SEA LEVEL OSCILLATIONS IN A SALT MARSH LAGOON SYSTEM, NORTH INLET, SOUTH CAROLINA

Trabajo recibido el 1° de noviembre de 1982 y aceptado para su publicación el 22 de noviembre de 1982.

RESUMEN

INTRODUCCIÓN

The water level in estuaries and lagoons fluctuates over a wide range of periods. The most obvious oscillation is usually the semidiurnal or diurnal tide due to astronomical forcing. However, meteorological forcing and variability in fresh water runoff may also cause large fluctuations of sea level on the scale of a few days. Astronomical forcing additionally causes sea level oscillations of a few hours, 24.8 hours, 14 days, 18.6-19-1 Years, and longer. There also exist 20 000-100 000 year periodic variations due to glaciation cycles. Further, there exist seasonal sea level cycles (Pattullo et al., 1955) because of water temperature changes. Particular locations may experience sea level cycles with periods from 20-365 days. Provis and Radok (1979) identified such phenomena along the Australian coast but did not give a definitive explanation.

The purpose of this paper is to focus attention on the existence of meteorologically induced sea level variability with periods most1y from 3 to 14 days. We will do this by analyzing 3 years of hourly water elevation data from North Inlet, South Carolina, a salt marsh lagoon system along the U.S. Atlantic coast (Fig. 1).

There are several important ecological implications of such non-tidal sea level oscillations. The extent of flooding of adjacent marshes, mud flats, and grass beds depends on sea level height. Kjerfve and McKellar (1980) suggested that the exchange of nutrients and organics between marsh and adjacent lagoon waters is a function of the degree to which marshes are flooded. Also, it may well, be that meteorological sea level variations drive material exchanges of dissolved and susp-ended constituents between the coastal ocean and lagoons/estuaries.

Fig. 1. Location of North Inlet, South Carolina.

North Inlet, South Carolina (Fig. 1) is a 32.2 km² salt marsh (Spartina alterniflora) lagoon system bordering on the Atlantic Ocean. Numerous winding tidal creeks, intersect the marsh. Salinities are sually 31-34 ppt because of the lack of significant fresh water runoff.

Analog tide measurements were made with a Weather Measure F551 recording water level gauge. A stilling well. filtered water level fluctuations shorter than 5 minutes. Three years of data were used for our analysis a) 1 june 1974-31 May 1975; b) 1 September 1978-31 August 1979; and c) 1 September 1979-3l August 1980. These three 365-day series were digitized at an hourly sampling rate. They were smoothed using an 8th order, bidirectional, Butterworth filter with a 48 hr half-gain period to illustrate the non-tidal. low-frequency varíations in sea level. Oscillations with periods shorter than 24.82 hours are attenuated by a factor of more than 2.5 x 10^-5 while the semi-diurnal tide is suppressed by a factor of 3 X 10^-10.

Fig. 2. Gain of an Sth order, bidirectional Butterworth filter with a 48 hour gain equal to 0.5.

The digital Butterworth filter is easily realized in serial form (Ackroyd, 1973). As recursive filters produce data phase shifts, each series was serially filtered in both a forward and reverse direction. The resulting filter has no associated phase shifts (Roberts and Roberts, 1978). Its gain, H, is given by H (f) = ( 1 + (tan πfΔ t/tan πfcΔt) ²n)-¹ where f (cph) is frequency; fc (cph) is the frequency at which the filter gain is 0.5; and n is the order of filter as well as an indicator of the number of points lost at each end of the filtered data because of the filtering operation. The frequency gain of the 8th order, 48 hr half-gain period filter is shown in Fig. 2.

Fast Fourier Transform (FFT) variance spectra were calculated for each unfiltered time series. The spectrum (Fig. 3) is the resultant of the three annual data sets, calculated with 30 degrees of freedom per spectral estimate.

Fig. 3. Rariance spectrum of the North Inlet water level. This resultant spectrum wa3 constructed as the mean of three annual spectra with each variance estimate having 30 df. The sampling rate is 1 hr and, thus, the Nyquist frequency is 0.5 cp1i. M., S, and N are the major semi diurnal tidal peaks, K, and 01 the dominant diurnal tidal peaks and the peaks at higher frequencies, M2 and M4 are important overtides. The 90%, confidence interval is shown.

The spectrum is dominated by tidal cons, tituents, as can be expected. North. Inlet experiences a semidiurnal mesotide with a mean range of 1.4 m, which explains 75% of the total vaiance in the sea level records. Diurnal constituents (24.8 hour period) account for 13% of the total variance. The diurnal range varies from 6 cm (equatorial tides) to 35 cm (tropic tides) (Palmer et al., 1980). Although several overtide species (cf. Doodson and Warburg, 1941) are evident in the spectrum (e.g., M₄ - 6.2 hours period, M₆ - 4.2 hour period) , neither the M₄ nor the M₆ constituent contributes significantly to the overall spectral variance.

The tidal fortnightly cycle, a sea level variation associated with the spring-neap cycle due to the declination of the moon, is apparent in the spectrum (Fig. 3) and contains approximately 1% of the overall spectral variance. High spectral values at periods of 27 and 50 days could correspond to variability in the length of day and global angular momentum, respectively.

The low-pass filtered sea level records are shown in figure 4. It is apparent that there exist significant sea level variations with periods from 2 to 14 days. In fact, these highly energetic oscillations account for 4% of the overall variance, exhibit a range from 25 to 80 cm, and occur throughout the year in all three data sets (Fig. 4). However, as these variations do not occur at precise periods, there exists no corresponding distinct peak in the sea level spectrum (Fig. 3) ; rather the frequency band as a whole is energetic.

Although 2-14 day sea level oscillations exist throughout the year, the winter fluctuations are more intense and more frequent as compared to summer occurrences. Brooks and Mooers (1977a) found 7-10 day and 12-14 day periodici ties in the winter and summer time records, respectively. The seasonal difference is clearly evident in the North Inlet summer and winter sea level spectra (Fig. 5) as well. This difference is due to the regular passage of meteorological fronts during late fall and winter. Such fronts are less intense and less regular during the summer, consistent with the lower variance in the 2-14 day band during summer.

Fig. 4. Low-pass filtered sea leved records for North Inlet for 1974-75, 1978-79, and 1979-80.

Non-tidal low-frequency sea level fluctuations are associated with meteorological forcing. However, the dynamics of such oscillations are poorly understood, and they are not easily predictable. Kjerfve et al. (1978) showed that sea level variations from 2-10 days in North Inlet are coherent with atmospheric pressure variations and wind stress, thus dependent on the passage of frontal systems. A high correlation between water elevation and wind stress along the U.S. east coast was also demonstrated by Blanton and Maddox (1978) and Chao and Pietrafesa (1980). Sea level variations may be manifestations of continental shelf waves, as well. Brooks (1978) found a coupling between atmospheric forcing and sea level variation at periods of 3 and 10 days, approximating the 12-day period of zero group speed barotropic continental shelf waves (Brooks and Mooers, 1977b) and spinoff eddies that propagate across the continental shelf (Lee and Mayer, 1977).

Fig. 5. Comparison of the low-passed filtered sea level spectra for December-February (winter) and june-August (summer) for the mean of the three years. Each. Spectral estimate has associated 18 df.

Our main objective in this paper is to focus attention on the existence of energetic wavefeatures in lagoonal/estuarine sea level records. These have periods from 2 to 14 days, may have heights up to 80 cm, and are caused by direct meteorological forcing or indirectlly by meteorological forcing transferred via the coastal ocean. These waves, (Fig. 4), which cannot be predicted, may have important ecological implications. Associated with the sea level oscillations are water discharges, also on time scales from 2 to 14 days, which may be responsible for net exchanges of waterborne materials between coastal ocean and lagoon/estuary. In addition, an astronomical high tide occurring at the time of the passage of a lowfrequency wave crest will cause more extensive flooding of adjacent marshes, swamps, and grass and mud flats than a high tide at the time of a lowfrequency wave trough passage. As Kjerfve and McKellar (1980) suggested, it is particularly during times of wetlands flooding that nutrients and detritus are likely to be exchanged between lagoon waters and the perimeter of the tidal watershed. Thus, measurement of material fluxes or computation of material budgets based on only a few tidal cycles may result in biased or totally erroneous conclusions. It may be necessary to relate measurements to the underlying lowfrequency sea level fluctuations to verify that material flux estimates are not biased toward times of either rising or falling low-frequency sea level.

One notable event occurs in the beginning of the 1979-80 data set (Fig. 4c). The anomalously high low-pass filtered water elevation in early September corresponds to the passage of Hurricane David along the southeastern U.S. coast.

Although meteorological sea level variations account for only 4% of the total variance, such variations are still significant. The percentage is small because North Inlet experiences a large tidal range. However, meteorologically induced sea level fluctuations of similar magnitudes can be expected for most coasts. Thus, in regions experiencing microtides, the low-frequency sea level waves can easily account for 50% of the variance, depending on local conditions, and thus achieve a larger relative importance. Such microtidal regions are common throughcut the Gulf of Mexico and Caribbean Sea.

The remaining 7% of the sea level variance is associated with seasonal fluctuations of periods in excess of a month. The characteristic annual sea level variation is a high in September-October and a low level in February (Fig. 6). The range of this cycle is on the average of 25 cm. According to Pattullo et al. (1955), most of the seasonal oscillation is due to steric changes in the water column, primarily as a result of the annual temperature cycle.

Fig. 6. Annual sea level curve for North Inlet with mean values for cach month ± the standard error of the mean, based on the three data years.

Aviso de Privacidad

Cerrar esta ventana