Date of Degree


Document Type


Degree Name



Earth & Environmental Sciences


James Booth

Committee Members

Z. Johnny Luo

Allan Frei

Catherine Naud

Subject Categories

Atmospheric Sciences | Meteorology


Streamflow, Precipitation, Storm Surge, Cyclones


As global temperatures continue to rise, the effects of anthropogenic climate change will impact the magnitude and frequency of pluvial, fluvial, and coastal flooding events. If we want to accurately predict changes in these flooding events, we need to fully understand them in the current climate. As such, this research examines the relationship between hydrometeorological hazards and the characteristics of the storm types, such as extratropical cyclones (ETCs) and tropical cyclones (TCs), that produce such hazards. Through the use of observational and reanalysis data, the work herein utilizes a cyclone-hazard association algorithm and extreme value analysis to assess the extent, probabilities, and causes of hydrometeorological hazards.

Chapter 2 examines the characteristics of ETCs and TCs and the large-scale atmospheric circulation patterns that resulted in the top 100 basin-scale 1-day precipitation, multi-day precipitation, and 1-day streamflow events from 1950 to 2012 in the Catskill Mountains. Of the top 100 precipitation and streamflow events, over 70% are associated with ETCs. These ETC-related events are more likely to take a more meridional path during the cool season and a more zonal path during the warm season. While TCs generate a much smaller number of events, TCs that pass over the region are more likely to generate a top 100 event than ETCs. ETCs associated with precipitation events have relatively less moisture associated with them compared to TCs, but stronger forcing for rising motion. Due in part to TCs, heavy precipitation events occur more often in the warm season whereas high streamflow events occur mainly in the cool season. Despite this difference, there is an overlap among approximately 43% of the top 100 events across all three hydrometeorological hazards: 1-day precipitation, multi-day precipitation, and 1-day streamflow. These overlapped events represent many of the very strongest events. A differentiating atmospheric feature for high streamflow events that do not also result in heavy precipitation is a larger area of high temperature anomalies that are present at least 48 h in advance of high streamflow events, indicating the possible role of snowmelt in these events.

Chapter 3 analyzes the statistical relationship between storm surge and TC characteristics, including TC proximity, intensity, path angle, and propagation speed, for 12 sites along the east coast of the United States. Storm surge is influenced differently by these TC characteristics, with some locations more strongly influenced by TC intensity and others by TC proximity. When conditional sorting is applied to isolate strong TCs that are close to a site, the correlation for individual and combined TC characteristics increases. For TCs passing within 500 km of a tide gauge, between 6% and 28% of TCs resulted in surge exceeding the 1-year return level at each site. If only the closest and strongest TCs are considered, the percentage of TCs that generate surge exceeding the 1-year return level is between 30% and 70% at sites north of Sewell’s Point, VA, and over 65% at almost all sites south of Charleston, SC. In addition to TC proximity and intensity as influential TC characteristics on storm surge, the location of the TC influences the direction of the winds around the TC center toward the site. When examining the tracks of TCs that do and do not generate surge exceeding the 1-year return level, there is a distinct difference for locations north of Sewell’s Point. While the centers of these TCs generally propagate in a northeasterly direction, the TC center is often located to the southwest of each location at the time of peak surge and very rarely track inland. For more southern locations, there is not a clear distinction between TC tracks as there seems to be a greater dependence on TC intensity.

Chapter 4 assesses the influence of the atmospheric dynamics and frequency of ETCs and TCs along the east coast of the United States in the calculation of storm tide exceedance probabilities. Observational water level data combined with a historical dataset of TCs and ETCs are utilized to highlight the differences in storm tide return levels for TC- and ETC-related storm tides for New York, NY; Sewell’s Point, VA; and Charleston, SC. Before the storm tide analysis can be applied, the secular trends and annual cycles of the water levels must be assessed. Since 1950, sea level has risen at all sites in this analysis. The annual cycle of sea level peaks in early Fall when it coincides with the warmest ocean temperatures. Three sensitivity analyses are performed in our calculation of storm tide exceedance probabilities. We first examine variations in the exceedance threshold used when fitting TC-related storm tides. The number of events that exceed the threshold will influence how well the distribution fits to the data, from which we determine the top 30 events to provide the best fit, by balancing trade-offs between variance and bias. The second sensitivity analysis examines the influence of separating storm tide by cyclone type in calculating return levels. For New York and Charleston, low frequency TC-related storm tides are underestimated compared to the total storm tide, whereas higher frequency TC-related storm tides are overestimated compared to the total storm tide. Lastly, the influence of the most extreme storm tides in the calculation of return levels are assessed for TC- and ETC-related storm tides. The most extreme TC-related storm tide had the greatest impact on return levels at long return periods for New York and Charleston. ETC-related storm tide exceedance probabilities were not as strongly influenced by the removal of the most extreme events, with the variance at longer return periods greatest in New York and diminishing as latitude decreased.

The results from these chapters highlight the ability to infer statistical relationships between hydrometeorological hazards and cyclone characteristics. While no two events result in the same exact outcome, these general relationships can be utilized in weather forecasts as a first step in assessing the likelihood of future flood occurrences. Medium-range forecasts from numerical weather prediction models are likely to have smaller biases in the representation of synoptic-scale features and the mean sea level pressure distribution compared to the location and magnitude of hydrometeorological hazards. The utilization of extreme value analysis provides an assessment of the risk associated with these hazards. Understanding the probability of a cyclone resulting in a hazard based on whether the intensity of the cyclone exceeds some threshold, or the cyclone path moves in a certain direction aids forecasters in preparing the public for these types of events. Such documentation is especially important given that some of the factors that affect the magnitude of hydrometeorological hazards, such as rising sea levels and increasing storm intensities, are likely to become worse in the future as a result of anthropogenic climate change.