NERRS banner

System-wide Monitoring Program

Synthesis of the Water Quality Data from 1995 to 2000
Chapter 6: Impacts of Tropical Systems on Water Quality Data, 1995-2000.

Introduction
Water cchemistry of estuaries is invariably related to weather patterns at daily, seasonal, and inter-annual time scales. Precipitation affects water chemistry through the direct input of freshwater as well as the transport of sediments, nutrients, and contaminants into estuarine systems. Evapo-transpiration provides a mechanism for the removal of freshwater from estuarine systems. Changes in barometric pressure and heat exchange generate wind and waves that result in turbulent mixing, de-stratification of the water column, and input of terrestrial matter through the erosion of shorelines. Wind and wave forcing also create currents that transport water, sediments, nutrients, and contaminants into and out of estuarine systems.

Seasonal weather patterns fall into two general categories: (1) fronts that form over land and pass over the coast on their way out to sea, and (2) storms that form at sea and move toward the coast. On the Pacific Coast of North America, weather systems that affect the coast form at sea and move towards land. During the tropical season (15 May – 30 Nov), storms that form in the East Pacific Ocean rarely make landfall in the United States. During the non-tropical season, storms that form in the East Pacific Ocean often pass over the coast and can drop considerable amounts of precipitation, particularly in the northwest. During the tropical season (1 Jun – 30 Nov) on the eastern seaboard of North America, fast-moving tropical systems frequently make landfall in the United States, often depositing considerable localized precipitation and causing substantial erosion. Fronts that move slowly across the eastern seaboard, usually produce less precipitation than tropical systems, and then move out to sea, typify weather patterns that affect estuarine systems along this coast during the non-tropical season.

During the tropical seasons that occurred between 1995 and 2000, 75 named tropical systems were observed in the Eastern Pacific Ocean and 80 named tropical systems were observed in the Western Atlantic Ocean. None of the tropical systems in the Eastern Pacific Ocean made landfall on U.S. soil, compared to 25 tropical systems (31% of total named storms) that made landfall on U.S. soil from the Western Atlantic Ocean. Subsequently, 24 tropical systems passed directly over or close to at least one water quality monitoring site in the National Estuarine Research Reserve’s System-wide Monitoring Program. Due to the unique positioning of NERR SWMP sites and the movement patterns of these storms, multiple NERR SWMP sites were affected by the same storm on numerous occasions. Changes in and subsequent recovery of water chemistry at NERR sites in association with these storms is herein examined.

Methods
Movement patterns of tropical systems between 1995-2000 were obtained from the National Hurricane Center (www.nhc.noaa.gov) in order to determine the frequency, duration, and intensity of tropical systems that potentially impacted water quality parameters at NERR SWMP sites.

Daily mean and range of water quality variables [temperature, salinity, depth, turbidity, DO (% sat), and pH] two weeks prior to and four weeks following storm passage were plotted in MS Excel in order to visually examine the effects of each storm on the respective water quality parameters at each site. These plots were then sent to the respective NERR Research Coordinator for review and comment.

Short-term and long-term effects of each storm were quantified for each parameter at each site impacted by a tropical system. Short-term effects for water temperature, depth, and turbidity were calculated as the maximum one-day shift in daily mean values as the storm approached or passed over the NERR site. Short-term effects for salinity and pH were calculated as the maximum difference and duration of change in mean daily values between storm passage and the onset of recovery to pre-storm conditions. Long-term effects for salinity and pH were calculated as the total number of days required in order for mean daily values for these parameters to return to pre-storm conditions. Short-term effects using mean daily values for dissolved oxygen (% saturation) were less informative given large (40-80% saturation) daily ranges for this parameter; thus, daily range values for periods of noticeable departure from pre-storm conditions were used instead.

Frequency distributions of short-term effects (one-day change in mean daily values or daily range) were created in order to place the short-term effects of each tropical system into perspective. Water quality data were queried using a relational database (MS Access) and the resulting output data processed using the Histogram function in Microsoft Excel. Annual scatter pots of mean daily values were created in order to place the long-term effects of tropical systems into perspective.

Results
Between 1995 and 2000, 24 tropical systems passed over one or more Reserves in the National Estuarine Research Reserve program (Table 22). Similar numbers of tropical systems were encountered by NERRs in the Gulf of Mexico (n=19), Southeast (n=20), and Mid-Atlantic (n=17) regions (Figure 48). Seventy-eight percent of tropical systems that affected Jobos Bay were hurricane intensity. Eighty-four percent of tropical systems affecting NERRs in the Gulf of Mexico were tropical storm or hurricane intensity, more than double the percentage of systems of the same intensity that affected NERRs in the Southeast. Similarly, the percentage of tropical depressions or extra-tropical systems that passed over the Southeast NERRs (35%) was more than double the percentage of tropical depressions and extra-tropical systems that passed over the Gulf of Mexico NERRs. This discrepancy is attributed to five tropical storm and hurricane systems that affected NERRs in the Gulf of Mexico, made landfall, then moved up the eastern seaboard, steadily losing intensity. All tropical systems that affected Mid-Atlantic and Northeast NERRs also affected NERRs in the Southeast and/or Gulf of Mexico.

No data were collected for two hurricanes (Marilyn and Hortense) that only affected sites at the Jobos Bay Reserve in Puerto Rico. With these exceptions, data were collected for at least one NERR SWMP site for each of the remaining 22 tropical systems (Table 22 (PDF)). In all, 128 data sets from NERR SWMP sites monitored during tropical system events were examined for changes in water quality observations associated with storm passage.

Since no noticeable changes in parameters monitored or data were not collected for a total of 29 data sets, these data sets were excluded from analyses. No data were collected for an additional 3 data sets for water temperature, 3 data sets for salinity, 2 data sets for depth, 24 data sets for turbidity, 11 data sets for DO, and 6 data sets for pH (Table 23).

Noticeable changes were observed for at least one parameter in 99 data sets. Noticeable changes in water temperature were observed in 71 data sets (74%), salinity in 56 data sets (58%), depth in 65 data sets (67%), turbidity in 39 data sets (52%), DO in 23 data sets (26%), and pH in 37 data sets (40%). In 80% of the data sets, effects from tropical systems were observed one day prior to the system passing the NERR or on the day of passage (range = 5 days before to 2 days after) (See PDF for details).

A strong negative linear relationship (R2=0.97) existed between the mean one-day change in mean daily water temperature and storm intensity (Figure 49). Similarly, a strong negative quadratic relationship (R2=0.99) was also observed for the maximum observed one-day change in mean daily water temperature (Figure 49). Decreases in mean daily water temperature less than or equal to -1.5°C were observed for 70% of hurricanes, 47-50% of tropical storms and tropical depressions, and 29% of extra-tropical systems examined. Overall, tropical systems accounted for less than 0.5% of these precipitous drops in mean daily water temperature (£ -1.5°C) between consecutive days in data sets collected at NERR SWMP sites between 1995-2000 (Figure 50). (See PDF for details)

No relationship between storm intensity and short-term change in salinity was evident (Figure 51). Six storm systems (Bertha, Fran, Bonnie, Georges, Dennis, and Floyd) drastically altered mean daily salinity, and these altered patterns persisted for 0.5 to 3.5 months at multiple NERRs (Figures 52-56) (See PDF for details)

No relationships between storm intensity and water depth and turbidity were evident (Figures 57-58). Changes in pH and DO occurred in less than 50% of the data sets examined, but were observed in at least one data set for 16 of the 21 tropical systems examined. Changes in both pH and DO occurred in 13 data sets from nine tropical systems (Table 24) (See PDF for details).

Discussion
Abrupt decreases in water temperature prior to storm passage were consistently observed in data sets, with increasing cooling effects strongly related to increasing storm intensity. Changes in water temperature were not observed for 25 data sets. In one instance (Zeke’s Island during Hurricane Fran), mean daily water temperature showed no change during storm passage, but increased slightly (3.2°C) during the four days following storm passage. This pattern may have been related to excessive biological oxygen demand (BOD) levels in the Cape Fear River (Mallin et al. 1999). Minor changes in all water quality parameters were observed for the remaining 24 data sets. The lack of noticeable changes in water temperature and other parameters for these data sets were probably due to (1) storms passing to the west of the NERRs, (2) large distances between storms and NERRs and (3) rapid deterioration of storm intensity and subsequent short-term exposure (<12 hours) to storms due to rapid movement past NERRs.

The phenomenon of sea surface temperature cooling related to hurricanes in the open ocean has been extensively studied (Beckle 1974, Price 1983, Greatbach 1985, Sanford et al. 1987, Cornillon et al. 1987, Shay et al. 1989, Sakaida et al. 1998). In the open ocean, sea surface temperature cooling of 2-9°C associated with the maximum sustained winds of hurricanes has been reported. As cool waters below the thermocline are brought to the surface via upwelling, warm surface waters are replaced by cold water in the wake of the storm, which effectively creates a cold water ‘footprint’ of the storm track. Although wind mixing and subsequent upwelling are the primary mechanisms responsible for sea surface cooling, the magnitude of the cooling effect depends on the initial thermal stratification in the ocean and storm mobility (Chang and Anthes 1978). Localized mixing occurs over longer periods in slow moving storms; thus, greater sea surface temperature cooling should occur.

Water temperature cooling associated with hurricanes in the open ocean has received much attention; however, documentation of similar changes in water temperature associated with storm passage in coastal and estuarine water bodies is sparse. The mechanism of water temperature cooling may be related to wind mixing and/or heat exchange at the air-water interface, but is unknown at the present time. Detailed examination of these potential mechanisms may be possible during future tropical system events with the incorporation of weather station data into the NERR SWMP.

During Hurricane Dennis, Arendt et al. (2001) observed the same water temperature cooling response in 18m of water in lower Chesapeake Bay as documented at the Chesapeake Bay Virginia, Goodwin Islands site, on the opposite side of the Bay. Coincident with this water temperature cooling, Arendt et al. (2001) observed a three-day period of inactivity for several adult tautog, Tautoga onitis, at a shipwreck continuously monitored using ultrasonic telemetry equipment. These authors also observed similar changes in detection patterns of adult tautog at all monitored sites coincident with abrupt decreases and increases in daily mean water temperature between Nov 1998 and Sep 1999 (Arendt et al. 2001). At a minimum, these findings suggest that abrupt changes in water temperature may serve as indicators of change in other physical or chemical parameters that have short-term ecological consequences in certain habitats.

Habitats monitored by the NERR SWMP largely consist of shallow (mean depth = 2m), tidal creeks. Daily temperature variation in many of these systems can reach up to 10°C. Although large daily temperature variation is regularly experienced, abrupt shifts in mean daily water temperature (£ -1.5°C), similar to shifts observed during the passage of tropical systems in NERR data sets between 1995-2000, are less common and occur less than 6% of the time at most sites. Furthermore, the abrupt shifts in mean daily temperature observed during the passage of tropical systems represent <0.5% of the total occurrence of these types of temperature changes (Figure 50).

If wind mixing is also responsible for these additional shifts in mean daily water temperature, then shifts in mean daily water temperature may be useful as indicators of the frequency of strong wind mixing events at NERR SWMP sites. The primary ecological ramification of strong mixing events is the breakdown of water column stratification, which can lead to increases in hypoxia. Given the small size of the water bodies monitored by the NERR SWMP and the strong tidal amplitudes experienced at these sites, wind mixing effects may be masked by the twice daily flushing effects of tidal cycles at these sites. Wind mixing may, however, play an important role in controlling the magnitude of the daily variation in DO (% sat). Although changes to DO were only observed in 27% of the data sets examined, strong wind forcing during several storms appeared to drastically reduce daily variation (to less than 20%) in DO for several days.

Short-term changes to salinity and depth during the passage of tropical systems were variable and dependent on the fetch of approaching storms. Tropical systems approaching from open water were usually associated with an initial increase in depth and salinity due to storm surge, followed by a decrease in salinity and depth after storm passage due to precipitation and strong winds pushing water down the estuary. This scenario was particularly evident in data sets from NERRs in the Gulf of Mexico (APA, RKB, WKB) and the North Carolina NERR. Extra-tropical and tropical storm systems that moved up the coast after making landfall were frequently associated with decreases in salinity and depth as a result of precipitation and strong winds pushing water down the estuary. With the exception of the NOC NERR, these types of systems constituted the majority of systems encountered by NERRs on the eastern seaboard (Table 22).

With a few exceptions for salinity, changes to water quality parameters monitored by the NERR SWMP during the passage of tropical systems between 1995-2000 were abrupt, and short-lived. Long-term changes in salinity were evident for only a few storms (Figures 80-84). Altered salinity distributions and excessive runoff from these storms subsequently resulted in ecological disturbances in some of these estuaries. Following Hurricanes Fran (1996) and Bonnie (1998), biological oxygen demand (BOD) loads caused large-scale fish kills in the Cape Fear and Neuse River estuaries (Mallin et al. 1999, Burkholder et al. 1999). Co-occurrence of Hurricane’s Dennis and Floyd in September 1999 produced record precipitation and flushed many fish populations into sounds and coastal waters (Mallin et al. 2000). More than two years after Hurricanes Dennis and Floyd, the distribution of some estuarine fish and invertebrate assemblages have still not returned to pre-hurricane levels (L. Crowder, pers. comm.).