Forecast Validation of Three Sierra Nevada Precipitation Events

Rebekah I. Banas
Mentor: Dr. Heather Dawn Reeves

Abstract:

In this paper, three precipitation events which occurred in the Sierra Nevada Mountains are investigated. These three events are all associated with atmospheric rivers, from which the resulting precipitation can be quite heavy. The 24-, 48-, and 72-h NAM model forecasts of temperature, dewpoint, and 24-h accumulated precipitation are analyzed at four Sierra Nevada stations. Temperature and dewpoints errors from the model output show a tendency to overestimate the temperature and underestimate the dewpoint during these three events. For one event, the precipitation was strongly underestimated (by about half). The other two events show no specific trend in either time or space.

Full Manuscript

 

The basis for this study:

• Atmospheric rivers can result in heavy precipitation, especially in mountainous regions on the US west coast, like the Sierra Nevada Mountains
• These rivers vary and so does the resulting precipitation amounts/types
• Forecast accuracy for these events is thus very important

What this study adds:

• Exploratory look into the NAM forecasts with some preliminary results
• Provides basis for more in-depth investigation on the predictability of atmospheric rivers and heavy precipitation in the Sierra Nevadas

Comparison of Estimated and Observed Storm Motions to Environmental Parameters

Eric Beamesderfer
Mentors: Kiel Ortega, Travis Smith, John Cintineo

Abstract:

This study explores current storm motion techniques and analyzes their accuracy with respect to different environmental parameters. Current motion estimates are compared to observed motions and different environmental parameters. The parameters investigated are the heights of the lifted condensation level (LCL) and the level of free convection (LFC), the mean relative humidity from the surface to 0°C and the storm relative helicity (SRH) from 0-3 km. Deviate estimates were seen by each storm motion estimator for the different environmental parameters. Also, it was evident that some storm motion estimators were superior to others. However, overall the observed motions in this study were dissimilar to the environmental.

Full Manuscript

The basis for this study:

• Storm motion estimates are either storm type specific or are inaccurate due to the fact they are determined by the kinematic field only
• Numerical modelling studies show that thermodynamic variables are also important to storm motion, yet no work has been completed with observations

What this study adds:

• Overall, all motion estimators are fairly inaccurate
• Investigated thermodynamic variables did not reveal additional information about storm motion estimate errors
• Storm relative helicity, a variable derived from the kinematic field, showed the most influence on storm motion estimate errors. However, helicity values are closely related to storm mode (i.e., supercell vs. non-supercell) thus storm mode may be more important than specific environmental variables

Quantifying Changes in Extreme Precipitation at Houston and Oklahoma City by 2041–2065 Using the Canadian Regional Climate Model (CRCM)

Daniel J. Brouillette
Mentors: Dr. Yang Hong, Lu Liu

One-half-degree gridded daily-projection precipitation model output from two combinations of the Canadian Regional Climate Model (CRCM)—one driven by the Community Climate System Model (CCSM) and another by the Canadian Global Climate Model 3 (CGCM3)—was obtained from the North American Regional Climate Change Assessment Program (NARCCAP). Gridded observational daily precipitation data were used as a reference to a 1971-1995 historic period and as a basis for validating the projection data. Validation suggested strong bias in the projection data, which necessitated that they be bias-corrected using a mean-value technique. Both the observational and projection data were ranked and assigned percentile values as a means of identifying and quantifying possible changes in extreme precipitation during a historic 1971-1995 and a future 2041-2065 period over two 1/2-degree grid squares centered over Houston and Oklahoma City. Overall results of the percentile analysis suggested that, for the highest percentile rankings, the daily precipitation values associated with a given percentile ranking will increase by the 2041-2065 period. For more moderate percentile rankings, the tendency toward change was less clear. For lower percentile rankings (approximately the 80th), there was indication that the values associated with a given percentile ranking will decrease by the future period. Analysis also suggested that a more sophisticated bias-correction procedure based on rain rate is necessary.

Full Manuscript

 

The basis of this study:

• Extreme precipitation has been increasing in frequency and intensity in the southern Great Plains in the last 50 years and is likely to continue to do so according to many studies.
• Daily-projection climate model output can be used to study such changes in the future by the 2041-2065 period (over an historic 1971-1995 period) quantitatively for Houston and Oklahoma City.

What this study adds:

• Extreme precipitation may increase in intensity in the future at Oklahoma City and, particularly, Houston.
• There is indication that overall precipitation may decrease in the future at the two locations.
• The overall suggestion is that drought will be more common and interspersed by more intense extreme precipitation events in the future at the two locations.

Deriving Population Exposure Fatality Rate Estimates for Tornado Outbreaks Using Geographic Information Systems (GIS)

Amber R. Cannon
Mentors: Kim Klockow, Randy Peppler, Dr. Harold Brooks

In this study we looked at several issues regarding the derivation of population exposure and fatality rate estimates during widespread tornado outbreaks. The two events studied were the April 3, 1974 and April 27, 2011 tornado outbreaks in the state of Alabama. We attempted to determine if tornado warnings have become more effective over time at reducing the number of fatalities. We used GIS to perform an analysis on these outbreaks. We found that the effectiveness of tornado warnings did improve between the two outbreaks, and we can have reasonably high confidence in using county level population data to compare recent and historical outbreaks, although the higher resolution of the census track data is preferred for studying a single tornadic event. We also found that the accuracy of fatality rates is directly related to the accuracy of the path data. Finally, GIS can be used to innovatively evaluate tornado warning effectiveness.

Full Manuscript

 

The basis for this study:

• The tornado outbreak in April of 2011 caused a high number of fatalities.
• Effectiveness of current tornado warning procedures was called into question.
• Are present-day tornadic outbreaks stronger or more deadly than historical outbreaks?

What this study adds:

• Rubrics for 1) outbreak population and fatality rate analysis and 2) the fatality rate for a single event provide a basis for comparison between two or more events.
• Tornado warning effectiveness in Alabama has improved from 1974 to 2011 based on these outbreaks.
• County level population appears reasonable for comparing historical and present day outbreaks.
• GIS can be used to innovatively evaluate tornado warning effectiveness.

Choosing the Most Accurate Thresholds in a Cloud Detection Algorithm for MODIS Imagery

Tracey A. Dorian
Mentor: Dr. Michael W. Douglas

We use a cloud detection algorithm that detects cloudy pixels from MODIS images by characterizing individual pixels as cloudy or non-cloudy based on the brightness values of the pixels and a predetermined threshold. The algorithm then produces mean fields of daytime cloudiness over different geographical regions. Although the cloud climatologies produced initially appeared realistic, it was found that the algorithm largely underestimated the cloud frequencies over some regions when using a threshold of 215. Analyzing various MODIS images and recording cloudiness over different sectors served as the “ground truth” data which we compared to the algorithm output. After comparing the subjective estimates and the algorithm output for four regions of the world, we found that the algorithm underestimates cloudiness over these additional regions and that lowering the thresholds to 170-190 over oceans and 190-215 over land generally identified the thick clouds most accurately. Studying more regions or extending research on certain regions will allow us to better understand how the algorithm behaves with certain types of cloudiness and geography. Even though the thresholding technique is somewhat arbitrary, by better understanding how the algorithm behaves we can modify the algorithm to ensure that the output more accurately describes cloud climatologies around the world. If we are able to do this, then our algorithm could be used for many applications such as validating the numerical model simulations of cloud climatologies or assessing climate and potential climate change.

Full Manuscript

 

The basis for this study:

• High resolution cloud climatologies are needed for mesoscale model validation, solar energy mapping, and other applications. 
• MODIS satellite imagery has not been fully exploited to produce high resolution cloud climatologies.

What this study adds:

• We have improved cloud detection by adjusting threshold values for our cloud detection algorithm. 
• The MODIS-based climatologies are now more realistic -- especially in stratus regions over oceans. 
• Both advantages and limitations exist in simple techniques to identify clouds from MODIS imagery.

Impact of AQUA Satellite Data on Hurricane Forecast: Danielle 2010

Travis J. Elless
Mentors: Dr. Xuguang Wang, Dr. Ting Lei, Dr. Govindan Kutty Mohan Kumar

This study focuses on the impact of AQUA satellite data from AIRS and AMSU on the forecast of hurricane Danielle by the Global Forecast System (GFS) model. The data assimilation method adopted to ingest the data is the Gridpoint Statistical method (GSI) which is based on the three dimensional variational (3DVAR) data assimilation technique. Two experiments were carried out to investigate the impact of AQUA satellite radiance observation on the forecast of the hurricane Danielle. The first experiment (Control), assimilated all the available data while the second experiment (No AQUA) incorporated all the observations but the AQUA satellite data. Data assimilation cycling started one week prior to hurricane genesis, on 15 August 2010 06 UTC. The root mean square track forecast error shows slightly negative impact at the early lead time and slightly positive impact at later lead time. However, the root mean square intensity forecast errors by the Control are shown to be lower than No AQUA for all forecast hours, indicating positive impact of the AQUA data on the intensity forecast.

Full Manuscript

 

The basis for this study: 

• An accurate 120 hour forecast would increase time for hurricane preparedness.
• Assess impact of AQUA satellite data on a hurricane forecast.

What this study adds:    

• Assimilation of Aqua data improved the intensity forecast of Hurricane Danielle
• There was no apparent impact on the forecast track

Sensitivity of Microphysical Parameters on the Evolution of a Supercell

Samuel P. Lillo
Mentor: Dr. Edward R. Mansell

Due to limited computational resources, critical microphysical processes must be accounted for in models through parameterization schemes. These schemes use many constants that have large uncertainties and may vary in nature spatially and temporally. By perturbing individual parameters within a single scheme, an ensemble can be created to attempt to account for the uncertainty in the model physics.

Five ensembles are created to test the sensitivity of a simulated supercell to the following parameters: cloud condensation nuclei (CCN) concentration, the efficiency of cloud water collection by graupel and hail, the fraction of liquid water allowed on graupel and hail, rime density function, and the drag coefficient as a function of particle density. A range of values was chosen for each parameter to represent the uncertainty that exists within the model microphysics. All ensembles exhibited growing variance through the simulation. Monotonic association to the storm evolution was most prominent in the CCN ensemble, in which there were notable variance in the track and intensity of the supercell.

Full Manuscript

 

The basis for this study:

• Parameterization schemes in models use many constants that have large uncertainties.
• Storm scale models have been proposed to assist in advanced warnings.
• For these models to be useful, we need to understand how the uncertainty surrounding these constants can impact model forecasts.

What this study adds:

• Perturbing microphysical parameters in the model can affect the motion, intensity, and severe characteristics of the storm.
• These results demonstrate the need for storm-scale ensemble physics diversity using multi-moment microphysics in future Warn-on-Forecast applications.

Observations of a Supercell and Weak Tornado Made With a Rapid-Scan, Polarimetric Mobile Radar

Alex Lyakhov
Mentors: David Bodine, Dr. Robert Palmer

A rapid-scan, X-band, polarimetric, mobile Doppler radar is used to collect horizontal reflectivity, differential reflectivity, cross correlation coefficient and radial velocity data of a supercell that produced two EF-0 tornadoes in Osage County, OK on 18 June 2011. Volume scans of the first tornado, which lasted a few minutes, were acquired every 30 s. Analysis of data reveals several common polarimetric radar signatures associated with supercells including the low-level inflow, low-level hail and differential reflectivity arc signatures. The low-level inflow and differential reflectivity arc signatures both decreased in prominence around the time of tornado dissipation. No tornadic debris signature was noted, likely owing to the fact that the tornado was too weak to loft heavy debris, suggesting it is difficult for polarimetric radars to detect weak tornadoes. A Three Body Scatter Spike was also evident in the data, suggesting the presence of large hail aloft. Doppler velocity data reveal that mid-level mesocyclone intensification is not a pre-requisite for tornadogenesis. A trend of increasing azimuthal shear with time up to tornado dissipation is observed in the lowest two elevation scans, as well as within the low and midlevel mesocyclones. Azimuthal shear decreased after tornado dissipation in the lowest two scans and the low-level mesocyclone, but not with the midlevel mesocyclone. Furthermore, an anticyclonic circulation accompanied the cyclonic mesocyclone. A hook signature in the reflectivity field was observed with the mesoanticyclone, which later morphed into a linear feature.

Full Manuscript

 

The basis for this study:

• There is little data of weak, brief, supercell tornadoes from rapid-scan, polarimetric radars.
• Little is known about how rapidly evolving processes—such as tornadogenesis— correlates with polarimetric signatures.
• May be possible to correlate distinct polarimetric and velocity signatures with tornado intensities.

What this study adds:

• This case study showed that the polarimetric inflow and differential reflectivity arc signatures weakened rapidly around tornado dissipation.
• Azimuthal shear decreased within the low-level mesocyclone following tornado dissipation, but did not immediately decrease within the midlevel mesocyclone.

Investigation of Radar Variables and Near Surface Environments for Developing a Surface Hail Fall Product

Sarah K. Mustered
Mentor: Kiel Ortega

In 2006, the Severe Hazards Analysis and Verification Experiment (SHAVE) was formed to collect high resolution severe weather reports. The resulting dataset is a detailed and accurate record of surface hail fall. Currently, the exact processes that determine the size and distribution of hail at the surface is relatively unknown. While there are numerous products that address the presence of hail cores aloft, a gridded surface product is missing. The benefit of a gridded surface product is a more accurate understanding of surface hail size at any particular location. In this study, we incorporated the near surface environment (NSE) from the 20 km RUC analysis and existing radar products with SHAVE hail reports to determine if the NSE could be a beneficial component of a surface hail fall product. We found that the current resolution and reliability of the NSE data is too low for the addition of NSE variables to add significantly to the accuracy of the existing radar products.

Full Manuscript

 

The basis for this study:

• Surface hail fall poses a significant threat to life and property, but the processes that affect hail size and spatial distribution are not well understood.
• Current hail products focus on hail detection aloft.
• A gridded surface hail product is needed to better track severe hail events.

What this study adds:

• The addition of NSE data to radar variables did not stratify hail size categories.
• Use of high spatial resolution verification data will require more novel searching and matching techniques.
• While addition of NSE data did not stratify the hail sizes, it did highlight certain parameter spaces which are more (or less) conductive to large hail production.

A Comparison of Wind Estimates From CASA and NEXRAD Radars During Severe Wind Events

Adam Taylor
Mentors: Dr. Jerald A. Brotzge, Dr. Frederick H. Carr

The Engineering Research Center for Collaborative Adaptive Sensing of the Atmosphere (CASA) Oklahoma test bed has proven the usefulness of a high-resolution, rapidly updating network of radars for a variety of applications. The aim of this project is to quantify the value of the CASA network for the detection of severe wind events. A comparison is made between the performance of Next Generation Doppler Weather Radar (NEXRAD) and CASA radial wind measurements in relation to Oklahoma Mesonet reports of high wind gusts. Two factors inhibit the accurate measurement of winds from weather radar: (1) The viewing angle of the radial velocity beam, and (2) the beam height above ground level. Results show that the CASA radar network performed better overall for detecting and analyzing high wind events within the test bed. CASA dual-Doppler data improved the measurement of winds by 7.27 m/s over all NEXRAD measurements.

Full Manuscript

 

The basis for this study:

• Severe wind events are not well detected with the current NEXRAD radars
• CASA radars provide an unprecedented dataset of low-altitude wind speed measurements of hydrometeors
• Aim to quantify the differences between estimations of surface winds from CASA and NEXRAD

What this study adds:

• Viewing angle was the most important source of error
• Dual Doppler and Profile correction is around 10% more accurate than best case NEXRAD and almost 25% better than all NEXRAD estimates