The Effect of a Tidal Cycle on the Dynamics of Nutrients in a Tidal Estuary in the Seto Inland Sea, Japan

Transcription

1 Journal of Oceanography, Vol. 54, pp. 65 to The Effect of a Tidal Cycle on the Dynamics of Nutrients in a Tidal Estuary in the Seto Inland Sea, Japan SHIGERU MONTANI, PAOLO MAGNI, MEGUMI SHIMAMOTO, NAO ABE and KOICHI OKUTANI Department of Bioresource Science, Kagawa University, Miki , Japan (Received 16 May 1997; in revised form 10 September 1997; accepted 29 September 1997) A 24 hour time series survey was carried out during a spring tide (tidal range ca. 2 m) of May 1995 on a tidal estuary in the Seto Inland Sea, Japan, in the context of an integrated program planned to quantify the dynamics of biophilic elements (carbon, nitrogen and phosphorus) and the roles played by the macrobenthos on the processes. Three stations were set along a transect line of about 1.4 km, which linked the river to the rear to the innermost part of the subtidal zone. Every hour, at each station, measurements were made of surface water temperature, salinity and dissolved oxygen concentration, and surface water was collected for the determination of nutrients [NH 4 + -N, (NO 3 + NO 2 )-N, PO 4 3 -P and Si (OH) 4 -Si]. During the ebb flow, riverine input of silicate and nitrate+nitrite significantly increased the concentrations of both the intertidal and the subtidal stations. Conversely, during the high tide, river nutrient concentrations were lowered by the mixing of fresh water with sea water. As a result, best (inverse) correlations were found at the river station for salinity against silicate (y = 2.9 Sal , r 2 =0.879) and nitrate+nitrite (y = 1.3 Sal , r 2 =0.796). In contrast, ammonium nitrogen concentrations were higher at intermediate salinities. Indeed, no significant correlation was found between salinity and ammonium. The effect of the macrobenthos, which is abundant on the intertidal flat, is discussed as a biological component that influences the processes of nutrient regeneration within the estuary. The effect of the tidal amplitude is an important one in determining the extent of the variations in nutrient concentrations at all three stations, which were stronger between the lower low tide and the higher high tide. Keywords: Tidal estuary, nutrient regeneration, macrobenthos, nutrient export, river input, salt intrusion, Seto Inland Sea. 1. Introduction An important aspect of the high variability of tidal estuaries is related to the short-time effects of a tidal cycle on the physicochemical characteristics of the water. On a time scale of hours, ebb advection of fresh water and salt water intrusion during the flood determine strong changes of several fundamental parameters of the water column, such as salinity (Uncles and Stephens, 1996), nutrients (Hernandez-Ayon et al., 1993; Balls, 1994) and suspended particulate matter (SPM) (Renshun, 1992). The extent of such variations will vary significantly depending on spring-neap tidal state or amplitude (Vörösmarty and Loder, 1994; Yin et al., 1995a, b; Uncles and Stephens, 1996). Moreover, matters may be greatly complicated by other environmental variables, such as precipitation rate, affecting fresh water discharge (Schubel and Pritchard 1986, Page et al., 1995), winds (Yin et al., 1995b) and current velocity (Renshun, 1992). There have been several approaches to the quantification of such variability. In the Tweed estuary on the east coast of Scotland, longitudinal and vertical surveys were undertaken from the seaward end to the up-estuary limit of saline intrusion at different tidal amplitudes and freshwater inflow (Uncles and Stephens, 1996). On the same Scottish coast, longitudinal surveys of nine rivers (including the river Tweed) were completed at all different seasons, within 3 5 days of each other (Balls, 1994). Comparisons between nutrients within an individual estuary were made on the basis of a strictly controlled sampling regime, around high water at the spring tide. In the Colorado River Delta (Hernandez-Ayon et al., 1993), too, seasonal changes of the spatial distribution of nutrients were compared at different seasons by sampling during the flood. In all these studies, transect lines were set along distances of several km. The time taken to travel between samplings, tidal-water displacements, variations in tidal velocities and a possible wide range of river discharge may have represented major drawbacks in this work. Time series surveys have also been undertaken as a valuable approach to quantifying the shortterm variability of estuarine waters (Yin et al., 1995a, b). Copyright The Oceanographic Society of Japan. 65

2 However, less information seems to be available about the effect of a tidal cycle on short distances, with coordinate and real-time (not elapsed time) information at different stations. A task of primary importance when studying nutrient distributions within an estuary is an evaluation of both inland inputs and in situ regeneration. Balls (1994) indicated that, on rivers along the Scottish East Coast, nutrient concentrations were related to land use. He reported nitrate + nitrite and phosphate concentrations up to 600 µm and 7 µm, respectively, in river catchments with intensive agriculture and low freshwater input. On the Forth River, Balls (1992) found a broad mid-estuarine maximum of ammonium concentration during the summer months. This maximum was attributed to a benthic input from decomposing materials in the sediment, suggesting that ammonium import by fresh water discharge was less significant than nitrate + nitrite. Also, in a Southern Californian salt marsh, estimations of the tidal export of ammonium were often more than 100 times higher than estimated loading from stream discharge, and it was independent of loading rates (Page et al., 1995). Possible explanations for ammonium regeneration within the marsh were given, such as dissimilatory reduction of nitrate, microbial ammonification of organic material in marsh sediments, and excretory products of the macrobenthos. The present study is part of a long-term, integrated monitoring program which aims at quantifying the dynamics of biophilic elements (carbon, nitrogen, phosphorus) in a small tidal estuary in the Seto Inland Sea, Japan. Within this program we followed, on a monthly basis, the quantitative seasonal changes of abundance and faunal composition of the intertidal and subtidal benthic communities to evaluate the roles played by the autochtonous macrobenthos in the processes. The first objective of this study was to evaluate the impact of river runoff and salt intrusion on the nutrient balance of this estuary during a complete tidal cycle. Since this is a mixed-semidiurnal type estuary (with pronounced differences between two successive low and high tides) we aimed to quantify the influence of the tidal amplitude on the extent of fresh water intrusion. Our second objective was to verify the sources of the different nutrients: from inland, outside or within the estuary. 2. Materials and Methods The survey was carried out during a spring tide of May It started at 10:00 on the 30th and ended at 10:00 on the next day. We conducted our 24-h time series observation on a river station (Stn. A), an intertidal station (Stn. B) and a subtidal station (Stn. Y3) set along a transect line (Fig. 1). Distances between stations were about 1.1 km and 0.3 km, respectively. At the river and the intertidal stations, samples were obtained from a small track on the shoreline, while at the subtidal station we used a research vessel, stationed Fig. 1. Study area and location of the sampling stations along the tidal estuary: Stn. A (river station), Stn. B (intertidal station) and Stn. Y3 (subtidal station). close to the low-tide shoreline of the intertidal flat. Surface water temperature, salinity and dissolved oxygen concentration were measured and surface water was sampled simultaneously at each station every hour. On the subtidal station, with a depth of ca. 10 m at high tide, the vertical profile was also monitored every two hours. The water samples were analyzed to determine nutrients [NH N, (NO 3 + NO 2 )-N, PO 4 3 -P and Si (OH) 4 -Si], pigments (Chl a and pheo-pigments), and particulate organic carbon (POC) and nitrogen (PON). In this article, attention will be restricted to the superficial horizontal distributions of nutrients. At the river station, hydrological measurements were done with a portable D.O. meter (UK 2000) and a portable salinometer (YSI Model 30). At the intertidal station, a STD recorder (Alec ACL-208 DK) was used for temperature and salinity, and a STDDO recorder (Alec ADO 1050-D) was employed at the subtidal station. On the intertidal station, dissolved oxygen concentration of surface water was not measured, but a CTD recorder (Alec AST-1000) was placed 66 S. Montani et al.

3 about 10 cm from the bottom sediment. Water samples from the three sampling stations were brought to the laboratory simultaneously, every 6 h, and, soon after, filtered through Nucleopore filters (pore size: 0.4 µm). Nutrient concentrations in the dissolved phase were determined in duplicate using a nutrient autoanalyzer (Technicon II) according to the method of Strickland and Parsons (1972). 3. Results 3.1 Tidal cycle During the sampling period, two lower and two higher tidal levels were predicted (the Maritime Safety Agency, Fig. 2). The first high tide, at 10:57, was +197 cm and the second one, at 0:24, was +239 cm. The difference in tidal height was used to distinguish them as a lower and higher high tides, respectively. Correspondingly, a lower low tide at 17:28 and a higher low tide at 6:41 occurred, which were +35 cm and +105 cm, respectively. The largest difference of the tidal level was 204 cm between 17:28 and 0:24. At lower low tide, the water column was reduced to few tens of centimeters at Stns. A and B, while it was ca. 7 8 m at Stn. Y Precipitation rate Figure 3 shows the daily precipitation rate and solar radiation reported for May 1995 (Takamatsu Meteorological Bureau), since this might possibly be correlated with the strength of the river runoff. Heavy rain occurred during this month, totalling 260 mm. A peak of 77 mm d 1 occurred on the 14th. During the survey, the weather was fine, following two days of little rain. 3.3 Temperature At the start, 1 h before the lower high tide, the temperature was high at Stn. A (20.4 C), intermediate at Stn. B (18.8 C) and low at Stn. Y3 (17.1 C) (Fig. 4a). Until 20:00 (2.5 h after lower low tide), temperatures remained significantly high at Stn. A, reaching a maximum in the afternoon of up to 24.8 C. In the evening, with the surging tidal flood and in coincidence with decreasing air temperature, surface water temperature significantly decreased at all the three stations. The temperature was uniform at 24:00, 0.5 h before higher high tide, varying from 17.2 C (Stn. B) to 17.8 C (Stn. A). Short after the higher high tide, temperature was lowest at Stn. Y3 (16.8 C, at 1:00 and 2:00), while at Stn. A it increased slightly up to 18.6 C (3:00), as low salinity water (Fig. 4b), not completely mixed during the flood, was sampled at Stn. A. After the higher low tide, a marked pattern was progressively re-established, as on the previous day. At 10:00, temperature was again higher at Stn. A (20.6 C), intermediate at Stn. B (19.0 C) and lower at Stn. Y3 (18.4 C). 3.4 Salinity During the first 6 h of measurements, up to 2.5 h before lower low tide (15:00), there was a strong salinity gradient (Fig. 4b). Salinity was significantly low at Stn. A, with a mean of 4.5 ± 3.8 psu. On the lower part of the estuary, salinity was 22.7 ± 2.5 psu and 29.7 ± 2.5 psu, at Stns. B and Y3, respectively. During the ebb flow, fresh water runoff progressively and strongly reduced the salinity at both Stns. B and Y3. From 16:00 to 20:00, salinity was lowest at Stn. A (mean of 2.2 ± 0.1 psu) and dropped to 7.6 ± 3.9 psu and 20.9 ± 2.1 psu, at Stns. B and Y3, respectively. Conversely, Fig. 2. Spring-tide cycle during the 24 hours survey of May 30 31, Nutrient Dynamics in a Tidal Estuary 67

4 Fig. 3. Precipitation and solar radiation recorded during May during the flood, salt intrusion started to move back into the estuary. Around the higher high tide (0:24), salinity was distributed homogeneously along the transect line. At 23:00 and 24:00, salinity varied from 30 psu to 31 psu at all stations. From 1:00 to 5:00, soon after the higher high tide, salinity remained steadily high at Stns. B and Y3 (31.1 ± 0.6 psu and 30.4 ± 0.7 psu, respectively), but dropped to 18.1 ± 5.4 psu at Stn. A. Shortly after the higher low tide (at 6:41), at Stns. B and Y3, salinity also dropped from 30.6 psu to 14.8 psu (7:00) and from 30.0 psu to 26.3 psu (8:00), respectively. At the next ebb flow, the spatial differences were re-established along the estuary, although values were higher than those of the previous day (Fig. 4b). 3.5 Silicate [Si (OH) 4 -Si] The fluctuation of silicate concentration (Fig. 4c) depended strongly on the river runoff and the tidal cycle. During the first 4 h of sampling, at lower high tide, the concentrations varied markedly along the estuary. It was high at Stn. A (86.2 ± 6.5 µm), intermediate at Stn. B (32.4 ± 11.5 µm) and low at Stn. Y3 (11.5 ± 2.3 µm). During the ebb flow, the silicate concentration at Stn. A, which was less affected by sea water, increased up to a maximum of 116 µm (0.5 h after lower low tide). From 14:00 to 19:00, mean concentration was 109 ± 7.5 µm. The fresh water spread progressively toward the lower part of the estuary and strongly affected silicate concentrations at the other two stations. During this period, mean concentration at Stns. B and Y3 increased up to 70.0 ± 16.3 µm and 54.1 ± 10.5 µm, respectively. Conversely, between 21:00 and 22:00, low-silicate water started to flush backwards into the estuary and strongly affected the silicate concentration at Stn. A, which was lowest before the higher high tide period. At this station, however, the silicate concentration increased rapidly again, as low salinity (and higher temperature) water had started moving down estuary with the new ebb. At the end period of the survey, marked differences along the estuary were reestablished, with the silicate concentration progressively decreasing from Stn. A to Stn. Y3. On the other hand, these differences were less than those found at the start of the survey (Fig. 4c). 3.6 Nitrate + nitrite nitrogen [(NO 3 + NO 2 )-N] At the start, nitrate + nitrite nitrogen concentration (Fig. 5a) was much higher at Stn. A (53.3 µm) than at Stns. B and Y3 (8.9 µm and 6.3 µm, respectively). During the next three hours, as lower low tide approached, there was a marked gradient between the stations. The concentration of nitrate + nitrite nitrogen was highest at Stn. A (39.4 ± 10.3 µm), intermediate at Stn. B (29.8 ± 4.3 µm) and lowest at Stn. Y3 (5.4 ± 1.2 µm). From 14:00 to 20:00, during the ebb flow, maximum values were observed at both Stn. A (53.4 µm) and at Stn. B (70.0 µm). During the same period, the concentration at Stn. Y3 also progressively and strongly increased from 26.4 µm to 47.4 µm. Conversely, during the flood, high-salinity water caused a significant drop of nitrate + nitrite nitrogen concentration at all stations. Stn. B had low concentration for a prolonged period, while at Stn. Y3 the lowering effect of high salinity- 68 S. Montani et al.

5 Fig. 4. Time series horizontal distribution of surface water temperature (a), salinity (b) and silicate concentration [Si(OH)4-Si] (c). Fig. 5. Time series horizontal distribution of surface water concentrations of nitrate + nitrite nitrogen [(NO3 + NO2 )-N] (a), ammonium nitrogen (NH4 + -N) (b) and phosphorus (PO4 3 -P) (c). Nutrient Dynamics in a Tidal Estuary 65

6 water was overcome, from 1:00 to 4:00 (Fig. 5a), by intertidal water with a much higher nutrient concentration. At Stn. A, soon after the higher high tide (0:24), low salinity water (salinity 10.6 psu at 2:00, Fig. 4b) coincided with an increase in nitrate + nitrite nitrogen concentration up to 33.5 µm. During the next ebb flow, the low salinity and high silicate and nitrate + nitrite nitrogen water mass had moved down the estuary. This particularly influenced concentrations at Stn. B, where nitrate + nitrite nitrogen concentration increased up to a peak of 44.4 µm, at 7:00. At 9:00, nitrate + nitrite nitrogen concentration was rather homogeneous along the estuary. At 10:00, at the end of the survey, the same distributional pattern as the previous day was present, although to a more limited extent. 3.7 Ammonium nitrogen (NH 4 + -N) As Fig. 5b shows, during the first 4 h of sampling, ammonium nitrogen concentration was relatively low and similar at Stns. A and Y3, with a mean of 6.9 ± 3.7 µm and 9.6 ± 2.0 µm, respectively. By contrast, the concentration was significantly higher at Stn. B, with a mean of 23.4 ± 5.7 µm. During the ebb flow, fresh water did not lower ammonium concentration at Stn. B. Conversely, intertidal water strongly affected the concentration at Stn. Y3, which increased significantly. From 14:00 to 21:00, mean concentration was 28.3 ± 7.1 µm and 29.5 ± 4.6 µm, at Stns. B and Y3, respectively. During this period, the ammonium concentration at Stn. A was the lowest, as it was in salinity. From 14:00 to 20:00, both salinity and ammonium concentration at Stn. A were the least, with a mean of 2.3 ± 0.2 psu and 3.0 ± 0.8 µm, respectively. With the new surging flood, high salinity water changed the nutrient distributional pattern. At Stn. Y3 from 21:00 to 23:00, ammonium concentration dropped from 31.7 µm to 4.8 µm. At Stn. B from 20:00 to 22:00, it dropped from 43.3 µm to 6.9 µm. At Stn. A, ammonium concentration increased by intertidal water brought up stream. After the higher high tide, the influence of sea water started to diminish and concentration increased both at Stn. Y3 and at Stn. B. It also increased further at Stn. A. From 6:00, at the higher low tide, the concentration in the lower part of the estuary was again higher than those in the upper part. As with salinity, the limited strength of the higher low tide gave higher ammonium concentrations than those found during the lower low tide. 3.8 Phosphate (PO 4 3 -P) During the first four hours of sampling, the phosphate concentration (Fig. 5c) at Stns. A and B was higher than that at Stn. Y3. Mean concentrations were 2.5 ± 0.4 µm, 2.6 ± 0.7 µm and 1.3 ± 0.6 µm, respectively. During the ebb tide, the phosphate concentration increased up to 5.3 µm at Stn. A and 5.2 at Stn. B (18:00). That increase strongly influenced phosphate concentration at Stn. Y3. From 21:00 22:00, as the tidal level rose, phosphate concentration at both Stn. Y3 and Stn. B decreased to the lowest values. At Stn. Y3, 0.5 h before and after the higher high tide, the concentration was 1.0 µm. At Stn. B from 22:00 to 2:00, phosphate concentration was constant at 1.4 µm. But the concentration at Stn. A reduced to a very limited extent (2.5 µm) for a shorter period. At this station, soon after midnight, low salinity water lead to an increase in phosphorus concentration until the end of the survey. Fig. 6. Dissolved oxygen concentrations of surface water at Stns. A ( ) and Y3 ( ), and bottom water at Stn. B ( ). 70 S. Montani et al.

7 3.9 Dissolved oxygen Dissolved oxygen concentration (Fig. 6) varied among 3.8 mg l 1 and 10.1 mg l 1 at Stn. A, among 1.9 mg l 1 and 6.9 mg l 1 at Stn. B (ca. 10 cm above the bottom sediment) and among 6.1 mg l 1 and 9.8 mg l 1 at Stn. Y3. A wide fluctuation of dissolved oxygen concentration was observed at both Stns. A and B, a more limited one at Stn. Y3. At Stn. A it was highest at 13:00 (10.1 mg l 1 ) and decreased progressively during the ebb period, with a minimum at 21:00 (3.8 mg l 1 ). At Stn. B from the start to 3:00, near-bottom water was normoxic (76.0 ± 10.2% of air saturation). However from 4:00 to the next sunrise, a significant drop of concentration was observed, with a minimum of 1.9 mg l 1 (23.5% of air saturation) at 9:00 (Fig. 6). 4. Discussion 4.1 Physical parameters and environmental variables Fresh water runoff strongly influenced the superficial temperature and salinity of the estuary. Contours of temperature and salinity clearly showed the presence of water masses of different origin (Figs. 4a and b). Station A was characterized by superficial water with high temperature and low salinity. Plots of temperature against salinity for individual samples (Fig. 7), indeed, gave a more significant (inverse) regression line at Stn. A (r 2 = 0.738). From Stn. A to Stn. Y3, this correlation becomes weaker, as a result of the mixing between fresh water and intertidal water. Water temperature also depended on heat transfer through the sea surface, at all stations being highest between 16:00 and 17:00 and lowest between 1:00 and 2:00 (Fig. 4a). Low salinity water import was stronger at the lower low tide (+35 cm) than at the higher low tide (+105 cm). Correspondingly, salt intrusion was stronger at the higher high tide (+239 cm) than at the lower high tide (+197 cm). Due to a comparable precipitation rate before and during the survey (Fig. 3), the river salinity was not further affected by external variables, which allows to suggest a strong influence of the tidal amplitude on salinity distribution. This is in accord with the report by Uncles and Stephens (1996) who showed that salinity intrusion was a strong function of the tidal state and a weaker function of fresh water inflow. 4.2 River runoff and nutrient regeneration Silicate [Si (OH) 4 -Si] Silicate showed the best correlation with salinity at Stn. A (r 2 = 0.879) (Fig. 8). This fact indicates that fresh water runoff was the main source of silicate to the tidal estuary. Such correlation became weaker with distance (r 2 = and r 2 = at Stns. B and Y3, respectively) as a result of a partial and progressive mixing. Occasionally (as between 2:00 and 4:00, Fig. 4c), silicate concentration was higher at Stn. Y3 than at Stn. B. The vertical profile at Stn. Y3, which we monitored every 2 h (Montani and Magni, in preparation), revealed that the increase of all nutrient concentrations was most significant at the surface water and related to the effect of the tidal export from the intertidal flat. We can infer that the high silicate concentration at Stn. Y3 depended on a delayed effect of the earlier ebb flow, when the tidal flood was still lowering the concentration at Stn. B. High concentrations of the other nutrients were similarly found during the same period (Figs. 5a, b and c). Irrespective of the station, plot of silicate against salinity was significant (r 2 = 0.836, not shown) and essentially linear, indicating the conservative behavior of silicate. In our study no significant addition (Balls, 1994) or removal (Edwards and Liss, 1973; Liss and Pointon, 1973) of silicate was observed within the estuary. Fig. 7. Temperature/salinity plots for individual stations (Stn. A: ; Stn. B: ; Stn. Y3: ). Nutrient Dynamics in a Tidal Estuary 71

8 4.2.2 Nitrate + Nitrite nitrogen [(NO 3 + NO 2 )-N] Nitrate + nitrite nitrogen had the best correlation with salinity at Stn. A (r 2 = 0.796) (Fig. 9). This correlation was low (r 2 = 0.381) at Stn. Y3. Irrespective of the station, correlation with salinity was not significant (r 2 = 0.555, not shown). Balls (1994) reported conservative mixing particularly for nitrate + nitrite nitrogen concentration, as evidence of its riverine origin. Also Page et al. (1995) found that nitrate nitrogen concentration decreased with increasing salinity, reflecting the dilution of nitrate enriched fresh water by tidal sea water. In our study, such a pattern was partly confirmed by a progressively broader regression line and a weaker significance of the plots of nitrate + nitrite against salinity. Due to the normoxic condition at the nearbottom water of the intertidal flat (Fig. 6), we also suggest that oxidation of ammonium nitrogen, produced within the intertidal zone, may enter the pool of nitrate + nitrite nitrogen Ammonium nitrogen (NH 4 + -N) No significant ammonium nitrogen import through Fig. 8. Silicate/salinity plots for individual stations (Stn. A: ; Stn. B: ; Stn. Y3: ). Square regression lines for individual stations are indicated. Fig. 9. Nitrate + nitrite nitrogen/salinity plots for individual stations (Stn. A: ; Stn. B: ; Stn. Y3: ). Square regression lines for individual stations are indicated. 72 S. Montani et al.

9 Fig. 10. Ammonium nitrogen/salinity plots for individual stations (Stn. A: ; Stn. B: ; Stn. Y3: ). Square regression lines for individual stations are indicated. fresh water inflow was observed during this survey. Accordingly, plots of ammonium nitrogen against salinity showed no significant correlations, particularly at Stns. B and Y3 (Fig. 10). The correlation result was positive (r 2 = 0.546) at Stn. A and inverse at Stns. B and Y3. Ammonium nitrogen concentration at the three stations was low with low salinity, but generally high with mid-salinity (between 25 psu and 27 psu), with a wide range at salinity values over 28 psu. Ammonium nitrogen concentration was higher than at Stn. A on the lower part of the transect line. Previous studies (Balls, 1992; Page et al., 1995) have shown that ammonium nitrogen is regenerated within the estuary rather than imported. Balls (1992) attributed a broad mid-estuarine maximum during the summer months to a benthic input from decomposing materials in the sediments. Page et al. (1995) indicated as a source of ammonium nitrogen, dissimilatory reduction of nitrate, entering the marsh in stream flow under anaerobic conditions, and microbial ammonification of organic materials in marsh sediments. In our study, we suggest that the intertidal origin of ammonium is caused, to some extent, by benthic biological processes of nitrogen excretion by the macrobenthos. This consideration is related to the dominant presence on this intertidal flat of two bivalves, Ruditapes philippinarum and Musculista senhousia, whose density and biomass have been monitored monthly since April On May 30th, the two bivalves were found in amounts up to individuals m 2 and 60 gdw m 2 (Magni et al., in preparation). In order to quantify the contribution of the macrobenthos on the nutrient budget of this estuary, we carried out complementary laboratory experiments on the nutrient excretion rate by these bivalves. The bivalves excreted an average of ca. 20 µm gdw 1 h 1 (at 20 C) of ammonium nitrogen (Magni et al., in preparation), which would result in 1200 µm of ammonium nitrogen m 2 h 1 released from the intertidal flat as an excretion product on May 30th. This estimation supports our hypothesis about the possible significant contribution of the macrobenthos on the processes of nutrient regeneration Phosphate (PO 4 3 -P) Balls (1994) showed that conservative mixing of nutrients, particularly phosphate, was a function of estuarine flushing time, as related to the particle-water interaction and chemical speciation (Froelich, 1988). He found an increase of phosphate concentration for mid-salinity, followed by a seaward decrease, most evident in turbid and slowly flushed estuaries. For the River Forth (about 40 km long), this fact was attributed to a release from suspended particles or reduced sediments, within the phosphate buffer mechanisms. In our study, we found no significant correlations between phosphorus and salinity, but a broad distribution of concentration along the salinity gradient (Fig. 11). The short distance of our transect line and the coincident sampling time at individual stations would make the processes of chemical speciation of phosphate less effective, as indicated by Balls (1994). We suggest that this biophilic element should also be partly regenerated through biological processes of nutrient regeneration on the intertidal flat (Magni et al., in preparation), and not solely exported by the river runoff. In our laboratory experiments, ca. 5 µm gdw 1 h 1 (at 20 C) of phosphate was excreted by the bivalves, which would result in 300 µm of phosphate m 2 h 1 released from the intertidal zone on May 30th. Nutrient Dynamics in a Tidal Estuary 73

10 Fig. 11. Phosphorus/salinity plots for individual stations (Stn. A: ; Stn. B: ; Stn. Y3: ). Square regression lines for individual stations are indicated. 4.3 Stoichiometric nutrient balance of the tidal estuary According to the tidal cycle and the nutrient distributional patterns, we divided the time-series survey into two flood and two ebb periods (Table 1). Elemental ratios of the nutrient concentrations, N/P, Si/N and Si/P, were determined to assess the temporal and spatial variations of the nutrient stoichiometric balance of this estuary. As Table 1 shows, during the first flood (10:00 12:00) the N/P ratio was relatively higher at Stn. A (maximum of 30.0 at 10:00) than at the other stations. This situation was due to the riverine import of nitrate + nitrite (Fig. 5a), but neither of ammonium (Fig. 5b) nor phosphate (Fig. 5c). With the next ebb (13:00 20:00) this ratio dropped significantly at Stn. A. This was caused by the strong variations of both ammonium and phosphate concentrations which were the lowest (Fig. 5b) and the highest (Fig. 5c), respectively. During this period, on the other hand, the N/P ratio increased at Stn. B and resulted significantly higher than that at Stn. A (Table 1). At Stn. B, the riverine import of nitrate + nitrite was accompanied by a progressive increase of the ammonium concentration up to a maximum of 43.3 µm at 20:00 (Fig. 5b). During this low tide period, the water level at Stn. B was reduced to few tens of centimeters, which possibly resulted in a closer detection of near-bottom processes, including the metabolic activity of the macrobenthos. The second flood represented the period of maximum mixing of the concentrations. At Stn. Y3, concentrations of individual nutrients were lowered by sea water intrusion, which was strongest between 23:00 and 1:00 (Figs. 5 and 6). On the other hand between 2:00 and 4:00, high salinity water with high nutrient concentrations (Figs. 4c and 5) was sampled at this station. As we will show in detail in a subsequent paper (Montani and Magni, in preparation), the vertical profile of Stn. Y3, monitored every 2 h, indicated that such a reverse nutrient distributional pattern between Stns. Y3 and B, with higher concentrations at Stn. Y3 than at Stn. B (Figs. 4c and 5), was however due to a late effect of the tidal export at Stn. Y3. It was stronger at the surface layer and probably related to water masses and tidal current displacements between Stns. Y3 and B. Accordingly, at Stn. Y3, the two contrasting situations (23:00 to 1:00 and 2:00 to 4:00) found during this flood determined the wide fluctuation of the N/P ratio from 7.4 (23:00) to 22.8 (2:00) and resulted in a high mean ratio of 17.3 ± 5.2 (Table 1). Vice versa at Stn. B, the N/P ratio dropped to values similar to those at Stn. A (Table 1). At the second ebb, the tidal amplitude was more limited than that at the first one (Fig. 2). This governed the less marked differences between stations, particularly at Stns. Y3 and B closest to each other. Accordingly, the N/P ratio did not reveal significant distributional patterns (Table 1). Si/N ratio was high at Stn. A, most significantly during the first flood (10:00 12:00) and the first ebb (13:00 20:00) (Table 1). This was the result of the stronger riverine input of silicate (Fig. 8) than that of the nitrogen pool (Figs. 9 and 10). During the second flood, all nutrient concentrations decreased with high salinity water intrusion. This gave rise to a much more balanced elemental ratio at Stn. A, as we found at both Stns. B and Y3 (Table 1) and on a transect line of the subtidal zone (Magni et al., in preparation), which approximated the Redfield et al. (1963) ratio. No dramatic changes were observed with the second ebb, but values decreased progressively from Stn. A to Stn. Y3 (Table 1), as related to the riverine import of silicate to the estuary and nitrogen regeneration within estuary. 74 S. Montani et al.

11 Table 1. Temporal variations of the nutrient stoichiometric balance at individual stations according to the tidal status. Tidal status N/P (ratio) Si/N (ratio) Si/P (ratio) Stn. A Stn. B Stn. Y3 Stn. A Stn. B Stn. Y3 Stn. A Stn. B Stn. Y3 Flood (10 h 12 h) (±5.4) (±4.4) (±8.4) (±0.2) (±0.1) (±0.2) (±4.3) (±1.7) (±4.5) Ebb (13 h 20 h) (±1.8) (±2.6) (±3.0) (±0.3) (±0.1) (±0.1) (±4.5) (±3.6) (±3.6) Flood (21 h 4 h) (±2.1) (±1.5) (±5.2) (±0.4) (±0.0) (±0.3) (±5.8) (±0.9) (±3.0) Ebb (5 h 10 h) (±2.3) (±2.3) (±1.3) (±0.2) (±0.1) (±0.1) (±2.8) (±3.2) (±0.9) At Stn. A, the Si/P ratio was highest during the first three hours of the survey (Table 1), due to the riverine input of silicate (Fig. 4c) which did not coincide with high phosphate concentration (Fig. 5c). At both Stns. A and B from the ebb period (13:00 20:00) to the next flood (21:00 4:00), the Si/P ratio decreased markedly to a similar (Stn. A) or lower (Stn. B) level than that at Stn. Y3 (Table 1). During the second flood, the reverse nutrient distributional pattern between Stns. Y3 and B (2:00 and 4:00, Figs. 4c and 5) was highlighted by both N/P and Si/P ratios, which were higher at Stn. Y3 that at Stn. B (Table 1). During the second ebb at Stn. B, wide fluctuations of both salinity (Fig. 4b) and nutrient concentrations (Figs. 4a and 5) were found. The limited tidal amplitude of this ebb (Fig. 2) contributed to mantaining on the intertidal flat a more mixed situation of water masses of different origin and to determine less marked distributional differences among stations. As a result, the N/P mean ratio was similar at Stn. B to that at Stn. Y3, while the Si/P ratio was more similar to that at Stn. A. Overall, the temporal and spatial variations of these ratios were as strong as those of individual nutrients. Unbalanced N/P, Si/N and Si/P ratios depended significantly on the riverine import of silicate and nitrate + nitrite nitrogen. Moreover, the tidal amplitude was closely related to the nutrient distributional pattern: the stronger tidal amplitude during the first part of the survey determined more marked spatial differences along the transect line, while a weaker ebb during the second part of the survey related to a longer residence time of water masses of different origin at individual stations which delayed their mixing. The ratios more closely approximated the Redfield ratio (N:Si:P = 16:16:1) at Stns. B and Y3 than at Stn. A. The hypothesis that the macrobenthos plays a significant role in the nutrient balance of this estuary, as discussed, would further enhance the daily variability of the stoichiometric nutrient balance of this estuary along the water column. 5. Conclusion The nutrient balance of the investigated estuary was rapidly influenced by the tidal cycle, as indicated by the strong changes of both nutrient concentrations and their ratios at individual stations on a time scale of 1 2 h. The effect of the tidal amplitude was important in determining the extent of the variations in nutrient concentrations at all three stations, which were stronger between the lower low tide and the higher high tide. Main sources of nutrients came either from the fresh water runoff (silicate, and partially nitrate + nitrite and phosphate) or the intertidal flat (ammonium, and partially nitrate + nitrite and phosphate), but not the subtidal zone. Further studies are in progress to estimate the contribution of the macrobenthos to the processes of nutrient regeneration within the estuary. Acknowledgements We wish to extend our heartfelt thanks to Dr. K. Tada, Associate Professor of Kagawa University, Mr. T. Hamazaki, captain of the research vessel Calanus, and all the members of the Marine Biochemistry Laboratory, Kagawa University, for their cooperative and tireless dedication during sampling activities and the sample handling and treatment. References Balls, P. W. (1992): Nutrient behaviour in two contrasting Scottish estuaries, the Forth and the Tay. Oceanologica Acta, 15, Balls, P. W. (1994): Nutrient inputs to estuaries from nine Scottish east coast rivers; influence of estuarine processes on inputs to the North Sea. Estuar. Coast. Shelf Sci., 39, Edwards, A. M. C. and P. S. Liss (1973): Evidence for buffering of dissolved silicon in freshwaters. Nature, 243, Froelich, P. N. (1988): Kinetic control of dissolved phosphate in natural rivers and estuaries: a primer on the phosphate buffer mechanism. Limnol. Oceonogr., 33, Nutrient Dynamics in a Tidal Estuary 75

12 Hernandez-Ayon, J. M., M. S. Galindo-Bect, B. P. Flores-Baez and S. Alvarez-Borrego (1993): Nutrient concentrations are high in the turbid waters of the Colorado river Delta. Estuar. Coast. Shelf Sci., 3, Liss P. S. and M. J. Pointon (1973): Removal of dissolved boron and silicon during estuarine mixing of sea and river waters. Geochim. Cosmochim. Acta, 37, Page, H. M., R. L. Petty and D. E. Meade (1995): Influence of watershed runoff on nutrient dynamics in a southern California salt marsh. Estuar. Coast. Shelf Sci., 41, Redfield, A. C., B. H. Ketchum and F. A. Richard (1963): The influence of organisms on the composition of sea water. p In The Sea, Vol. 2, ed. by M. N. Hill, John Wiley, New York. Renshun, Z. (1992): Suspended sediment transport processes on tidal mud flat in Jiangsu Province, China. Estuar. Coast. Shelf Sci., 35, Schubel, R. J. and D. W. Pritchard (1986): Responses of upper Chesapeake Bay to variations in discharge of the Susquehanna river. Estuaries, 9, Strickland, J. D. H and T. R. Parsons (1972): A Practical handbook of seawater analysis. Bull. Fisher. Res. Board Canada 167, 310 pp. Uncles, R. J. and J. A. Stephens (1996): Salt intrusion in the Tweed Estuary. Estuar. Coast. Shelf Sci., 43, Vörösmarty, C. J. and T. C. Loder (1994): Spring-neap tidal contrasts and nutrient dynamics in a marsh-dominated estuary. Estuaries, 17, Yin, K., P. J. Harrison, S. Pond and R. J. Beamish (1995a): Entrainment of nitrate in the Fraser river estuary and its biological implications. II. Effects of spring vs. neap tide and river discharge. Estuar. Coast. Shelf Sci., 40, Yin, K., P. J. Harrison, S. Pond and R. J. Beamish (1995b): Entrainment of nitrate in the Fraser river estuary and its biological implications. III. Effects of winds. Estuar. Coast. Shelf Sci., 40, S. Montani et al.