Introduction

Hurricane Katrina was a Category 4 storm on the Saffir-Simpson scale when it made landfall near New Orleans, Louisiana on Monday, August 29, 2005. The storm surge was estimated at ~3-3.5 m near the city of New Orleans. Hours after the passage of the storm, several levee breaks occurred in Orleans Parish, portions of which lie below sea level. This resulted in water levels in excess of 3 m in some portions of the city of New Orleans accompanied by massive damage to property and loss of life. Many residents and first-responders were exposed to the floodwater over the course of several days and serious concerns about potential pollutants in the floodwater have been expressed in various forums. These include concern over toxic chemical constituents, potentially pathogenic microorganisms, and the attendant public health implications resulting from exposure to these waters.

Potential sources of pollutants following flooding include sewage leaking from the wastewater collection system, septic tanks, and wastewater treatment plants, gasoline leaking from submerged vehicles and fuel stations, chemicals leaching from industrial facilities, businesses, and submerged homes, and decaying vegetation and other organic debris. Water quality impacts have been reported for previous hurricanes but have primarily been limited to depleted dissolved oxygen (DO) due to elevated BOD loading (1, 2), increases in fecal coliforms (3, 4), and transient increases in nutrient loading (1, 2). Water quality effects are storm-dependent. For example, while Hurricane Fran (1996) and Hurricane Bonnie (1998) in North Carolina resulted in depleted DO and massive fish kills (1-4), Hurricane Floyd (1999), a much larger storm, produced milder impacts presumably due to dilution by the large volumes of resultant floodwater (3). Unlike these storms, Hurricane Katrina flooded a major urban area and different types of chemical and microbial hazards are associated with this event. The closest analogy may be the "Prague" flood of 2002, the massive flooding of the Elbe River basin that resulted in release of dibenzodioxins and furans to the watershed due to flooding of the Spolana chemical plant (5).

The objective of the current study was to present a snapshot of the chemical and microbiological characteristics of the water immediately following the flood caused by Hurricane Katrina. The focus of the initial effort was to collect as complete a dataset of chemical and microbial constituents as possible that encompassed street-to-street variability and sampling across the depth of the floodwater column. Samples were collected in the main area of the city of New Orleans, the "East Bank", where primary human contact with the floodwaters occurred during rescue operations. Samples were also collected from the 17th Street drainage canal, before and after pumping began, to facilitate assessment of potential impacts on Lake Pontchartrain, the receiving body for floodwaters pumped from the city of New Orleans.

Materials and Methods

Sampling. Floodwater samples were collected from two Orleans Parish "East Bank" locations: in the West End and Lakeview neighborhoods of the Lakeview district on Saturday, Sept. 3, 2005 and in the Tulane-Gravier neighborhood of the Mid-City district on Wednesday, Sept. 7, 2005 (Figure 1). Sampling locations in the Lakeview district are denoted by points numbered 1-21 in Figure 1, and sampling locations in the Mid-City district are denoted by points labeled as NO-1-NO-9. The Lakeview district is bound by Lake Pontchartrain to the north and the 17th Street Canal to the west. The Lakeview district is composed of newer residential neighborhoods and is near one of the larger canal breeches at the 17th Street Canal. It was selected for sampling strictly based on accessibility and safety issues immediately after the flood. Once these analyses were completed, a second area was targeted for sampling that would provide contrast to the Lakeview district with respect to land use, contact time with the floodwaters, and distance from the levee breaks. The Tulane-Gravier neighborhood in the Mid-City district was selected to provide this contrast. It is a mixed-use area that includes low-income residential, light industry, and a health-care education complex. It is located 6-8 miles from any of the levee breaks and possesses some of the lowest elevations within New Orleans' bowl-shaped topography. Both the Lakeview and Mid-City locations are located within the leveed "East Bank" of Orleans Parish. This contiguous urban area had the highest population of the flooded areas of the city and had the greatest number of residents exposed to the floodwater. Additional samples for some constituents were collected in Lake Pontchartrain on Saturday, September 3 for comparison purposes. Sampling locations in Lake Pontchartrain are labeled as LP-1-LP-3 in Figure 1.

Figure 1 Top: Overview of New Orleans sampling sites. Bottom left: Lakeview district sampling sites. Bottom right: Mid-City sampling sites.

Samples were obtained using a flat-bottomed boat equipped with a small outboard motor. Samples were collected at street intersections using the cross-street patterns as a makeshift grid. When approaching an intersection targeted for sampling, the motor was turned off and the boat was allowed to drift in, minimizing mixing of water that was collected. For the Lakeview sampling event, surface samples were collected. For the Mid-City sampling event, surface and bottom samples were collected. Bottom samples were obtained using a low-flow peristaltic pump with collection tubing placed on the bottom of the floodwater column at street level. These data were supplemented by dissolved oxygen profiles, fecal coliforms, and chemical oxygen demand (COD) sampling along the 17th Street Canal (Figure 1 points labeled as C1-C15), the location of a major levee breach and, after the levee repair, one of the routes of floodwater discharge into Lake Pontchartrain. Samples collected along the length of the 17th Street Canal on Sept. 6, 11, and 13, 2005, were obtained by surface sampling along the bank and by boat. GPS data were recorded for all sampling locations. Data presented in this paper met all quality control criteria and sample preservation requirements described in the referenced methods.

Field Parameters (DO, pH, ORP, Conductivity, Tem perature, Turbidity, and Alkalinity). Dissolved oxygen was measured in the field by titration using the Winkler method. Temperature, conductivity, oxidation-reduction potential (ORP), and pH were measured using a calibrated Myron L Ultrameter. Alkalinity was measured immediately in the field using a calibrated sulfuric acid titrant with bromocresol green as an indicator. Turbidity was measured by a calibrated HACH 2100 A turbidity meter (Hach, Loveland, CO).

Organic Constituents (COD, Volatiles, and Semivolatiles). COD was measured by the closed reflux colorimetric method using Hach 8000 reaction vials according to the manufacturer's recommended protocol. Semivolatile organ ics were measured using EPA Method 625 by Pace Laboratories (Kansas City, KS). Volatile organic compounds were measured using EPA Method 624 by Pace Laboratories. In addition, separate analyses were performed to search for volatiles not included on the Method 624 list. For these analyses, an EPA Method 624 technique was applied to 5-mL samples on an Agilent 6890/5973N GC-MS operating in full-scan mode. Spectra from unknown peaks were compared with NIST library spectra and identifications were assigned based on comparisons between observed retention times and expected retention times and match quality statistics computed by the software from comparisons between NIST spectra and spectra obtained from the unknowns. Concentrations were estimated using total ion responses from the unknowns and the response of the known internal standard, fluorobenzene, using a response factor of unity.

Metals. Metals (Pb, As, Cu, Cr, Zn, Ni, and Cd) were analyzed using ICP-MS (Perkin-Elmer ELAN 9000) using EPA Methods 200.8 and 6020.

Nitrogen and Phosphorus. Total Kjeldahl nitrogen (TKN) was measured using EPA Method 351.4. Total phosphorus was measured using EPA Method 365.1.

Oxygen Uptake Rate. Oxygen uptake rate (OUR) was measured by Standard Method 2710B (6) using a YSI oxygen-sensitive electrode in the field.

Microbial Parameters. Fecal coliforms were enumerated using a most probable number (MPN) technique (Standard Method 9221 (6)).

Data Reduction and Analysis. Statistical comparisons were performed on selected data sets after confirming normality (Shapiro-Wilk test) and equality of variances (F test). When either normality or the equality of variance assumption was violated, data were log-transformed and assumptions were retested prior to performing the comparison. When comparing Lakeview district surface floodwater parameters to Mid-City district surface floodwater parameters, t tests (difference between means) were utilized. When comparing Mid-City district surface floodwater to bottom floodwater, paired t tests were utilized. All comparisons were performed at = 0.05.

Results

Basic Water Chemistry. Surface floodwater from the Lakeview district neighborhoods was near neutral in pH with an average of 6.94 and a range from 6.71 to 7.24 (Table 1). Alkalinity ranged from 63.6 to 286.4 mg/L as CaCO₃ with an average of 153.9 mg/L as CaCO₃. Conductivity in the Lakeview area averaged 9.8 mS/cm compared with 42.9 mS/cm for full-strength seawater at 15 C and standard pressure (7). Floodwater temperatures closely followed air temperatures during sampling, ranging from 29.4 to 33.1 C.

Four days later in the Mid-City district, surface floodwater pH values were slightly more acidic with an average of 6.82 and a range of 5.92-7.29 (Table 1). Alkalinity averaged 94.0 mg/L as CaCO₃, ranging from 61.2 to 114 mg/L as CaCO₃. Water temperatures were again high, averaging 30.8 C. In the Mid-City district, surface floodwater was more turbid, averaging 11.0 NTU versus 6.27 NTU at the Lakeview district.

In the Mid-City district, water at the bottom of the floodwater column had pH (average = 6.82), temperature (average = 30.7 C), alkalinity (average = 92.9 mg/L as CaCO₃), and turbidity (11.6 NTU) values that were not statistically different from the surface floodwater at the same location (Table 1). ORP values were statistically lower in the bottom samples (-35.9 mV versus 26.6 in the surface water) (paired t test at = 0.05).

Dissolved Oxygen and Oxygen Demand. At both Lakeview and Mid-City locations, dissolved oxygen (DO) concentra tions in the surface floodwater were below saturation. At Lakeview, DO measured at the surface ranged from below detection (<0.02 mg/L) to 3.3 mg/L with an average of 1.6 mg/L (Table 1). At Mid-City, DO concentrations in surface floodwater ranged from 2.8 to 7.5 mg/L with an average of 4.8 mg/L (Table 1). Higher dissolved oxygen in the Mid-City location is attributed to the large amount of boat and helicopter traffic associated with search and rescue activities on September 7th.

DO concentrations of water sampled from the bottom of the floodwater column were all below detection (<0.02 mg/L) with the exception of location NO-1, where the DO concentration was 1.3 mg/L (Table 1). Lower DO concentra tions at the bottom of the floodwater were expected due to the floodwater oxygen demand and limited reaeration potential at these depths, which averaged slightly over 1 m (Table 1). Differences in DO between surface and bottom samples were well correlated with more negative ORPs at depth.

COD in samples collected at Mid-City locations on Sept. 7, 2005, ranged from 33.9 to 104 mg/L (average = 79.9 mg/L) for the surface samples and from 37.9 to 121 mg/L (average = 76.9 mg/L) for the bottom samples (Table 2). COD concentrations were not statistically different between the surface and bottom of the floodwater column at the Mid-City district.

Nitrogen and Phosphorus. Total phosphorus (TP) concentrations measured in surface floodwater samples collected at Mid-City on Sept. 7, 2005, ranged from 0.30 to 0.59 mg/L and averaged 0.39 mg/L (Table 2). Concentrations of TP collected from the bottom of the floodwater column at Mid-City ranged from 0.28 to 0.91 mg/L, averaging 0.49 mg/L (Table 2). TP concentration in the bottom floodwater was statistically greater than that in surface floodwater at Mid-City. This may reflect release of phosphate from dissolution of iron oxyhydroxide minerals under anaerobic conditions in the bottom floodwater. While the phosphorus concentra tion exceeded the U.S. EPA desired goal to prevent nuisance plant growth in surface water bodies (0.1 mg/L), the values measured in floodwater samples are within the range of TP concentrations measured in urban streams in the Acadian-Pontchartrain Drainages during the time interval of 1999-2001, where the median concentration was 0.45 mg/L (8). As an additional point of reference, the total phosphorus concentration measured in the Mississippi River near St. Francisville, LA, ranged from 0.10 to 0.45 mg/L during the period of 1998-2001 (8).

Total Kjeldahl nitrogen (TKN, the sum of ammonia and organic nitrogen) concentrations measured in surface floodwater samples collected at Mid-City on Sept. 7, 2005, ranged from 1.2 to 8.1 mg/L and averaged 2.5 mg/L (Table 2). Concentrations of TKN collected from the bottom of the floodwater column at Mid-City were very similar, ranging from 1.0 to 4.4 mg/L and averaging 2.4 mg/L (Table 2). The total nitrogen (TN, the sum of all forms of nitrogen) concentration in the Mississippi River at the same location ranged from 1.2 to 3.7 mg/L with a median value of 2.1 mg/L during this interval (8). TN concentrations measured in urban streams in the Acadian-Pontchartrain Drainages during the time interval of 1999-2001 had a median concentration of 1.6 mg/L, with the range of the 10th and 90th percentiles spanning from 0.4 to 3.0 mg/L (8). Nitrate and nitrite concentrations, though not measured in New Orleans floodwater samples, were likely quite low given the low DO concentrations (Table 1). Thus, TKN values may be comparable to the TN concentrations in the floodwaters.

Volatile and Semi-Volatile Organic Compounds. In the Lakeview district, volatile organic compounds (VOCs) detected in the surface floodwater samples consisted of only gasoline components, namely benzene, toluene, and ethylbenzene (Table 3). Xylenes are not on the EPA Method 624 list of analytes. A visible sheen was noted at many of the sampling locations and this sheen was captured in the surface grab samples. The maximum concentration was observed at location 18, where total VOC concentrations were 30 g/L. Only benzene exceeded the U.S. EPA drinking water maximum contaminant level (MCL, 5 g/L) at location 18. VOC concentrations are 2-3 orders of magnitude higher than those previously reported for stormwater runoff (9).

In the Mid-City area, a similar profile of VOCs was observed. Benzene, toluene, and ethylbenzene were detected in concentrations in the surface floodwater similar to that observed in the Lakeview district (Table 3). Concentrations of VOCs were detected in only 2 of 9 surface floodwater samples (22% of samples taken) in Mid-City, while VOCs were detected in 5 of 21 samples (23% of samples taken) in the Lakeview district.

In the bottom samples from the Mid-City district, methylene chloride, a common chlorinated solvent, was detected at two locations in addition to the gasoline components detected in the surface floodwater samples (Table 3). In addition to the volatiles on the EPA Method 624 list, separate analyses were conducted to search for unknowns. A number of compounds were detected (Table 4) but in concentrations estimated to be of the same order of magnitude (g/L) as the data quantified in Table 3. These compounds included reduced sulfur species (dimethyl sulfide and dimethyl disulfide) associated with the decay of organic matter under anaerobic conditions, and methanethiol, which is placed in natural gas and propane as an odorant to warn of leaks. A series of compounds were also detected that are associated with household chemicals. 2-Methyl, 1-propanol is used in aerosol paints, insecticides, and a variety of other household products. Butyraldehyde is used in caulking compounds and sealants. Xylenes are gasoline components. 2-Propenoic acid, 2 methyl-, methyl ester is found in synthetic resins and rubber adhesives. All of these compounds would logically be expected at the sites sampled.

Only 1 semivolatile compound was detected in Lakeview surface floodwater, at location 21 (5.2 g/L of bis(2-ethylhexyl) phthalate, a widely used plasticizer) (Table 5). Similarly, in the Mid-City floodwater, few semivolatiles were detected (Table 5). Bis(2-ethylhexyl) phthalate, the compound detected in Lakeview, was observed at 3 locations at g/L levels. Phenol was also detected at one location with a concentration of 7.3 g/L.

Metals. With the exception of arsenic, metals concentra tions in surface floodwater sampled from the Lakeview district did not exceed MCLs for drinking water (Table 6). Concentrations of lead (mean = 3.2 g/L), for example, were less than the 15 g/L MCL but within the range (1.5-58.0 g/L) reported for a series of recent stormwater runoff events in Louisiana (10). Lead concentrations exceeded the range of dissolved concentrations in the Mississippi River (0.1-0.5 g/L) (11). Arsenic concentrations (mean = 30 g/L) consistently exceeded the MCL of 10 g/L.

Surface floodwater metals concentrations in the Mid-City district were significantly higher than those in the Lakeview district for all metals except Cu ( = 0.05) (Table 6). For example, concentrations of Pb were approximately an order of magnitude higher in the Mid-City surface floodwater samples (mean = 28 g/L compared with 3.2 g/L at Lakeview). Several factors could contribute to the increased metals concentrations observed here including a longer period of contact with the floodwaters in the Mid-City district and differences in the land-use profiles of the two neighbor hoods sampled.

Concentrations of metal in the floodwater at the bottom of the water column at Mid-City were very similar in magnitude to those of the surface floodwater (Table 6). Concentrations of Ni, Cr, and As were statistically higher in the bottom floodwater, but differences did not exceed 15% of the total concentration. Overall in the Mid-City district, MCLs were consistently exceeded for As and Pb, and individual sampling locations exceeded the MCL for Cr (100 g/L). Concentrations of Zn, Cd, and Cu were comparable to the ranges reported in stormwater runoff in recent Louisiana events (10). Pb concentrations were higher than stormwater ranges (10), Mississippi River water (11), and the MCL.

Microbial Constituents. Fecal coliform concentrations at all locations in the surface floodwater from the Lakeview and Mid-City districts were at least 1 order of magnitude higher than the water quality standard for primary contact of 200 MPN/100 mL (Table 7). The mean concentrations at Lakeview (n = 7) and Mid-City (n = 9) were on the order of 10⁵ MPN/100 mL, and the median concentration at both locations was on the order of 10⁴ MPN/100 mL. This is considerably higher than the concentration of 2.4 × 10³ MPN/100 mL (95% confidence interval of 1.0 × 10³ to 9.4 × 10³ MPN/100 mL) measured in Lake Pontchartrain (sample location LP-2 in Figure 1) offshore from the University of New Orleans Research and Technology Park on September 3, 2005, prior to the start of pumping to remove wastewater from the city.

Presence of high concentrations of fecal coliforms, an indicator of fecal contamination by warm-blooded mammals, serves as a general indicator of the presence of potential human pathogens in the floodwaters. Thus, it is possible that rescue workers, emergency response personnel, and stranded population remaining in New Orleans during and after the flood were exposed to human pathogens.

Prior to flooding by Hurricane Katrina, urban stormwater runoff from the New Orleans metropolitan area has histori cally contained high concentrations of fecal coliforms, and a state advisory to avoid swimming and other primary contact sports has been in effect for the south shore of Lake Pontchartrain since 1985 (12). Fecal coliform concentrations reported here for the floodwater following Hurricane Katrina are generally within the range (6.0 × 10³ to 1.6 × 10⁵) reported by McCorquodale et al. (12) for stormwater discharges from New Orleans drainage canals to Lake Pontchartrain during normal wet weather flow (i.e., nonflooding events). High fecal coliform concentrations in drainage canals during normal wet weather flows has been attributed to sanitary sewer cross-flows with the stormwater collection system.

Floodwater Pumped into Lake Pontchartrain. On days prior to the start of pumping to remove floodwater from New Orleans, the DO concentration measured at the surface of water in the 17th Street Canal was below 0.5 mg/L on either side of the idle #6 pumping station. After the start of pumping on September 5th, the DO measured in the canal began to increase. A plot of dissolved oxygen concentration as a function of position in the 17th Street Canal on Sept. 6, 11, and 13, 2005, (Figure 2) revealed considerable spatial variability in DO following the start of pumping to drain floodwaters from the city. It is evident from the data that the pumping stations impart sufficient turbulence and mixing to aerate the water. DO concentrations decreased at locations downstream from pumping locations, consistent with microbial consumption of biodegradable organics. This is also consistent with the fact that oxygen uptake rate (OUR) measured in water sampled from the 17th Street Canal on Sept. 14, 2005, at a location close to the point of discharge into Lake Pontchartrain (Figure 1, location C7), was 0.83 mg/(L hr).

Figure 2 Dissolved oxygen concentration measured at various distances along the length of the 17th Street drainage canal.

The COD concentration of 12 samples collected from various locations along the length of the 17th Street Canal on Sept. 13, 2005 ranged from 50 to 180 mg/L, with an average of 72.5 mg/L, close to the average measured in the Lakeview and Mid-City neighborhoods on Sept. 3 and 7, 2005, respectively. The concentration of fecal coliforms measured in a surface sample collected on Sept. 14, 2005, from the 17th Street Canal near the outlet into Lake Pontchartrain (Figure 1, location C7) was 1.30 × 10⁴ MPN/100 mL (95% confidence interval of 5.0 × 10³ to 3.9 × 10⁴ MPN/100 mL), within the range measured in the Lakeview and Mid-City neighborhoods sampled on Sept. 3 and 7, 2005, respectively.

Discussion

Data indicate that Hurricane Katrina floodwater in these two neighborhood "snapshots" is a brackish, well-buffered water with low concentrations of volatile and semivolatile organic pollutants. Chemical oxygen demand and nutrients are elevated, but similar in magnitude to those of normal stormwater. Several metals, including Pb, As, and Cr, approached, and in some cases exceeded, drinking water standards. Fecal coliforms were very elevated compared with primary contact water standards but, again, were similar in magnitude to what is found in typical stormwater runoff from the area. Collectively, these data indicate that Katrina floodwater is similar to normal stormwater runoff but with somewhat elevated Pb and VOC concentrations. What distinguishes Katrina floodwaters are their large volume and the human exposure to these pollutants that accompanied the flood rather than extremely elevated concentrations of toxic pollutants.

The absence of a major toxic source can be explained by several factors. First, the dilution provided by the estimated 100 000 acre-feet (U.S. Army Corps of Engineers official estimate) of floodwater within the East Bank of New Orleans served to lower the observed concentrations. The leveed areas of Orleans Parish studied here do not include any large chemical plants or refineries with major chemical storage which are located to the south and east of the city. Gasoline, the major organic source located within the breached area, was in very short supply as the hurricane approached and the majority of residents evacuated. Although thousands of cars and lead-acid batteries were submerged, the floodwater had enough alkalinity to resist drops in pH that would have mobilized more metals.

Another important factor is the role of mass transfer. A lasting image of Hurricane Katrina media coverage is the aerial shots of people wading through iridescent sheens of fuel in the affected areas. Our study measured only low concentrations (in the range of <1 to 19 ppb) of benzene, toluene, and ethylbenzene even in places where there was a visible oil sheen. Small leaks of gasoline released below the surface rise quickly to produce an iridescent sheen that may consist of surface films with thickness in the micrometer to millimeter range depending on the rate of entry. Despite the apparent large extent of floodwater surface coverage by these films, the initial mass fraction ratio of hydrocarbons-to-water was likely much lower than 1%. The total cumulative mass fractions of benzene, toluene, and ethylbenzene in gasoline are typically in the 10-15% range. Given these factors and assuming a 1-mm hydrocarbon layer in 3 m of floodwater, the mass ratios of these constituents would be in the range of 33-50 ppm initially and distributed in both the nonaqueous (sheen) and aqueous phases. Several chemodynamic processes in the floodwaters, including chemical reactive degradation, adsorption to suspended solids and impervious surfaces, and evaporation to air, continually operate to further reduce these levels. Of these fate mechanisms, only evaporation will be considered further here.

The evaporation chemodynamics of both the soluble and sheen fractions is expected to be very rapid. For surface water low in total suspended solids and of depth h (m), the evaporation half-life of chemicals can be estimated by eq 1

Presence of a hydrocarbon sheen on the floodwater surface complicates modeling of the evaporation process, but in the case of the BTEX fraction, the effect is likely small. The films are very thin but contribute significantly to the mass retrieval in water grab samples. On the basis of laboratory experiments using thin oil layers, h_o (mm), the 1st order evaporation rate constant can be estimated by eq 2

In addition to concerns over direct human exposure to floodwater, the effect of floodwater discharge on aquatic life in Lake Pontchartrain is also of concern. On the basis of our results, two potential adverse impacts should be considered: (1) the depletion of DO in the lake as a result of oxygen demand exerted by the biodegradation of organic materials entering the lake, and (2) the potential for metal toxicity to fish. Although the COD levels measured in the floodwater are lower than those of typical municipal wastewater (14) and stormwater (15), it is not clear from the data collected what fraction of COD is due to soluble components and what fraction was associated with particulate matter, and it is not clear what fraction is biodegradable. With relatively high turbidity (Table 1) a fraction of the COD was likely due to particulate matter. Regardless of the exact fraction of COD corresponding to particulate matter, it is clear from the OUR and DO measurements that some fraction is biodegradable. Low DO concentration in the floodwater combined with residual biological oxygen demand has the potential to adversely impact aquatic life following discharge in Lake Pontchartrain, the water body to which the floodwater has been pumped. It should be noted however, as depicted in Figure 2, that the pumping operation aerates the water to some degree and allows consumption of at least a portion of the COD prior to discharge in the lake. From this perspective, it is clear that the pumping system provided some degree of treatment prior to floodwater discharge into the lake, but determining adverse effects of low DO within the lake (e.g., fish kills) will require further monitoring.

Metal toxicity to fish can be evaluated conservatively by comparing floodwater metal concentrations with U.S. EPA national recommended water quality criteria (16) for freshwater and saltwater ecosystems (Table 6), recognizing that Lake Pontchartrain is an estuarine system lying between these two endpoints. Average copper and zinc concentrations in both Lakeview and Mid-City floodwater greatly exceed both acute and chronic criteria for freshwater and saltwater ecosystems (Table 6). Concentrations of Pb, Cd, Ni, and As in the Lakeview and Mid-City district exceed some of the criteria, but not all. Clearly, there is the potential for metal toxicity in Lake Pontchartrain resulting from floodwaters pumped into the lake. Long-term bioconcentration and biomagnification of metals in aquatic species is also possible but can only be ascertained from long-term monitoring.

Caution should be exercised when extrapolating the results presented here to other flooded parishes and districts within New Orleans because of differences in land use in these and other areas. These data represent water quality conditions in these neighborhoods between 5 and 9 days after the inundation from the flood, but prior to significant pumping of the floodwater back into Lake Pontchartrain. Importantly, these areas represent locations where many first-responders and residents were exposed to floodwaters. Although some conclusions can be drawn about the quality of the floodwater based on this data set, more detailed human exposure, waste load allocation, and ecological risk assessment calculations for Lake Pontchartrain should be conducted prior to reaching ultimate conclusions regarding the environmental impacts of Hurricane Katrina.

Acknowledgment

We acknowledge Louisiana Water Resources Research Institute for support and Pace Laboratories who provided tremendous service for samples on the organic analytes under difficult circumstances. We also acknowledge the Louisiana State University Police Department for security and CK Associates for field support.