure data will be laid out and contoured in ARCMAP. Surface wind speed/direction will also be drawn on each mobile mesonet data point, and the average wind speed/direction over the 2-3 minute period the radar image was valid will be placed on each RADWIN.
This data should show small thermal and moisture gradients in high-resolution around the supercell.
Next, parameters will begin to be derived from the RAP analysis data. The surface data points will be input into the RAP analysis sounding (Figure 6) replacing the surface data provided by the sounding. Using Bufkit, the actual LCL heights, CAPE, CIN, 0-6km Bulk Shear and 0-1km SRH will then be calculated and contoured on the radar image. Again, this will show the small features and changes in high-resolution that may not be picked up on by the RAP analysis.
Figure 6: A analysis sounding will be taken from the archived Bufkit data closest to the tornadic storm just prior to initiation. Surface data collected in the experiment will be input into the lowest level of the sounding which will then calculate the true values of the severe weather parameters.
The calculated values will then be compared to the RAP analysis information. The difference in these values inside the inflow region of the supercell will be taken showing just how much these values change right in the inflow of the supercell. Ideally, this will give a picture of how much of an impact the inflow region of the supercell has on the parameters that create severe weather. Also, biases in the RAP analysis should be seen.
IV. Conclusion
The full project from planning to the data analysis stage should allow for a well-rounded and precise experiment to better understand the severe weather parameters right in the vicinity of the supercell and the tornado. Instruments will be well tested and ready for the experiment during the pre-project planning stage. During the field experiment, the SPECTRA team will use the deployment plan to get instruments in and around the tornado and supercell. During the post-experiment data analysis stage, the severe weather parameters will be calculated using the RAP analysis and Bufkit data. This will allow a more accurate representation to be drawn to get a more detailed look at the severe weather parameters the supercell and tornado is moving into.
Budget
The budget for this project is all self-funded through multiple means. The primary source of funding comes from freelance work with Live Storms Media, selling weather footage to media outlets. As shown in Figure 8, this accounts for the largest source of funding for the research project.
Figure 8: The budget for SPECTRA. The entire project including the instrumentation, as well as the fuel and hotel costs for the field experiment, it estimated to be $2,431.69. The project is funded by research awards and freelance weather video, which will pay for almost the entire project.
Another source of funding for the project came from previous awards won through research. These monetary awards were won through a multitude of research presentations. That money is being put right back into future research.
Overall, the cost for one RADWIN is $313.55. SPECTRA currently has two of these completed instruments valued at $627.10. There are two more partial RADWINs each valued at $152.40 apiece, for a total of $304.80. In addition, the mobile mesonet instrument by Davis was already purchased prior to the experiment (Figure 9).
Figure 9: Itemized list for the RADWIN instrument. This includes the quantity for each component and the cost of each component for a single RADWIN.
During the actual experiment, we are estimating total mileage to be around 5,000 miles. Using a fuel calculator provided by FuelEconomy.gov, we estimate total gas usage to be around 200 gallons. To compensate for the rapid change in gas prices in the summer time, we are anticipating gas prices to be $3.00 per gallon, higher than the anticipated forecast. Using the average MPG of the Dodge Journey, the vehicle used for this project, the expected gas expense will come in at $599.79.
In addition to the gas expense, hotels are going to be another expense for the project. The project timeline would call for overnight stays at hotels for 11 nights. The estimated hotel cost should come in at approximately $900.
The total expense for this project is estimated at $2,431.69. This includes the instrumentation, as well as the fuel and hotel expense during the field experiment. Factoring in the funding for the project, there remains only $31.61 left in out of pocket expenses which will be paid for by Midwest Weather Center (Figure 8).
Mobile mesonet deployment on an Iowa supercell.
A closer look at the mesonet with the anemometer propeller removed for long-distance transport.
Mesonet collecting data on a low-precipitation and high-based supercell.
RADWIN collecting data on the Canadian, Texas tornado on May 27, 2015.
References
David C. Dowell and Howard B. Bluestein, 2002: The 8 June 1995 McLean, Texas, Storm. Part II: Cyclic Tornado Formation, Maintenance, and Dissipation. Mon. Wea. Rev., 130, 2649–2670.
Bruce D. Lee, Catherine A. Finley, and Christopher D. Karstens, 2012: The Bowdle, South Dakota, Cyclic Tornadic Supercell of 22 May 2010: Surface Analysis of Rear-Flank Downdraft Evolution and Multiple Internal Surges. Mon. Wea. Rev., 140, 3419–3441.
Richard L. Thompson, Corey M. Mead, and Roger Edwards, 2007: Effective Storm-Relative Helicity and Bulk Shear in Supercell Thunderstorm Environments. Wea. Forecasting, 22, 102–115.
Matthew J. Bunkers, Jeffrey S. Johnson, Lee J. Czepyha, Jason M. Grzywacz, Brian A. Klimowski, and Mark R. Hjelmfelt, 2006: An Observational Examination of Long-Lived Supercells. Part II: Environmental Conditions and Forecasting. Wea. Forecasting, 21, 689–714.
James Marquis, Yvette Richardson, Paul Markowski, David Dowell, Joshua Wurman, Karen Kosiba, Paul Robinson, and Glen Romine, 2014: An Investigation of the Goshen County, Wyoming, Tornadic Supercell of 5 June 2009 Using EnKF Assimilation of Mobile Mesonet and Radar Observations Collected during VORTEX2. Part I: Experiment Design and Verification of the EnKF Analyses. Mon. Wea. Rev., 142, 530–554.
Rebecca J. Mazur, John F. Weaver, and Thomas H. Vonder Haar, 2009: A Preliminary Statistical Study of Correlations between Inflow Feeder Clouds, Supercell or Multicell Thunderstorms, and Severe Weather. Wea. Forecasting, 24, 921–934.
Thompson, R. L., 1998: Eta model storm-relative winds associated with tornadic and non-tornadic supercells. Wea. Forecasting, 13, 125-137.
Michael C. Coniglio, 2012: Verification of RUC 0–1-h Forecasts and SPC Mesoscale Analyses Using VORTEX2 Soundings. Wea. Forecasting, 27, 667–683.
Thomas M. Hamill and Andrew T. Church, 2000: Conditional Probabilities of Significant Tornadoes from RUC-2 Forecasts. Wea. Forecasting, 15, 461–475.
Joshua Wurman, David Dowell, Yvette Richardson, Paul Markowski, Erik Rasmussen, Donald Burgess, Louis Wicker, and Howard B. Bluestein, 2012: The Second Verification of the Origins of Rotation in Tornadoes Experiment: VORTEX2. Bull. Amer. Meteor. Soc., 93, 1147–1170.