Impact of hurricanes on the flux of rainwater and Cape Fear River water dissolved organic carbon to Long Bay, southeastern United States



[1] The hurricane flux of rain and river water dissolved organic carbon (DOC) to Long Bay located on the southeastern coast of the United States was determined for four hurricanes that made landfall in the Cape Fear region of North Carolina. Riverine flux of DOC following hurricanes Fran (1996) and Floyd (1999) represented one third and one half of the entire annual river flux of DOC to Long Bay, respectively. The majority of this DOC was recalcitrant and not available for biological consumption. The high flux of DOC from hurricane Floyd resulted from extremely high precipitation amounts (in excess of 50 cm) associated with the hurricane and subsequent flooding. High riverine DOC fluxes were observed following hurricane Fran but not hurricanes Bertha (1996) and Bonnie (1998). The westerly path of Fran deposited rain inland along the Cape Fear River watershed, causing high river flow conditions, while Bonnie and Bertha took an eastern path, resulting in a minimal effect to the Cape Fear River flow rates. The rainwater flux of total DOC to Long Bay from the four hurricanes was not as dramatic as that observed for riverine fluxes. However, unlike river water DOC that is refractory, rainwater DOC is highly labile. Rainwater from the four hurricanes in this study deposited 2–5 times the DOC deposited in an average storm. This represented a flux of 3–9% of the entire annual budget of bioavailable DOC to Long Bay being deposited over a 1 or 2 day period, likely spurring short-term secondary productivity following the hurricanes.

1. Introduction

[2] Oceanic dissolved organic carbon (DOC) is the largest reservoir of surficial organic carbon on earth [Menzel, 1974; Williams, 1975; Duce and Duursma, 1977; Mopper and Degens, 1979; Hedges, 1992]. The most significant external sources of this oceanic carbon include both rainwater [Willey et al., 2000, and references therein] and river water [e.g., Mantoura and Woodward, 1983; Hedges, 1992]. Rivers export approximately 0.2 × 1015 g C yr−1of DOC to the oceans each year [Schlesinger and Melack, 1981; Meybeck, 1982; Ittekkot, 1988]. The refractory nature of this riverine DOC often results in conservative mixing curves between river water and seawater [Mantoura and Woodward, 1983; Aminot et al., 1990; Avery et al., 2003] suggesting a direct injection of river-derived DOC to coastal waters. Rainwater DOC flux to coastal waters is approximately the same order of magnitude as riverine DOC inputs [Willey et al., 2000]. The majority of this atmospherically deposited DOC is bioavailable and likely contributes to secondary productivity in the coastal oceans [Avery et al., 2003].

[3] DOC fluxes from rivers and rainwater to the oceans are typically based on average annual river flow rates and rainfall amounts. However, catastrophic events such as hurricanes often inject large volumes of rain into watersheds over relatively short time periods. The combination of high rainfall and the resulting high river flow suggests that these extreme weather events have the potential to deposit large quantities of DOC into the coastal ocean during short time periods. Despite the potential importance of these events on oceanic DOC budgets, the flux of rainwater and river water DOC resulting from hurricane events has not been determined. Four hurricanes made direct landfall in the Cape Fear River region between 1996 and 1998. This provided us with the unique opportunity to quantify the hurricane rain and river water flux of DOC to Long Bay along the southeastern coast of the United States. We report for the first time the impact of hurricanes on DOC deposition to the coastal ocean where these short-term fluxes rival the entire annual inputs of DOC from rain and river water.

2. Methods

2.1. Study Site

[4] In order to compare the flux of rain and river water DOC to coastal waters, the boundaries of a receiving basin must first be identified. For this study, Long Bay was chosen as the study site since it receives the majority of the Cape Fear River estuary flow and it is close to our rain collection site on the University of North Carolina at Wilmington (UNCW) campus (Figure 1). Rainwater collected at the UNCW station and in the adjacent Onslow Bay during a research cruise had very similar composition [Willey and Cahoon, 1991]. Long Bay is roughly 100 km long and extends to the shelf break at about 100 km offshore giving it an approximate area of 10,000 km2.

Figure 1.

(top) Eastern United States seaboard with hurricane storm tracts: squares, Fran, 1996; triangles, Bertha, 1996; crosses, Bonnie, 1998; and diamonds, Floyd, 1999. Shaded area indicates the extent of the Cape Fear River drainage basin. (bottom) Lower Cape Fear River estuary near Wilmington, North Carolina, and sampling locations.

2.2. River Water Collection

[5] The Cape Fear River estuary receives river flow from three branches just above Wilmington, North Carolina (Figure 1). The combined flow of the Cape Fear River and Black River enters from the west and the flow from the North East Cape Fear River enters from the east. The majority of the flow into the estuary, and eventually Long Bay (∼90%), comes from the combined flow of the Cape Fear River and the Black River. River water samples were collected just above the split in the estuary at station Nav near Navassa, North Carolina (for Cape Fear and Black Rivers) and at station NCF6 (for the North East Cape Fear River) (Figure 1).

2.3. Rainwater Collection

[6] The rainwater sampling site was an open area encircled by longleaf pine, wiregrass, and turkey oaks on the UNCW campus (34.23°N and 77.88°W), 8.5 km from the Atlantic Ocean. All rainwater samples were collected on an event basis using Aerochem Metrics (ACM) Model 301 Automatic Sensing Wet/Dry Precipitation Collectors. Rainwater samples for DOC analysis were collected in 4 L Pyrex beakers placed within the high-density polyethylene (HDPE) collector buckets. The sampling beakers and all glassware in contact with rain samples were muffled at 550°C for a minimum of 5 hours prior to use. Sampling beakers were replaced with a dry muffled beaker after each event.

2.4. DOC Analysis

[7] Dissolved organic carbon was determined by high-temperature combustion (HTC) using a Shimadzu TOC 5000 total organic carbon analyzer equipped with an ASI 5000 autosampler (Shimadzu, Kyoto, Japan). Standards were prepared from reagent grade potassium hydrogen phthalate (KHP) in Milli-Q Plus Ultra Pure Water. Samples and standards were acidified to pH 2 with 2 M HCl and sparged with carbon dioxide free carrier gas for 5 min at a flow rate of 125 mL min−1 to remove inorganic carbon prior to injection onto a heated catalyst bed (0.5% Pt on alumina support, 680°C, regular sensitivity). A nondispersive infrared detector measured carbon dioxide gas from the combusted carbon. Each sample was injected 4 times. The relative standard deviation was ≤3%. The detection limit for this instrument is 5 μM. All samples were run in triplicate.

2.5. Bioavailability Experiments

[8] Bioavailability experiments on river water and rainwater were conducted using methods described by Avery et al. [2003]. Initial DOC concentrations were determined on rain and river samples shortly after collection. Samples were placed in the dark and incubated at in situ temperature. Aliquots were removed for DOC analysis at approximately 1 week intervals. DOC consumption was considered complete when constant DOC concentrations were obtained for consecutive measurements. This required between approximately 1 to 5 months. Bioavailable DOC was considered to be the DOC loss during the incubation. River water that was filtered (0.2 μm) and stored in a refrigerator showed no loss in DOC after 1 month, confirming that DOC loss in bioavailability experiments was biological consumption.

3. Results and Discussion

3.1. Rainfall and River Flow Rates

[9] Precipitation amount directly affects the rainwater flux of DOC during a hurricane. Rainfall amount also has an important roll in controlling riverine DOC flux to the coastal ocean by increasing river flow rate and therefore the rate of DOC deposition. Figure 2 shows the average monthly river flow rates for years in which a hurricane occurred and long-term average monthly flow rates for those months. Of the four hurricanes presented in this study, both Fran and Floyd resulted in months with combined (Black + NE Cape Fear + Cape Fear) elevated river flow rates into Long Bay while flow rates for the months in which Bertha and Bonnie occurred were indistinguishable from average months. The high river flow rates following Floyd resulted from extremely high precipitation amounts during this hurricane. Rainfall amounts for Floyd were 2 to 4 times that of any other hurricane presented in this study (Table 1). High river flow conditions were observed after Fran, which had a relatively low amount of precipitation, while hurricane Bonnie had no impact on river flow rates even though it had higher precipitation amounts. This seeming disparity can be explained by the storm tracks of the hurricanes as well as the amount of rainfall and degree of ground saturation that occurred during the months preceding the hurricanes. In addition to the amount of rain from each hurricane, Table 1 shows the amount of rain received in Wilmington, North Carolina, during the 2 months prior to each hurricane (U.S. Weather Service). Hurricane Bertha occurred 2 months before Fran, resulting in Fran having the highest rainfall amount in the 2 months preceding any of the hurricanes in this study. The ground was likely more saturated with water prior to Fran than any of the other hurricanes resulting in more runoff and higher river flow rates. Prior to hurricanes Bertha and Bonnie, rainfall amounts were similar to the average for Wilmington (∼25 cm in 2 months).

Figure 2.

Average monthly flow rates for the Black, North East Cape Fear, and Cape Fear Rivers for months in which a hurricane occurred and average long-term monthly flow rates for July and September for the Black River (1952–2000), North East Cape Fear River (1940–2000), and the Cape Fear River (1969–2000) (U.S. Geological Survey river flow data obtained from

Table 1. Precipitation Amounts From Hurricanes From Our Rain Collection Station at University of North Carolina at Wilmington (Bertha, Fran, and Bonnie) and From the National Weather Service Wilmington, North Carolina (Floyd), and Amount of Precipitation in the 2 Months Preceding Each Hurricanea
HurricaneDatePrecipitation Amount, cmPrecipitation Amount in Preceding 2 Months, cm
  • a

    Our rain gauge failed during Floyd. Precipitation amounts for the 2 months preceding each hurricane are from National Weather service, Wilmington, North Carolina.

Bertha12 July 19961128
Fran4 September 19961347
Bonnie26 August 19982420
Floyd15 September 19994932

[10] The track of Fran also contributed to its large impact on the river. While hurricanes Bertha, Bonnie, and Floyd took a more easterly path after making landfall, Fran went inland essentially following a path directly up the Cape Fear River watershed depositing rain upstream (Figure 1). This is evident in the river flow data for the individual river systems entering the Cape Fear Estuary (Figure 2). The river flow following Fran contained a relatively larger proportion of Cape Fear River flow as opposed to Black River and North East Cape Fear River flow, indicating sources upstream of Wilmington. Approximately two thirds of the flow entering the ocean was from the Cape Fear River. Although a significant amount of rain was deposited inland in the upper Cape Fear River watershed, a larger fraction of the river flow following Floyd was from the North East Cape Fear River and the Black River indicating more local watershed sources compared to Fran. Roughly half the flow to Long Bay was from the combined North East Cape Fear River and Black River following Floyd.

3.2. Posthurricane River DOC Flux

3.2.1. Bertha and Bonnie

[11] Following hurricanes Bertha and Bonnie there was an elevation in the flow rate of the North East Cape Fear River compared to its average flow rate for the months in which the hurricanes occurred (Figure 2). The Cape Fear River flow however was slightly lower during these months. As a result, there was essentially no change in the net river flow into Long Bay following hurricanes Bertha and Bonnie and therefore no anomalous flux of DOC to Long Bay following these hurricanes. The flux of DOC into Long Bay for the two weeks following the hurricanes were calculated from the average daily flow rates for the 3 rivers and the average DOC concentrations previously reported for the river water entering the Cape Fear River Estuary (1 mM [Avery et al., 2003]). This calculated value represents the actual flux of DOC to Long Bay since DOC is conservatively mixed in the estuary and there are no large sources or sinks within the estuary [Avery et al., 2003]. The resulting DOC flux to Long Bay for the 2 weeks following the hurricanes represented 3% of the annual flux for hurricane Bertha and 4% of the annual flux for hurricane Bonnie (Figure 3). Assuming a constant river DOC flux, the rivers should deposit 4% of the annual flux in a 2-week period. Therefore there was no elevated river flux of DOC to Long Bay resulting from these two hurricanes.

Figure 3.

Flux of riverine dissolved organic carbon to Long Bay for the 2 weeks following each hurricane. Fluxes were calculated using USGS average daily flow rates for the 14 days following each hurricane and the average DOC concentration for river water entering the Cape Fear River estuary (1 mM [Avery et al., 2003]) for Bertha, Fran, and Bonnie and the average measured DOC concentration of river water entering the Cape Fear River estuary (1.6 mM) for Floyd. Both bioavailable and refractory DOC are shown.

3.2.2. Fran

[12] Cape Fear River flow rates following hurricane Fran were some of the highest ever recorded with values approaching 1300 m3 s−1 or 13 times the average flow rate for September (USGS). Black River and North East Cape Fear River flow rates averaged 200 m3 s−1 each or roughly 15 times the average flow rates for these rivers in September. Using USGS published river flow data, the flux of water for the 2 weeks following Fran was 13.1 × 108 m3 for the Cape Fear River, 2.7 × 108 m3 for the North East Cape Fear River, and 2.8 × 108 m3 for the Black River. The sum of these 3 rivers deposited a total of 18.7 × 108 m3 of water to Long Bay. This represents roughly one third of the total annual river water flux to Long Bay being deposited in only 2 weeks. Using the average DOC concentration for the Cape Fear Estuary headwaters (1 mM [Avery et al., 2003]) and river flow rates presented above, the calculated DOC flux to Long Bay for the 2 weeks following hurricane Fran was 21.0 × 109 g C, or approximately one third of the annual Cape Fear River flux of DOC to Long Bay (Figure 3).

3.2.3. Floyd

[13] Both rainfall amount and Cape Fear River flow were tremendous as a result of Hurricane Floyd. Flow rates along the main branch of the Cape Fear River reached values approaching 1100 m3 s−1, roughly 10 times the average rate for September (USGS). North East Cape Fear River flow rates (800 m3 s−1) and Black River flow rates (750 m3 s−1) were approximately 40 times the average rate for September (USGS). Using USGS published river flow data, we calculate a main branch Cape Fear River water flux during the 14 days following the hurricane of 10.1 × 108 m3, a North East Cape Fear River water flux of 5.0 × 108 m3 and a Black River flux was 4.2 × 108 m3, amounting to a total river flux of 19.3 × 108 m3. As was the case for hurricane Fran, this represents roughly one third the total annual flux of water to Long Bay occurring in only 2 weeks. Cape Fear estuary DOC concentrations were determined 1 week and 2 weeks following hurricane Floyd (see section 3.3). A concentration of 1.2 mM was obtained for the first sampling period and 1.9 mM for the second sampling period. Using the average DOC concentration for our two sampling periods following the hurricane (1.6 mM) and river flow rates presented above, the calculated DOC flux to Long Bay for the 14 days following hurricane Floyd was 36.4 × 109 g C, or approximately one half of the annual river flux of DOC to Long Bay (Figure 3).

3.3. Rainwater DOC Flux

[14] DOC is a major component of rainwater with average values for continental rain of 161 ± 21 μM and 23 ± 3 μM for marine rain [Willey et al., 2000]. DOC values for hurricanes Bertha, Fran, and Bonnie were remarkably similar with a volume-weighted average of 78 ± 2 μM [Willey et al., 2000]. Hurricane Floyd rainwater DOC concentration was 11 ± 3 μM, which was considerably lower than previous hurricane rain and marine rain in general. The lower concentration likely reflects a rainout effect resulting from the exceptionally high volume of rain received from this hurricane (47 cm). The previous three hurricanes, Bertha, Fran, and Bonnie, had rainfall amounts of 11 cm, 13 cm, and 24 cm, respectively (Table 1). The rainwater flux of DOC to Long Bay resulting from each hurricane was calculated by using the DOC concentration and the depth of rain received in Wilmington (Figure 4).

Figure 4.

Rainwater dissolved organic carbon flux to Long Bay from each hurricane, average storm flux and annual storm flux. Flux was calculated using the area of Long Bay (10,000 km), the precipitation amount from each hurricane, and the volume-weighted average DOC concentration for Bertha, Fran, and Bonnie (78 ± 2 μM), and the DOC concentration for Floyd (11 μM). Both refractory and bioavailable DOC are shown.

[15] The rain from the hurricanes reported in this study deposited between 1 to 10% of the annual rainwater flux and 2–5 times the average storm flux of DOC to Long Bay. The hurricane with the largest impact was Bonnie, which had an average hurricane DOC concentration but roughly twice the volume of Bertha and Fran (Figure 4). Bonnie deposited 10% of the annual rainwater DOC flux during a 2-day period. The rainwater flux of DOC deposited in Long Bay as a result of Hurricane Floyd was the lowest (1%) despite the high volume of rain from this storm due to extremely low DOC concentration. In contrast to the posthurricane river DOC flux, the rainwater DOC flux during the hurricane was not as dramatic compared to annual fluxes. Although a great deal of rain was received during the hurricanes the low concentrations of DOC decreased the overall impact of hurricane rain.

3.4. Bioavailable DOC

[16] The bioavailability of rain and river DOC is important since it determines the ultimate fate of the organic material that is deposited. The bioavailable fraction can lead to oxygen depletion and fuel secondary productivity while the recalcitrant fraction acts as a source of organic carbon that can be stored for long periods of time in the oceans. Bioavailabilty experiments were performed on river water following hurricane Floyd. The bioavailability of river water entering the Cape Fear Estuary averages 9 ± 4% (Figure 5). Two DOC decomposition experiments were conducted on the river water following Floyd in order to determine the bioavailability of the posthurricane river DOC. Samples were collected on 23 September 1999 and on 28 September 1999 when the river was close to its maximum flood stage. The DOC concentration of the first sampling period was 1.2 mM, which is near the upper range of normal river values [Avery et al., 2003]. The percent labile DOC (14%) was also similar to that previously obtained for the river under normal conditions (Figure 5). The sample collected during the second sampling period displayed an elevated DOC concentration of 1.9 mM, roughly double normal river values. The percent bioavailability of this DOC (33%) was the highest ever obtained for the Cape Fear River (Figure 5).

Figure 5.

Bioavailabilty of river and rainwater dissolved organic carbon determined by measuring total loss of DOC during incubation experiments. Rainwater and average river value are from Avery et al. [2003]. Average river value measurements were from river samples collected on 19 February 1999, 19 April 1999, and 16 July 1999.

[17] The disparity in the concentrations and bioavailability of DOC between these two posthurricane sampling periods most likely reflects compositional differences in the DOC present in the estuary. Immediately following the hurricane, the similarity of the DOC bioavailability and concentration to normal river values suggests the DOC in the estuary at this time was coming from the same local sources as under nonhurricane impacted conditions.

[18] Two weeks after the hurricane made landfall markedly different results were observed. There was significantly more DOC in the river (1.9 mM versus 1.2 mM) and a much greater percent of it was bioavailable. The much greater bioavailability in the 28 September sampling period suggests the sources of DOC in the estuary during this sampling period were much different than under nonhurricane impacted conditions. During the two weeks lapse following the landfall of the hurricane the floodwaters in the piedmont of North Carolina had sufficient time to reach the Cape Fear Estuary near the sampling site. The DOC in the river during this later sampling period could therefore have contained some nonwetland DOC sources resulting from runoff and possibly point sources from the piedmont. Much of this DOC could have originated from municipal sewer systems, which released large quantities of untreated sewage following the hurricane due to the excessive rainfall and over burdening of the plants with excess water. Mallin et al. [1999] reported a large decrease in estuarine dissolved oxygen concentrations with values approaching 0.0 mg L−1 in the weeks following Hurricane Floyd. This is not surprising considering the DOC concentration in the estuary was roughly twice usual values and the DOC was approximately 3 times more labile, resulting in a potential biological oxygen demand 6 times higher than normal. Assuming a high dissolved oxygen concentrations (10 mg L−1 or 0.31 mM) and 33% of the 1.9 mM DOC being bioavailable (0.66 mM), and a 1:2 ratio of C:O during respiration, the bioavailable DOC could consume all the O2 present.

4. Conclusions

[19] The importance of riverine and rainwater DOC flux under normal conditions has been previously reported for Long Bay [Avery et al., 2003]. On an annual basis, river water deposits roughly 4 times the amount of DOC compared to rainwater. However, when bioavailability is taken into account, rainwater deposits two times as much bioavailable DOC to this coastal bay. The relative importance of hurricane riverine DOC flux, and the bioavailability of that flux, depends on the characteristics of the individual hurricanes. These characteristics include the amount of rainfall, the path the hurricane takes, and the rainfall conditions preceding the hurricane.

[20] The impact of DOC flux from hurricanes Fran and Floyd was strikingly different than that of Bertha and Bonnie. Both Fran and Floyd deposited an extraordinary amount of riverine DOC to Long Bay, mainly in the form of recalcitrant DOC. The storm tract of hurricane Fran was such that it deposited a tremendous amount of rain in the Cape Fear River watershed which was already saturated with water from hurricane Bertha a few weeks earlier. These conditions resulted in high river flow rates with a correspondingly high DOC flux. Hurricane Floyd deposited high rainfall amounts in the coastal watershed as well as the Cape Fear River watershed resulting in 100 year flood conditions. The flooding resulted in the transport of nonwetland sources of DOC to Long Bay. This DOC was not only higher in concentration to normal river values but was also higher in bioavailability. Hurricanes Bertha and Bonnie took a more easterly path than Fran after making landfall and were preceded by relatively normal precipitation conditions. As a result, the riverine flux of DOC was no different than under nonhurricane conditions.

[21] The flux of hurricane rainwater DOC depends on the amount of rain and the DOC concentration. Highest rainwater deposition values were obtained for hurricane Bonnie, which had an average hurricane DOC concentration, and a slightly above normal amount of rainfall. When rainfall amounts become extreme, as was the case during Hurricane Floyd, the rain DOC concentrations can become so diluted that the flux is significantly reduced.

5. Summary

[22] The results of the current study demonstrate that over a relatively short time period, posthurricane riverine DOC can provide a significant flux of DOC to the coastal ocean, most of this DOC being recalcitrant and representing potential organic carbon available for long-term storage On the other hand, rainwater associated with hurricanes deposits a substantial amount of bioavailable DOC. Since a large amount of refractory DOC is deposited to coastal oceans during a short period of time during hurricanes, secondary productivity is likely increased immediately following a hurricane possibly effecting biological systems. The impact of hurricane-derived recalcitrant riverine DOC and bioavailable rainwater DOC on the carbon budgets and biological systems illustrates the importance of examining coastal systems not only under normal conditions but also under nontypical extreme conditions.