Ingrid Luffman

Wake-up Call in East Tennessee?: Correlating Flood Losses to National Flood Insurance Program Enrollment (1978-2006)

The National Flood Insurance Program (NFIP) provides federally-backed insurance for properties in Special Flood Hazard Areas, yet many property owners do not enroll in the program. I compared flood losses and flood insurance enrollment for three Tennessee communities: Chattanooga, Elizabethton and Pigeon Forge, to investigate the relationship between flooding and NFIP enrollment. Normalized flood losses and insurance purchases were cross-correlated using lags of zero through nine years to investigate the relationship between flood losses in one year and NFIP enrollment in subsequent years. The correlation between flood losses and NFIP enrollment is significant (r = 0.39 and 0.42 respectively, p<0.05) in the year in which flood losses occurred for Chattanooga and Elizabethton. In Pigeon Forge, flood losses correlate to NFIP enrollment in the following year (r=0.43, p=0.02).

Groundwater Nitrate Concentrations and Its Relation to Landcover, Buncombe County, NC

High concentrations of nitrate (NO3) in groundwater can be harmful to human health if ingested, and the primary cause of blue baby syndrome, among other health impacts. In this study, the spatial distribution of NO3 in groundwater for 610 private drinking water wells in Buncombe County, North Carolina was modeled. While NO3 concentration in the sampled wells did not exceed the 10 mg/L limit established by the United States Environmental Protection Agency, some wells had NO3 concentrations approaching this limit (as high as 8.5 mg/L). Kriging interpolation was implemented within a Geographic Information System to predict NO3 concentrations across the county, and a cokriging model using land cover type. Cross validation statistics of root mean square and root mean square standardized for both models were compared and the results showed that the predicted NO3 map was improved when land cover type was integrated into the model. The cokriging interpolated surface with land cover as a covariate had the lowest root mean square (0.979) when compared to the kriging interpolated surface (0.986), indicating a better fit for the model with land cover. NO3 concentrations equal or greater than 2 mg/L were concentrated in 37% hay/pasture land, 34% developed open space, and 29% deciduous forest. The study did not reveal any statistically significant difference in the presence of high NO3 concentration between these landcover types, indicating they all relate to high NO3 content.

Iron and Manganese in Groundwater: Using Kriging and GIS to Locate High Concentrations in Buncombe County, North Carolina

For health, economic, and aesthetic reasons, allowable concentrations (as suggested by the United States Environmental Protection Agency) of the secondary contaminants iron (Fe) and manganese (Mn) found present in drinking water are 0.3 and 0.05 mg/L, respectively. Water samples taken from private drinking wells in rural communities within Buncombe County, North Carolina contain concentrations of these metals that exceed secondary water quality criteria. This study predicted the spatial distribution of Fe and Mn in the county, and evaluated the effect of site environmental factors (bedrock geology, ground elevation, saprolite thickness, and drinking water well depth) in controlling the variability of Fe and Mn in groundwater. A statistically significant correlation between Fe and Mn concentrations, attributable to bedrock geology, was identified. Prediction models were created using ordinary kriging and cokriging interpolation techniques to estimate the presence of Fe and Mn in groundwater where direct measurements are not possible. This same procedure can be used to estimate the trend of other contaminants in the groundwater in different areas with similar hydrogeological settings.