|
|
REU Home | 2009 Home | Students & Mentors | Projects | Schedule | Travel Tips | Photos |
|
NWC REU 2009May 26 - July 31 |
||
|
|
Project ProposalsThis page is in the process of being updated with the completed papers. Projects were directly funded by this REU, except as otherwise noted for two projects. Last updated August 12, 2009 Tornado False Alarms on Days With No Reported Tornadoes: A Climatological and Radar Survey PDFHannah Barnes A five-year study of tornado false alarms, as issued by the National Weather Service (NWS) from 2000 to 2004, found that 31.6 % of false alarms occurred on days when the Weather Forecast Office (WFO) was unable to confirm at least one tornado in their county warning area (CWA) from midnight to midnight (i.e., a zero day false alarm). Reviewing tornado warning data obtained from the National Atmospheric and Oceanic Administration (NOAA)/NWS, this study conducted a climatological and radar survey to diagnose situations when false alarms were issued on days with no confirmed tornadoes, i.e., marginally severe days that failed to produce tornadoes. The study was composed of three steps. First, zero day false alarms were compared to tornado day false alarms (days when tornadoes were confirmed within the WFO CWA). Climatological trends were identified in terms of time of day and year, geographic region, county population density, and distance from nearest WSR-88D radar. Second, the impact of the perceived largescale tornadic potential was explored by an examination of Watch information from the Storm Prediction Center. Third, Level 2 radar data were examined. Reflectivity and velocity radar data were used to identify the impact of storm morphology, purple haze, and variations in circulation intensity with height. This survey suggests four important trends: (i) zero day false alarms comprise a larger percentage of the total number of tornado false alarms in geographic regions less susceptible to tornadoes, (ii) zero day false alarms are more similar to one tornado day warnings than outbreak day false alarms in terms of the perceived large-scale tornadic potential, (iii) the circulation intensity of zero day false alarms and outbreak day false alarms at the lowest height scanned by a WSR- 88D radar is notably weaker than those associated with one tornado day warnings, and (iv) purple haze may be a considerable factor in zero day false alarms and outbreak day false alarms. Forecasting Southern Plains Wind Ramp Events Using the WRF Model at 3-KM PDFKristen Bradford Wind ramp events—extreme and rapid changes in wind power output due to abrupt changes in wind speed—are a growing concern for the wind energy industry; therefore, precise forecasting of these phenomena is crucial to the advancement of wind power in the United States. Weather Decision Technologies, Inc., (WDT) is partnering with NanoWeather, Inc., to create a wind forecasting system, called WindPredictorTM, in order to precisely predict winds (and, in turn, ramps) for the energy industry. WDT’s contribution to WindPredictor will be a customized version of the Weather Research and Forecasting (WRF) model, which is currently being run on a 3-km grid. This paper assesses the 3-km WRF’s performance regarding ramp event prediction. A comparison between surface wind forecasts and hourly METAR observations was utilized to assess its performance. David Goree A Social Perspective Of Warn On Forecast: Ideal Tornado Warning Lead Time and the General Public's Perceptions Of Weather Risks PDFStephanie Hoekstra In this study, 136 National Weather Center visitors were surveyed to assess their understanding and perception of weather risks. The majority of the respondents performed well overall. They seemed to be familiar with tornado seasons, however, they were not aware of the relative number of fatalities caused by several weather phenomenon each year in the United States. This study also aimed to pinpoint the ideal tornado warning lead-time for the general public, which was found to be 33.5 minutes. This justifies the fact that a longer lead-time of 1-2 hours, of which the possible future tornado prediction paradigm called warn on forecast could provide, would not be necessary for the general public. In fact, when asked what they would do if given a one-hour lead-time, respondents reported that taking shelter was a lesser priority than if given a 15-minute lead-time, and fleeing the area became a popular alternative. The majority also reported the situation would feel less life threatening if given a one-hour lead-time. Responses were analyzed according to several difference parameters, including age, region of residency, and educational level, however no significant conclusions can be drawn when evaluating how these variables can change the public’s perceptions of weather risks or their preferred ideal lead-time. Thus, the results of this study are informative to future studies, which evaluate the true impact of warn on forecast on the public, since the social perspective of a longer lead-time is often overlooked and underresearched. Verification of ESTOFEX Lightning and Severe Weather Forecasts PDFAlex Kowaleski The European Storm Forecast Experiment’s (ESTOFEX) daily 2006-2009 ordered lightning and severe weather forecasts were analyzed by using a two by two contingency table. Probability of detection (POD), frequency of hits (FOH), probability of false detection (POFD), critical success index (CSI), and bias were calculated. These scores were compared among seasons and years to determine how forecasting skill varied by season, and how it changed from 2006 to 2009. They were also compared among forecasters to determine if some forecasters were more skilled than others. It was determined that severe weather forecasts improved in both POD and FOH scores between 2006 and 2009, Forecasts of lightning, however, did not consistently improve during the forecasting period. It was also determined that ESTOFEX issued superior lightning and severe weather forecasts during summer, and their forecasts were less successful during fall and winter. The differences in forecasting success among forecasters, however, were not sufficiently large to determine if some forecasters were more skilled than others. An analysis of the Relative Operating Characteristics curves (ROCs) of ESTOFEX severe weather forecasts indicated that they were useful for decision-making. Evaluating Interaction in Visualization Tools in Meteorology PDFKaren Nielson Meteorological software is rapidly and frequently updated to keep up-to-date with current developments in the field. However, little research has been done concerning the accessibility and efficiency of such tools. In this study, we interviewed three meteorology researchers concerning their usage of the Warning Decision Support System – Integrated Information (WDSS-II) visualization tool, and then observed while completing everyday tasks using the tool. We expected to uncover a wide variety of interaction patterns during these observations. Responses and interactions were also analyzed to find the strengths and limitations of the software. We found that our participants have distinct but overlapping approaches to assigned tasks. Certain features, such as the mouseover readout, were popular with all participants, while other features were used minimally or not at all. Our findings may provide some insight that can be built upon in future studies. Warning the Public About Hail: Determining the Potential for Short-Term Damage Mitigation PDFLauren Potter Previous attempts to mitigate hail damage have focused primarily on protective efforts, such as insurance programs, which require long-term planning. Very little research has been done to determine what actions can be taken to prevent hail damage just before it occurs. In this study, hail reports taken from the National Climatic Data Center (NCDC) Storm Events Database were compared with NWS warnings for Colorado, Massachusetts, Oklahoma, and South Carolina during the years of 1999-2008. Warning accuracy and average lead time were determined for all hail reports during 1999 and 2000 as well as damaging hail reports between 1999 and 2008. It was found that in general, under the county warning system, sufficient lead time existed for mitigating action by the public. While less than 50% of NWS Severe Thunderstorm Warnings resulted in hail reports, it is likely that many hail events were not reported in the NCDC data base. Using data from the Severe Hail Verification Experiment (SHAVE), an average hail swath size of 863 square kilometers (333.2 square miles) was calculated from 14 hail swaths across the United States. The average polygon warning size covered more area than the average hail swath size, while county warnings covered a much larger area and would prompt an unnecessary number of people to take mitigating action. There are serious limitations as to the data sources available to make decisions for damage mitigation. A more thorough, more accurate data base would improve the economic cost-benefit analysis of various possible mitigation methods. First Look at Forecasters' Experiences With High-Temporal Resolution Phased Array Radar Data: An Evaluation Research Study PDFCristal Sampson The Phased Array Radar (PAR) is a research radar that is under consideration to replace the Weather Surveillance Radar-1988 Doppler (WSR-88D). As a new technology it is important to provide user insight into the development stage to ensure intended users have the most usable tool upon deployment and not only understand the operational utility of PAR. Results from experiments held in 2008 and 2009 have already assisted researchers developing PAR. The participants of these experiments evaluated real-time and archived cases; after each evaluation questionnaires were completed. The responses to two archived cases were analyzed in this paper using a data-driven method. The results show how high-temporal resolution data of PAR impacted the participating forecasters in a simulated warning environment. Suggestions are made to improve future research and development. Numerical Forecasting of Banded Snow: A Case Study PDFAstrid Suarez Banded snow is challenging to forecast using numerical models because the bands have varying temporal and spatial scales. Although numerous ingredients-based forecasting strategies have been developed, their successful application relies on accurate forecasting of the location and intensity of the ingredients themselves. One possible way to improve numerical forecasts of banded snow or banding indicators is through the use of convection-permitting and/or ensemble modeling techniques. The study examines three forecasts of a banded snow event in association with a shallow, short-lived low-pressure system over southern Indiana on 3 February 2009. The forecasts include two single deterministic experiments with 12- and 3- km grid spacing (S12km and S3km, respectively) and a 30-member 12 km ensemble forecast with a 12-h data assimilation training period (EnKF12km). Of the three experiments, the ensemble mean of EnKF12km provides the best forecast of the position and strength of the surface low pressure. Banding ingredients, including frontogenesis and low-level moisture, are also considered; and again, the EnKF12km experiment provides the best forecast of the position of these fields. The convection-permitting simulation, on the other hand, positions the indicators slightly too far to the south, but resolves the northwest to southeast oriented bands over southern Indiana. These findings are consistent with previous studies that suggest for spring time mesoscale convective systems the most effective forecasting strategy is to couple high-resolution and ensemble forecasts to assess the character and location of a given event, respectively. A Stochastic Daily Mean Temperature Model For Weather Derivatives PDFJeffrey Viel Weather derivatives are usually priced by analyzing the climatologic data of an underlying weather index. This research proved that when using heating and cooling degree days as an underlying weather index that climatology was not fully representative of future outcomes. Previous research attempted to develop techniques for daily mean temperature simulations, but these techniques were based on invalid statistical assumptions and lacked time dependencies in the residuals. Due to the fact that heating and cooling and degree days are an aggregate monthly metric and path dependent, it was important to simulate the complete behavior of a time series. This research provided an in depth statistical analysis of the daily mean temperature time series for eighteen cities from the Chicago Mercantile Exchange (CME). The residuals were used to develop two models to simulate a possible temperature time series for 2007. A distribution of ten thousand possible outcomes were created for each model, and then analyzed against the climatologic data sets. Ultimately, this research exhibited that techniques involving daily mean temperature simulation from the statistical analysis of the residuals could accurately produce likely outcomes of degree days for weather derivative contracts to be based upon. Diagnosis of Azimuthal Shear Associated With Tornadoes PDFTravis Visco Current rotation detection algorithms are prone to poor performance due to high local variance in a Doppler radial velocity field. A new linear least squares derivatives (LLSD) technique has been developed which is less prone to problems from such variability and outputs a number of fields diagnosing the radial wind field, including azimuthal shear and radial convergence. We subjectively diagnosed rotation within thunderstorms from seven storm days which yielded eighty tornadoes. Threshold values and trends of azimuthal shear and radial convergence fields were extracted from the tornado producing storms. These values and trends will be used to train a new rotationdetection algorithm. Impacts of Super-Resolution Data on National Weather Service Warning Decision Making PDFJonathan Vogel Super-resolution data provided by the Weather Surveillance Radar-1988 Doppler (WSR-88D) has changed the look and feel of radar data and impacted the warning decision making of National Weather Service (NWS) warning forecasters. Since the Build 10 upgrades to the WSR-88Ds in 2008, spatial resolution of radar data was enhanced from legacy resolution to super-resolution. The improvement should result in both more detailed storm features and storm feature identification at distances 50% greater in range. These details ought to have allowed for increased lead time, decreased false alarm ratio (FAR), and better warning decision making. However, no formal study was previously completed to determine if these expectations had been realized. For this study, a survey was sent to forecasters from all Weather Forecasting Offices (WFOs) in all regions of the NWS who were expected to have experience using radar data in a warning decision environment. While understanding of the technological aspects of super-resolution data appears to be lacking, 50% to 70% have seen a perceived improvement in storm feature appearance and 30% to 50% have seen a perceived improvement in identification at farther ranges. A majority also agrees there is potential for increased lead time and decreased FAR, but it is too early to tell numerically how super-resolution has impacted them. The most surprising change was that 60% have noticed positive changes to wintry precipitation echoes due to super-resolution data. Overall, there is strong NWS support for super-resolution data and mainly positive impacts in the warning decision environment. However, for some aspects, such as lead time and FAR, it is too early to clearly quantify results. Viability of Weather Dissemination Via Social Network Technologies PDFJustin Wittrock Social networks are a recent product of the ever-increasing advancement in online technology. Few studies have been conducted to determine the viability of disseminating weather information via social network technology. Instead, most of the studies that are currently available focus on weather information being distributed through radio, newspapers, TV stations, and the Internet. Social networks serve as a useful source for rapid weather communication, especially during the occurrence of severe weather phenomenon. This study depends on public perception toward using a social network as a medium for weather information distribution and determines whether it is feasible to do so. This purpose is fulfilled by conducting a web-based survey created through SurveyMonkey.com.
Copyright © 2009 - Board of Regents of the University of Oklahoma |
|