Introduction

Floodplain ecosystems are very sensitive to perturbations in the hydrologic regime, so that human impacts such as the regulation of river discharges may greatly effect the terrestrial ecology of floodplain habitats. Specifically, vegetation dynamics on floodplains are closely associated with the hydroperiod regime, which is defined as the extent, duration, and timing of inundation on a floodplain (Sharitz and Mitsch 1993). Species composition, productivity, and regeneration vary in response to complex interactions between river flow and floodplain geomorphology. However, hydroperiod regimes are neither consistent from year to year nor regular over the course of one year. These inconsistencies are of crucial importance to the ecology of the floodplain, because the timing of floods and hydrologic droughts affects seedling establishment and survivorship (Visser and Sasser 1995). Therefore, an understanding of the seasonal timing and expected recurrence of floods and hydrologic droughts is essential to the analysis of the ecology of the floodplain. These factors have become increasingly important to many floodplain ecosystems such as the Roanoke River floodplain in eastern North Carolina, where the construction of a series of dams starting in 1952 has initiated a regulated flow regime to the lower river.

Townsend and Walsh (1997) developed a digital elevation model (0.5-1 m vertical resolution) and derived a series of topographic surfaces to represent the potential for inundation across the floodplain. These models were evaluated using information derived from the analysis of satellite imagery from the microwave sensors, JERS-1 and Radarsat (Townsend and Walsh 1997, Townsend 1997). A series of SAR images was employed to represent the change in inundation across the floodplain during the course of a hydrologic year (Townsend 1997). For each image, the hydrologic record of the lower Roanoke River was examined to determine the frequency of occurrence during the post-damming period for the conditions represented in the image. The results were used in conjunction with the DEM to develop a model of a simulated hydroperiod regime for a hypothetical growing season which experienced the conditions represented by the SAR images (Townsend 1997). In another study, Konrad (1997) examined the hydroclimatology of the floodplain, and found that winter/spring flooding periods have been reduced in intensity, but extended in time. Additional analyses indicate that discharge levels in summer months have been reduced below pre-dam levels, but that the absolute minimum discharges from the pre-dam period no longer occur. In general, variance in discharge patterns has been reduced. This suggests substantial changes in the spatial and temporal pattern of flooding on the Roanoke.

The Roanoke River Bioreserve includes some of the best remaining examples of large river alluvial ecosystems on the eastern coast. Dams were constructed upstream along the piedmont portion of the river in the 1950s, and these have caused changes in the hydrology of the system. Specifically, water flow has been leveled out across the year so that there is less time with either very low or very high water levels, and more time with moderately high water (Konrad 1997).