When changes in LULC occur in a growing urban area, changes in the frequency, intensity, and amount of precipitation received can occur (Shepherd 2005). Beyond the mechanisms of precipitation enhancement (UHI, increased aerosols and surface roughness), urban areas have a further, and still poorly understood, effect on precipitation by altering the movement, growth, and demise of individual storm cells. Bornstein and Lin (2000) defined storm bifurcation as “a group of storms [that] move in two directions from a specific location (such as upwind of city)”. This phenomenon differs from storm splitting in that splitting is “a single initial storm [that] splits into two separated supercells, given appropriate vertical wind shear conditions” (Bornstein & Lin 2000, p. 515). While it is possible for storm splitting and bifurcation to occur in multiple types of rainfall events (frontal, convective, tropical), very few studies on the specific phenomenon of storm bifurcation have been undertaken. Even though individual synoptic situations in an urban area are complex, it can be concluded that when regional winds are strong, surface roughness of an urban area dominates the local synoptic regime (as opposed to a UHI-dominated regime in weaker synoptic settings).

The research question

The primary research objective is to find if urban areas across the southeastern United States influence heavy precipitation events (≥25 mm). Synoptic variations in precipitation delivery mechanisms differ from city to city in the southeast, thereby likely contributing to spatial variations in urban-precipitation interactions. Given the frequency with which extreme weather and climate impacts affect this region, an improved understanding of precipitation variability in urban regions and its application to emergency management and hazards analysis is vital. By employing multiple detection methods at multiple locations within the southeastern U.S., this research will provide a new, robust, and comprehensive assessment of urban-precipitation relationships for this part of the United States.

Precipitation enhancement study

Once the selection of urban test sites has been made, three tests will be employed to detect the existence of precipitation enhancement: (1) downwind vs. upwind, (2) temporal analysis, and (3) the contour test (Diem 2008). The downwind vs. upwind test requires mean wind direction for heavy precipitation days to spatially determine locations of upwind and downwind regions. If the downwind region receives more rainfall, it is possible that the precipitation has been enhanced by the previously discussed mechanisms of urbanization. The temporal analysis tests for trends in the inter-arrival times between events using the Poisson Process (Keim & Cruise 1998). If heavy precipitation events are occurring more frequently, then urban influences may be the cause. The contour test uses interpolated precipitation at non-urban stations to estimate precipitation within the urban core. Precipitation received at locations where observed precipitation exceeds the spatially interpolated value are likely candidates for urban enhancement. If these candidates are spatially biased toward the downwind side of the city center, a precipitation maximum may be a result of urban enhancement.

Bifurcation study

The number of cities is restricted to two plus a non-urban location serving as a control site due to the in-depth nature of this portion of the analysis. Radar-derived precipitation estimates will be utilized, in addition 900 hPa flow to represent wind speed and direction across the urban core, to study the spatial characteristics of each event for evidence of an urban signal in the form of bifurcation. It is expected that for bifurcation events, there will be a greater amount of rainfall received in the upwind, periphery, and downwind areas than within the city center. It is also possible that there will be an area of greater rainfall in the downwind region in the event that the storm has rejoined.

Synoptic link study

A circulation-to-environment manual classification (Yarnal 1993) of 500 mb daily composite geopotential height patterns will be conducted for each heavy precipitation event, and the corresponding synoptic-scale pattern will be classified as frontal, tropical (e.g. hurricane, tropical storm, or tropical disturbance), air mass, or transition to link bifurcation to the larger-scale synoptic processes that may either enhance or inhibit its occurrence. This approach will permit a comparative analysis of precipitation anomalies during different synoptic conditions as well as during bifurcation and non-bifurcation rain events.