Search

Heatwaves are a leading contributor of weather-related mortality, globally contributing to thousands of deaths each year. The impacts on humans may be direct or indirect through avenues such as heat stress, strained medical capacity, infrastructure breakdown, and reduced crop yields. While extreme heat is often measured by temperature and humidity, Wet Bulb Globe Temperature (WBGT) is commonly used to evaluate real-time heat stress risks in humans and correlates better with heat related illness, and is used by the Occupational Safety and Health Administration (OSHA) and the US Army. WBGT is a weighted average of air temperature, natural wet bulb temperature, and black globe temperature. A local hourly, daily, and monthly WBGT climatology will allow those planning outdoor work to minimize the likelihood of heat related disruptions. Further, understanding the characteristics of heat waves will allow emergency planners and responders to know what to expect when heat waves occur. Additionally, evaluating the predictability of WBGT heat waves allows an understanding of what advance warning may be possible and when confidence may be higher. In this study, WBGT is calculated from the ERA5 reanalysis and is validated by the Oklahoma Mesonet and found to be adequate. Two common methods of calculating WBGT from meteorological observations are compared. The Liljegren method has a larger diurnal cycle than the Dimiceli method, with peak WBGT about 1 °F higher. The high and extreme risk categories in the southern United States Great Plains (USGP) have increased from 5 days per year to 15 days from 1960-2020. Additionally, the largest increases in WBGT are occurring during DJF, potentially lengthening the warm season in the future. Heat wave definitions based on maximum, minimum, and mean WBGT are used to calculate heat wave characteristics and trends with the largest number of heat waves occurring in the southern USGP. Further, the number of heat waves is generally increasing across the domain. This study shows that heat wave days based on minimum WBGT have increased significantly which could have important impacts on human heat stress recovery. The predictive skill of WBGT heat waves is evaluated using ERA5 reanalysis and models from the S2S Project Database. North American atmospheric regimes are defined using K-Means clustering of detrended standardized 500 mb geopotential anomalies from ERA5 reanalysis. Additionally, heat wave types (e.g. Hot-dry or warm-humid) are defined using the standardized anomalies of temperature and humidity relative to other heat waves during the same season. An analysis of the predictive skill of atmospheric regimes, heat wave types, seasonality, and the impact of ENSO and the MJO is conducted using these datasets both with zero days lead time and at S2S lead times. Finally, regime statistics are calculated in S2S model forecasts to identify skill and forecast biases. Each regime has unique heat wave frequency, type, and seasonality characteristics. Heat waves may occur in the US Great Plains with greater than twice the climatological frequency in some regimes, however the increase is often seasonally dependent. Additionally, some skill is shown at discriminating between heat wave types in different regimes. Further, the regimes that are conducive to heat waves at short lead times differ from those that correlate with heat wave occurrence at longer lead times. When incorporating the ENSO or MJO phase heat wave in addition to the atmospheric regime, some additional predictive skill is observed, with the MJO providing more skill, as it introduces larger variation between regimes and provides information regarding timing of individual heat waves where the ENSO phase does not due to the much longer period. Combining both the ENSO and MJO phase provides the highest skill, with some ENSO/MJO/regime combinations having historically observed heat wave rates in excess of 50% at some lead times while other combinations are near 0% historically, thus leading to forecasts of opportunity when a signal for higher heat wave rates occurs. S2S models show statistically significant skill in forecasting most regimes up to 5 weeks lead time. However, the accuracy at predicting the correct atmospheric regime may not be useful beyond 1-2 weeks, which may limit the ability to incorporate regimes into heat wave forecasts at longer lead times, as this information is necessary for combination statistics that incorporate regimes.

Flash drought is the rapid intensification of drought like conditions. Initiated by meteorological processes that quickly desiccate soil they can be devastating to agriculture and local ecosystems. Flash droughts are driven by a complex interaction of many terrestrial and atmospheric processes making prediction challenging. The prediction of future flash drought events will require a better understanding of these complex interactions. In particular, synoptic scale processes that lead to changes in evaporative stress are poorly understood, and little is known as to how they impact the local processes. To better predict flash drought both remote teleconnections and the effects of evaporative stress on local terrestrial conditions must be identified. Additionally, the temporal lag from remote drivers to flash drought development needs to be quantified. Using current datasets, multiple methods focusing on synoptic and local spatial scales from seasonal to sub seasonal temporal scales were used to increase our understanding of conditions that drive flash drought. Seasonal atmospheric variables were used to identify areas in Europe primed for agricultural flash drought development in a later season. Additionally, sea surface teleconnections and synoptic scale atmospheric processes were examined before and during flash drought development to identify remote drivers. Lastly, the identified remote drivers were mathematically and statistically assessed for both their covariance with evaporative stress and their correlation with changes in evaporative stress in the Central US.

Increasing precipitation extremes across the globe will naturally lead to not only increased flooding but an increase in the occurrences of flash floods, with assessment and prediction of these events becoming more critical every year. Floods are a natural facet of the water cycle and serve as an unmistakable indicator of changes in that cycle globally. As such, the characterization and monitoring of floods also needs to be undertaken at a global scale especially in the face of an ever-changing climate. A key tool in that process is understanding the inherent capabilities of satellite-based precipitation products in their ability to model flood characteristics. Hence, this work focuses on assessing the innate differences in the high-resolution Multi-Radar Multi-Sensor (MRMS) system and the Integrated Multi-satellitE Retrievals for GPM (IMERG) suite of satellite products when used as precipitation forcings to simulate hydrologic outputs through the operational Ensemble Framework for Flash Flood Forecasting (EF5) hydrologic framework. The biases of precipitation presented by satellite products has been well studied, but less has been done to assess how significant these errors persist into physical hydrologic processes especially when compared to ground-based radar estimates. By using each precipitation product as an individual forcing for EF5, the simulated hydrographs can be post-processed where discrete simulated flood events are determined and matched, assessing the ability of each precipitation product to accurately generate reliable simulated representations of flood characteristics. Relationships between these flood characteristics, such as peak flow and flood duration, and physical basin characteristics are investigated. Steps are also undertaken in reimagining the calculation for a timing characteristic of floods in order to increase the accuracy of the estimates, allowing for future integration into the methodologies presented herein. Increasing knowledge regarding the capabilities or deficiencies of satellite precipitation products with respect to flash flood modeling will have implications on the ability to characterize flood events over areas with little or no coverage by ground-based precipitation monitoring networks and subsequently improve flood forecasting operations globally.

Flash droughts are highly impactful, subseasonal to seasonal events that pose a severe risk to agricultural production. The 2012 flash drought event in the United States is a prime example, as this event resulted in tens of billions of dollars of crop loss and caused longer-lasting effects to the overall US economy. This study examined impactful flash drought events across two agricultural regions of the central United States, spanning the 40-year period from 1981 to 2020. The two regions, the Southern Great Plains and Midwest, were selected given the agriculturally-dense areas in distinctively unique climate regions. Using the standardized evaporative stress ratio (SESR), flash drought events were selected based on several factors, including the overall spatial coverage, rapid onset, and spread across the region. The events, defined as abrupt agricultural flash droughts (AAFDs), provide critical information regarding the timing of rapid drought transition across different areas of the United States and the implications they bring to agricultural producers. Initial results show that over the last 20 years, AAFDs have increased in frequency across critical regions of agricultural growth. Within the Southern Great Plains, essential findings within the timing of events illustrate how AAFDs, in conjunction with a winter La Niña pattern, led to a consistently below-average harvest of winter wheat the following growing season. Across the Midwest domain, results indicate a greater seasonality in AAFD events, with a majority of events transpiring during the late Spring and early Summer months.

Spaceborne radars uniquely measure, provide the finest depiction of, and give the most accurate estimate of precipitation globally from space. The Global Precipitation Measurement (GPM) mission dual-frequency precipitation radar (DPR) is the successor to the Tropical Rainfall Measuring Mission (TRMM) precipitation radar (PR), expanding on its capabilities with a dual-frequency radar and coverage into the midlatitudes. The consistent ability to detect various precipitation magnitudes across satellite missions is critical to the study of global precipitation over various time periods. The precipitation delineation capabilities of spaceborne radars are characterized as functions of their reflectivity and the corresponding precipitation magnitude from the reference Ground Validation Multi-Radar/Multi-Sensor (GV-MRMS) over CONUS. The Heidke Skill Score, a measure of skill with respect to random chance, is computed to synthesize the capabilities of the spaceborne radars. This enables a finer depiction and interpretation of spaceborne radar capabilities than the bulk metrics widely used in the literature. Skill is more sensitive to changes in rain rate at lower rain rates and changes in reflectivity at higher rain rates. The TRMM-PR and GPM-DPR best delineate moderate precipitation while the GPM-KuPR detects precipitation with low to moderate skill. While both the TRMM-PR and GPM-DPR perform better than GPM-KuPR, the TRMM-PR performs better at lower reflectivity thresholds while the GPM-DPR performs better at higher reflectivity thresholds. Certain factors do not have a significant impact on the overall skill. Others have a significant impact, but the number of cases were small and did not greatly impact the overall skill. While the GPM-DPR struggles with the detection of precipitation, and the TRMM-PR performs the best overall, both the GPM-DPR and TRMM-PR have good delineation capabilities.

Precipitation microphysics in tropical cyclones (TCs) are often poorly represented in numerical simulations, which ultimately affects TC structure, evolution, and prediction. This provides a large incentive to better observe and understand the underlying microphysical processes in TCs in order to improve precipitation forecasts and improve warning operations. Recently, ground-based polarimetric radar observations have been able to capture the evolution and structure of precipitation in landfalling TCs in the United States, revealing numerous microphysical processes through the investigation of vertical profiles of dual-polarization radar variables. While ground radars are a useful tool for quantifying precipitation processes in TCs, they are unable to sample precipitation when TCs are over the open ocean. Therefore when ground radar networks are sparse or non-existent, space-borne radar can provide precipitation retrievals of TCs at snapshots in time. This is particularly useful for monitoring the evolution of precipitation in TCs prior to landfall. Specifically, this dissertation investigates precipitation microphysics in TCs using the NASA Global Precipitation Measurement (GPM) mission dual-frequency precipitation radar (DPR) on a global scale, and is complimented by polarimetric ground radar observations, disdrometer data, and reanalysis data when available.

Winter weather has substantial impacts on human activity, but it is challenging to predict more than 7-10 days in advance. Subseasonal-to-seasonal (S2S) prediction is a rapidly developing field of meteorology that could be used to improve winter weather prediction at longer time scales. The Madden-Julian Oscillation (MJO) is a major driver of global S2S variability. This study seeks to examine how the MJO affects the frequency of winter weather over the contiguous United States, which will help further improve winter weather prediction in the future. This is done by comparing the frequency of winter weather reports and snow accumulation data at surface meteorological stations in a given phase of the MJO to its climatology. We find that the MJO has significant impacts on increasing the frequency of extreme winter weather in the Northeast when MJO convection is in the western Indian Ocean (phase 1), the Southern Great Plains when MJO convection is in the central Indian Ocean (phase 2), and the Northern Great Plains when MJO convection is over the Maritime Continent (phase 4). These results are primarily driven by the forcing of the MJO on the jet stream and its subsequent impacts on geopotential height, as well as temperature and wind in the lower atmosphere. MJO impacts on the frequency of extreme winter weather can be seen up to 15 days in advance, suggesting utility in S2S forecasting.

Between 26 to 28 October 2020, an anomalous, high-impact, and early-season ice storm affected much of Oklahoma. This was the first time the National Weather Service (NWS) Forecast Offices in both Norman and Tulsa had issued an ice storm warning during the month of October and the event is the earliest ice storm in the last 40 years. Overall accumulations were consistent with past ice storm events, but the early season nature of this event led to higher impacts due to increased surface area for accumulations on trees that retained leaves from the growing season. This resulted in branch failures, many downed powerlines, and widespread power outages. Due to the early-timing and severe impacts the goal of this study is to investigate the evolution of the event and identify critical physical processes. First, the synoptic-scale and mesoscale features associated with the event were examined. At the synoptic-scale, a 500 hPa Alaskan ridge, ample 700 hPa moisture transport from the eastern Pacific Ocean region, and the progression of cold 850 hPa temperature anomalies from Canada were all noted 14 days before ice storm onset. At the mesoscale, deep-tropospheric ascent contributed to the significant, localized heavy precipitation across central Oklahoma that was collocated with an anomalous, shallow cold air mass. Second, the October 2020 ice storm was compared to 12 past ice storms in Oklahoma that occurred between 1996 and 2017 to investigate whether the synoptic-scale patterns of an early-season storm differ from normal winter season events. The results yielded similar overall patterns, however, the magnitude of anomalies was generally greater for October 2020 than other past cases. Finally, a potential predictability signal was identified where the geopotential height pattern of past ice storms exists simultaneously with cold 2-meter temperature anomalies over much of the CONUS spanning up to two weeks before event onset.

Machine learning, deep learning, and other artificial intelligence (AI) methods are becoming popular tools within the meteorological research community. However, despite the breadth of promising AI research and its increasing adoption within operational agencies, expert forecasters are often hesitant to fully embrace this relatively new technology. Operational forecasters have a practiced, expert insight into weather analysis and forecasting but typically lack the time, resources, or guidance for formal research and development due to the daily demands of their jobs. Conversely, many researchers have the resources, theoretical knowledge, and formal experience to solve complex meteorological challenges but may lack a full understanding of operation procedures, needs, requirements, and authority necessary to effectively bridge the research to operations (R2O) gap. To address these challenges and attempt to improve the R2O success of AI-derived products, this research investigates how operational forecasters evaluate new forecast guidance and how their perspectives about the R2O process differ from those of the research community. The results from these investigations are then used to derive a collaborative co-production framework intended to optimize the R2O process while improving researcher-forecaster communication throughout the development cycle. Finally, the benefit of this collaborative co-production framework is demonstrated by applying modern AI techniques in tandem with the expert knowledge of Storm Prediction Center forecasters to develop two new forecast products designed to predict lightning hazards and emulate county-based Severe Thunderstorm and Tornado Watches that dynamically evolve with the predicted time and location of the severe weather threat.