Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary

Estuarine, Coastal and Shelf Science xxx (2015) 1e10

Contents lists available at ScienceDirect

Estuarine, Coastal and Shelf Science journal homepage: www.elsevier.com/locate/ecss

Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary Clarisse Odebrecht a, Paulo C. Abreu a, Jacob Carstensen b, * a b

Institute of Oceanography, Federal University of Rio Grande (FURG), Brazil Department of Bioscience, Aarhus University, Denmark

a r t i c l e i n f o

a b s t r a c t

Article history: Accepted 3 March 2015 Available online xxx

In this study it was hypothesised that increasing water retention time promotes phytoplankton blooms in the shallow microtidal Patos Lagoon estuary (PLE). This hypothesis was tested using salinity variation as a proxy of water retention time and chlorophyll a for phytoplankton biomass. Submersible sensors fixed at 5 m depth near the mouth of PLE continuously measured water temperature, salinity and pigments fluorescence (calibrated to chlorophyll a) between March 2010 and 12th of December 2011, with some gaps. Salinity variations were used to separate alternating patterns of outflow of lagoon water (salinity <8; 46% of the time) and inflow of marine water (salinity >24; 35% of the time). The two transition phases represented a rapid change from lagoon water outflow to marine water inflow and a more gradually declining salinity between the dominating inflow and outflow conditions. During the latter of these, a significant chlorophyll a increase relative to that expected from a linear mixing relationship was observed at intermediate salinities (10e20). The increase in chlorophyll a was positively related to the duration of the prior coastal water inflow in the PLE. Moreover, chlorophyll a increase was significantly higher during austral spring-summer than autumn-winter, probably due to higher light and nutrient availability in the former. Moreover, the retention time process operating on time scales of days influences the long-term phytoplankton variability in this ecosystem. Comparing these results with monthly data from a nearby long-term water quality monitoring station (1993e2011) support the hypothesis that chlorophyll a accumulations occur after marine inflow events, whereas phytoplankton does not accumulate during high water outflow, when the water residence time is short. These results suggest that changing hydrological pattern is the most important mechanism underlying phytoplankton blooms in the PLE. © 2015 Elsevier Ltd. All rights reserved.

Regional index terms: Brazil Rio Grande do Sul Patos Lagoon South Atlantic ocean Keywords: saltwater inflow light limitation biomass accumulation choked lagoon sediment resuspension Geographic coordinates: 32
080 10.0000 S 52
060 09.0000 W

1. Introduction Phytoplankton variability in coastal ecosystems is determined by diverse range of factors and complex interacting site-specific processes (Cloern, 2001; Cloern and Jassby, 2010; Gallegos et al., in this issue). Differences among ecosystems may result from physical, geomorphological and hydrodynamic characteristics, but also nutrient enrichment, climatology and human disturbances. In many estuaries and coastal lagoons, phytoplankton biomass and species composition variability are strongly associated with hydrodynamics,

* Corresponding author. Department of Bioscience, Aarhus University, Frederiksborgvej 399, 4000 Roskilde, Denmark. E-mail address: [email protected] (J. Carstensen).

when basic growth requirements (light and nutrients) are plenty (Peierls et al., 2012; Thompson et al., in this issue). Coastal lagoons are shallow, dynamic and highly productive ecosystems separated from the ocean by a sand barrier that is penetrated by one or several channels allowing water exchange with the ocean. Coastal lagoons are classified as choked, restricted or leaky according to their degree of water exchange with the ocean (Kjerfve, 1986). Due to the restricted water exchange, choked lagoons have comparatively long water retention time and high phytoplankton biomass (Knoppers et al., 1991; Roselli et al., 2013). The Patos Lagoon in Southern Brazil (Fig. 1), is the largest (10,360 km2) choked coastal lagoon in the world (Kjerfve, 1986). The connection with the coastal ocean in the southern area of the Patos Lagoon has typical estuarine conditions, which influence the growth and distribution of all biota from primary producers to fishes (Seeliger et al., 1997; Odebrecht et al., 2010). In the Patos

http://dx.doi.org/10.1016/j.ecss.2015.03.004 0272-7714/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004

2

C. Odebrecht et al. / Estuarine, Coastal and Shelf Science xxx (2015) 1e10

Fig. 1. Geographical location showing the watershed of the lagoons Patos and Mirim in the Southwest Atlantic Ocean and the sampling station located at the mouth of the Patos Lagoon Estuary near the city of Rio Grande.

Lagoon Estuary (PLE), phytoplankton growth and dynamics are strongly light-limited and only partially influenced by nutrients (Abreu et al., 1994a, 1995; Haraguchi et al., in this issue) and grazing (Abreu et al., 1994b). However, hydrology is considered the key forcing function of phytoplankton variability at both shorter and longer time scales (Abreu et al., 2010). This strong connection to the hydrodynamics in the PLE makes phytoplankton variability in this estuary highly unpredictable on shorter time scales. However, peaks of chlorophyll a were frequently observed in oligohaline-mesohaline waters leaving the estuary after a certain period of coastal water entering the lagoon (Fujita and Odebrecht, 2007; Abreu et al., 2010), and these authors suggested that elevated chlorophyll a concentrations in this salinity range were likely the result of phytoplankton accumulation as a response to increasing water retention time. In this study it is hypothesised that the occurrence of short-term phytoplankton blooms in the shallow microtidal PLE are linked to the increasing water retention time. To test this hypothesis a statistical approach was applied to identify such periods with enhanced retention time from continuous time series of salinity and chlorophyll a monitored in the main navigation channel of the PLE for almost two years. In this study a phytoplankton bloom is defined as a significant increment of chlorophyll a measured in the short time scale (hours e days), hypothesised to result from phytoplankton biomass accumulation generated by increasing estuarine water retention time. The duration of coastal water intrusion into the PLE, characterised by high salinity in the navigation channel, was used as a proxy of water residence time. The aim of the study was to investigate if the magnitude of the phytoplankton bloom increased with the duration of coastal water inflow. This relatively short-term, although high-resolution, investigation was compared with data from a long-term (21 years) monitoring program to assess the general applicability of the results. 2. Study area The Patos Lagoon is a choked and shallow system (average depth 5 m) with a watershed of approximately 200,000 km2 extending 250 km along the microtidal coastline of the southern Brazilian plain (10,360 km2; 30
120 - 32
120 S; 50
400 - 52
150 W) (Fig. 1). The

Patos Lagoon is mainly oligohaline, whereas large salinity variability characterises the Patos Lagoon Estuary (PLE; about 1000 km2). In 1998, the PLE became a monitoring site of the Brazilian Long-Term Ecological Research Program (BR-PELD), and water quality and phytoplankton have been continuously sampled since 1993 (Abreu et al., 2010; Odebrecht et al., 2010). The hydrology of this ecosystem is mainly driven by wind strength and direction, rainfall and evaporation (annual water surplus 200e300 mm, Klein, 1997), whereas tides are negligible. The PLE exchanges water with the Atlantic Ocean through a narrow inlet (0.5e1.0 km wide; 14e18 m depth) that is frequently dredged for shipping purposes. In general, southerly winds push the coastal water into the PLE, whereas the dominant winds from the north (mainly NE), acting along the main axis, forces the water from the € ller et al., 2001). lagoon through the PLE to the coastal region (Mo During flood periods the seaward flow through the PLE, driven by large river discharge, is only reversed by strong south to westerly winds. Larger saltwater intrusions are typically associated with the passage of atmospheric fronts lasting from 6 to 10 days. Strong southerly winds can persist for 2e4 days, retaining the water in the PLE with increasing retention time as a consequence. Owing to changing winds and freshwater discharge, large shortterm salinity oscillations occur in the PLE, on the time scale of hours to days, alternating between fresh-oligohaline, over mesohaline to truly marine water; these strong variations influence the phytoplankton composition and abundance (Fujita and Odebrecht, 2007; Abreu et al., 2010). Except during drought periods, phytoplankton primary production in the Patos Lagoon is largely light limited, due high turbidity resulting from high seston input from land and sediment resuspension (Odebrecht et al., 2005). Depths in the PLE are less than 5 m, except for the main channel, and large shallow shoals (<1.5 m) prevail near the margins. However, even in these shoals, light limitation in the water column prevails from May to August, and sometimes also in austral spring/summer months (Abreu et al., 1994a). 3. Material and methods Submersible sensors for in situ continuous monitoring of water temperature and salinity (RBR Ltd.) were deployed at 5 m depth at

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004

C. Odebrecht et al. / Estuarine, Coastal and Shelf Science xxx (2015) 1e10

the Brazilian Navy pier (water depth 15 m) in the main inlet where the channel to the PLE is narrowest (800 m) (Fig. 1). In general the water column of the PLE main navigation channel, where sensors were deployed, is well mixed as shown by Abreu et al. (2010; Figs. 3 and 4). However, a pycnocline develops for a short period during the entrance of the salt wedge and may reach the sensor (Abreu et al., 2010, Fig. 4). Bi-directional flows are rarely observed in the channel, although freshwater outflow in the western side and saltwater inflow through the eastern side of the channel has been € ller and Castaing, 1999; Fernandes et al., 2005). described (Mo A fluorescence sensor (Turner Cyclops) was connected to the RBR data logger and programmed to collect data every minute. Every week the sensors were brought to the surface and logged data transferred to a computer. Moreover, monthly maintenance of the sensors in the laboratory was necessary (anti-corrosive paint, removing barnacles, etc.). In the present study, data from an almost two-years period (March 2010eDecember 2011) were analysed. Despite regular maintenance there were minor outfalls on the sensors causing gaps in the time series of several days to weeks. The continuous data from the sensors were aggregated to hourly observations by averaging. Water samples were collected during maintenance visits, filtered (Whatman GF/F Ø25 mm), and analysed for chlorophyll a in the laboratory using a Turner TD-700 Fluorometer after 24 h of dark extraction in acetone 90% without acidification (Welschmeyer, 1994). The continuous fluorescence measurements were translated into chlorophyll a using a calibration curve estimated from simultaneous observations of fluorescence and chlorophyll a (Fig. 2). Both chlorophyll a and the voltage from the fluorescence sensor were log-transformed and the calibration curve was established as a linear regression with errors associated with both variables (type II regression).

3.1. Identification of PLE outflow periods Alternating regimes of inflow and outflow conditions were identified from the continuous salinity sensor using appropriately chosen salinity thresholds for partitioning the time series into these two dominating flow patterns and transitions between them. First, a moving median filter of 23 h (smoother) was applied to the hourly

Fig. 2. Calibration of measured chlorophyll a versus fluorescence given by the voltage signal (type II regression). Consequently, coefficients of determination are listed for both variables. Note the log-scale on both axes.

3

salinity observations to remove short-term fluctuations (typically 2e4 h), caused by changing wind conditions. These fluctuations were short and generally representing movement of water in the inlet channel, rather than inflow and outflow conditions of the Patos Lagoon. Second, salinities less than 8 typically characterized conditions of outflowing water from the Patos Lagoon, salinities above 24 typically characterised inflowing water from the South Atlantic coast, and intermediate salinities between 8 and 24 represented the transition phase between the two alternating regimes. These salinity thresholds were found appropriate to robustly separate transitions between the two general flow regimes, whereas short-term events of changing flow patterns, not representing a change between the dominating flow regimes, resulted more frequently if lower or higher threshold values were chosen. Three different hydrological regimes were identified from the median-filtered salinity signal: 1) outflow from the Patos Lagoon (salinity <8), 2) inflow from the South Atlantic coast (salinity>24), and 3) outflow from the PLE in the transition phase between inflow and outflow from the Patos Lagoon (salinity decreasing from >24 to less than 8). The latter regime represented flushing of the PLE, typically lasting a few days, whereas the change from outflow to inflow was very rapid (a few hours). In addition, there were periods with short-term inflows and outflows where salinity changed from low (or high) to intermediate salinity and then back to low (or high). These short-term interruptions were included as part of the general outflow (or inflow) conditions. Since periods of inflows interrupt the prevailing outflow conditions from the Patos Lagoon and increase the retention time in the PLE, changes in chlorophyll a with decreasing salinity were investigated during the transition from marine inflow to lagoon outflow. The phytoplankton accumulation in the PLE was analysed using observations during outflow conditions from the PLE (intermediate salinities) combined with the six hours of high salinity before the onset of the outflow (South Atlantic coastal water) and the six hours of low salinity after the outflow (Patos Lagoon water). The six hours before and after the PLE outflow were used to characterise the water masses of the two dominating flow regimes; six hours was sufficiently long to obtain relatively precise estimates and sufficiently short to be representative of the conditions just before and after the PLE outflow. First, a linear regression of chlorophyll a versus salinity using the six hours before and after the PLE outflow was estimated to assess the conservative mixing relationship between the two end members; Patos Lagoon and South Atlantic coastal water. If conservative mixing dominated, observations during the PLE outflow would follow the same mixing relationships. Therefore, deviations from this conservative mixing pattern were analysed by fitting a spline model with three degrees of freedom (spline mixing curve) to the residuals from the conservative mixing curve and including predicted deviations from this linear relationship from the observations during the PLE outflow. For each event of PLE outflow the maximum deviation of the spline curve from the linear conservative mixing was identified and characterised by its salinity (Smaxdev) and the difference in chlorophyll a from the spline and linear mixing curves (Chlamaxdev). In addition the mean difference between the spline and linear mixing curves (Chlameandev) and the duration of the PLE outflow (TPLEoutflow) was calculated (number of hours to change the salinity from >24 to less than 8). The three latter characteristics of the PLE outflows were analysed and compared to the duration of inflow before the PLE outflow (Tinflow), constituting a proxy for the retention time in the PLE. It was further investigated if these relationships were season-specific by partitioning the PLE outflow events into austral spring-summer (22 September e 20 March) and autumn-winter (21 March e 21 September). Finally, the relationship between

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004

4

C. Odebrecht et al. / Estuarine, Coastal and Shelf Science xxx (2015) 1e10

Fig. 3. Four examples of identification of PLE outflow conditions from 2010 (A) and 2011 (BeD). Measured salinity is dashed grey line, median-filtered salinity is solid black line, and chlorophyll a is shown as green line. Thresholds of salinity used to separate the different regimes are shown as dashed blue lines. The different hydrological regimes are separated by dotted red lines and labelled in yellow as 1) outflow from the Patos Lagoon, 2) inflow of South Atlantic coastal water, and 3) outflow from the PLE. The change from outflow to inflow is also indicated but not labelled. All examples cover one week of data.

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004

C. Odebrecht et al. / Estuarine, Coastal and Shelf Science xxx (2015) 1e10

5

Fig. 4. The estimated linear and spline mixing curves for the four PLE outflow examples from Fig. 3. The six hours before and after the PLE outflow are used in the linear regression, whereas intermediate salinities are also included for the spline curve. Significance of the two types of mixing curves is inserted.

Chlamaxdev and Smaxdev was examined to identify salinity regimes favourable to larger blooms in the PLE. 3.2. Long-term monitoring station The general applicability of the results obtained from the two years with continuous measurements at the buoy was assessed using a long-term monitoring station located nearby in the inlet channel to the PLE. At this station, the Secchi depth has been recorded and water samples have been collected from the surface (0.5 m) on a monthly basis since 1993. Water samples were brought to the laboratory and analysed for dissolved inorganic nutrients (DIN ¼ ammonia þ nitrate þ nitrite, DIP ¼ phosphate, and DSi ¼ silicate), suspended particulate matter (SPM), and chlorophyll a according to standard protocols and described in Abreu et al. (2010). Changes in water quality across the salinity range were examined for the entire period and specifically for the two years with buoy data. A non-parametric general additive model (GAM) was fitted to the water quality data using all observations to assess the overall relationship with salinity. Finally, water quality samples were matched in time with the hydrological regimes found using the buoy data to identify specific water quality samples representing PLE outflows. 4. Results The salinity, temperature and fluorescence sensors were operational from 11th of March 2010 to 12th of December 2011, but there were several gaps in the time series ranging from a few hours up to almost two months. Throughout the entire time series salinity displayed the expected characteristic alternating pattern between outflow and inflow conditions to the PLE. Salinities in the inlet

channel to the PLE indicated that outflow from Patos Lagoon dominated (salinity < 8; 46% of the time), inflows of South Atlantic coastal water were also quite frequent (salinity >24; 35% of the time), and that transitions between these two main hydrological regimes accounted for 19%. Chlorophyll a, calibrated from the fluorescence sensor, varied between 0.6 and 206 mg L1 with an average of 6.4 mg L1. Common patterns of chlorophyll a response to input of saltwater in the estuary (Fig. 3) were the parallel increase of chlorophyll a and salinity (Fig. 3a), and most common an increase of chlorophyll a in oligo-mesohaline water after some period of saltwater intrusion (Fig. 3b, c, d). Using the algorithm for separating different hydrological regimes, a total of 28 PLE outflow events were identified, lasting between 5 and 55 h (TPLEoutflow). Most of these outflow events followed the same pattern (Fig. 3), starting with a relatively long period of outflow from the Patos Lagoon (hydrological regime #1), followed by a rapid change to South Atlantic coastal water inflow (hydrological regime #2), and then a gradually declining salinity during outflow from the PLE (hydrological regime #3). Although the change from low to high salinity was not always rapid, occasionally giving rise to longer periods of intermediate salinities (e.g. Fig. 3D), this hydrological regime was not considered in the present study. Most of the PLE outflow events (26 out of 28) had a significant departure from a linear mixing relationship and many of them were highly significant (18 events with P < 0.0001). However, there were differences between PLE outflow events regarding the magnitude of the departures and their distribution with decreasing salinity (Fig. 4). As an example, the PLE outflow in June 2010 (Fig. 4A) lasted 19 h with a mean deviation in chlorophyll a of Chlameandev ¼ 0.9 mg L1, and the largest departure of the spline mixing curve from conservative mixing was Chlamaxdev ¼ 2.0 mg L1, observed at a salinity of Smaxdev ¼ 20.6. For the three PLE outflow

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004

6

C. Odebrecht et al. / Estuarine, Coastal and Shelf Science xxx (2015) 1e10

examples in 2011 (Fig. 4BeD), Chlameandev was 2.0, 3.1, and 1.9 mg L1, respectively; Chlamaxdev was 4.6, 7.4, and 5.9 mg L1, respectively; and Smaxdev was 18.6, 15.6, and 18.7, respectively. The majority of PLE outflows lasted between 10 and 30 h, although two outflows took more than two days (Fig. 5A). Inflows prior to the PLE outflows generally lasted longer with broader

Fig. 5. Distributions of PLE outflow characteristics. A) Duration of outflow event (TPLEoutflow), B) mean deviation in chlorophyll a from conservative mixing (Chlamaxdev), C) maximum deviation in chlorophyll a from conservative mixing (Chlamaxdev), and D) salinity for maximum chlorophyll a deviation (Smaxdev).

spread from 12 h to more than 6 days (Fig. 5B). The mean deviation in chlorophyll a from conservative mixing had a mean of 0.7 mg L1 with a range from 5.5e5.6 mg L1 (data not shown). The max deviation in chlorophyll a from conservative mixing had a more right-skewed distribution with a mean of 1.5 mg L1 and ranging up to 13.0 mg L1 (Fig. 5C). Approximately one third of the 28 PLE outflows had lower chlorophyll a than expected from conservative mixing (negative values of Chlameandev and Chlamaxdev), but the general tendency was an increase in chlorophyll a relative to that expected from conservative mixing. The largest deviations in chlorophyll a typically occurred at intermediate salinities between 10 and 20 (Fig. 5D). Only 25 of the 28 PLE outflows could be associated with duration of the previous inflow period (Tinflow) due to data gaps. The duration of the PLE outflow was positively correlated, although the relationship was not strong, to Tinflow (Fig. 6A), confirming the expected relationship that flushing time of the brackish PLE water depended on the amount of South Atlantic coastal water having

Fig. 6. PLE outflow characteristics versus duration of the previous inflow period. A) Duration of outflow event (TPLEoutflow), B) mean deviation in chlorophyll a from conservative mixing (Chlamaxdev), and C) maximum deviation in chlorophyll a from conservative mixing (Chlamaxdev). In B) and C) season-specific regressions are shown. Note that the durations of the PLE outflow and the previous inflow are log-transformed.

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004

C. Odebrecht et al. / Estuarine, Coastal and Shelf Science xxx (2015) 1e10

entered the PLE. Both the mean and max of the chlorophyll a deviations from linear conservative mixing increased with Tinflow (Fig. 6B,C). Particularly, the largest increases in chlorophyll a were observed in October and November, and typically for Tinflow exceeding 2e3 days. In fact, increases in Chlameandev with Tinflow were significantly higher during austral spring-summer than autumn-winter (difference in intercepts of ~1.3 mg L1), whereas Chlamaxdev increased with Tinflow at a faster rate (slope) during spring-summer. There was also a tendency towards larger increases in chlorophyll a relative to conservative mixing with lower salinity, but this relationship did not change with the two investigated seasons (data not shown). Only three water quality samples from the regular monthly monitoring program could be matched with periods of PLE outflow (Fig. 7A). The matching sample with a previous inflow period of 87 h had a relatively high chlorophyll a, though the number of matching samples is too low to make any statistical inference. There was a slight gradual increase in chlorophyll a across the entire salinity range without any excursions at intermediate salinities; this intermediate range including PLE outflows as well as transitions from Patos Lagoon outflow to inflow and short-term interruptions of these two hydrological regimes. Inorganic nutrient concentrations generally declined with salinity, although DSi levels in the fresh-oligohaline water were lower than in the mesohaline water (Fig. 7BeD). The SPM concentration was high at low and high salinities and reached a minimum concentration at salinities between 5 and 10 (Fig. 7E). In fact, the turbidity of the water was reduced by ~50% at these intermediate salinities. Changes in SPM across the salinity range had a large effect on Secchi depths, increasing from 0.2 m at low salinity to 1 m for salinities around 10 before reaching 0.7 m at high salinity (Fig. 7F). Overall, the water quality data for the two years with buoy data (2010e2011) were similar to the long-term variations. 5. Discussion The single narrow channel that connects choked lagoons with the open coast functions as a “dynamic filter”, restricting the influence of tides in such systems. In these lagoons, the hydrology is mainly governed by the action of wind (Kjerfve, 1994). This is also the case of the Patos Lagoon Estuary (PLE), where winds strongly influence the phytoplankton short-term variability (hours to days), whereas rainfall affects the phytoplankton on larger time scales (seasonal; inter-annual). The rather random nature of meteorological forcing makes the prediction of phytoplankton variation very difficult, even with long-term data on phytoplankton and abiotic factors (Abreu et al., 2010). Despite the inherent high variability, some patterns emerged from short-term intensive studies on the phytoplankton in the PLE. In one of these, phytoplankton samples were taken each 2.5 h during 30 days in austral spring (OctobereNovember 2004) and 30 days in JanuaryeFebruary 2004 (austral summer) and it was recognized that phytoplankton biomass accumulation in the estuarine region was coupled to the wind direction and the freshwater outflow (Fujita and Odebrecht, 2007). The authors concluded that the phytoplankton blooms in austral spring resulted from the maintenance of species from the coastal region in the estuary during periods longer than the cell division time, provided the availability of light and nutrients. During summer, in spite of longer coastal water intrusion and high retention time, higher phytoplankton biomass was rarely observed, probably because the phytoplankton production was balanced by losses caused by grazing and/or sedimentation, or simply exhaustion of nutrients. The other intensive study measured phytoplankton biomass every 3 h for one week in each of the four seasons in the main channel (Abreu

7

et al., 2010) and the authors suggested that higher chlorophyll a values observed at short time scales would result from phytoplankton biomass accumulation due to the water retention in the estuary after longer saltwater inflow. These observations lead to the hypothesis tested in this study that the phytoplankton biomass accumulation in the PLE would result from the changing the delicate balance between phytoplankton biomass doubling time and the water retention time. Fig. 8 exemplifies how these two rates could be related in different hydrological conditions. During the passage of atmospheric fronts, southerly winds push the coastal water through the narrow channel of the PLE, blocking the estuarine water outflow (Fig. 8A). Increasing the water residence time inside the estuary would allow the accumulation of phytoplankton cells only if 1) losses are negligible (Lucas et al., 2009) and 2) phytoplankton doubling time is shorter than the water retention time. The transition from southerly to northerly winds leads to a change in the barotropic pressure gradient from the estuary towards the coastal region €ller et al., 2001, 2007; Fernandes et al., 2002, 2005), increasing (Mo the estuarine outflow and export of formed biomass (Fig. 8B). In this last case, the water retention time becomes smaller than doubling time, hampering the accumulation of phytoplankton biomass. To test this hypothesis, a two years dataset collected by sensors located in the main estuarine channel was used. Fluorescence sensors collected phytoplankton biomass information each minute and several situations of phytoplankton biomass accumulation were recognized. The use of in situ sensors and automatic systems are getting more common at present and generates an enormous amount of data that allows a better interpretation of the phytoplankton dynamics on a short time scale (Chen et al., 2010; Paerl et al., 2010; Navarro et al., 2011; Hall et al., in this issue). Moreover, the analysis of continuous data during periods from weeks to years allows a better comprehension on how processes that occur in short scales (hours-days) may influence the larger scales variations that emerge after weekly, monthly, or longer sampling periodicity. Continuous data also provide information on how well traditional monitoring programs with more infrequent sampling capture phytoplankton dynamics at different time scales. In general, our results demonstrated that the amount of oligoand mesohaline water in the PLE and the time before the prevailing outflow condition increase with longer saltwater inflows. However, one must bear in mind that salinity measurements made by the sonde are not the best indicator of water masses movement. Therefore, we consider the duration of the coastal water inflow (Tinflow), determined by salinity measurements, as a proxy for the retention time in the PLE (Fig. 6). It is quite difficult to measure the phytoplankton doubling time in situ due to methodological limitations. In a mesocosm experiment conducted in the PLE, the chlorophyll a took longer to double in bags with no manipulation (nine days in the Control), compared to bags that received the addition of N and P (24 h; Abreu et al., 1994b). Thus, it would be reasonable to consider that the phytoplankton (chlorophyll a) doubling time in the PLE would vary from one to nine days, being shorter when the levels of nutrients and light are high enough to allow rapid growth of phytoplankton. In contrast to tidal estuaries characterized by periodic saltwater intrusions, our results showed that, the coastal water inflow into the PLE is neither regular in terms of frequency of occurrence, nor in the duration of the saltwater inflow (Figs. 3 and 6). Moreover, it was possible to identify two main patterns of chlorophyll a increase in response to the saltwater inflow. The chlorophyll a increase coinciding with salinity increase was probably caused by benthic microalgae resuspension or advection with the salt-wedge intrusion into the estuary, as indicated by the rapid increase of

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004

8

C. Odebrecht et al. / Estuarine, Coastal and Shelf Science xxx (2015) 1e10

Fig. 7. Water quality data versus salinity from the long-term monitoring station at the inlet channel. The solid line is the estimated GAM relationship (P-value inserted in plots) between water quality and salinity based on all data. Three observations matching PLE outflows identified from the buoy data are highlighted for chlorophyll a and labelled with the duration of the inflow period prior to the PLE outflow. Dashed

chlorophyll a, parallel to the increment of salinity (e.g. Fig. 3A) and also as previously observed (Bergesch et al., 1995; Abreu et al., 2010). The most common pattern, however, was the measurement of high chlorophyll a values in meso- and oligohaline water leaving the estuary after a longer period of saltwater inflow. The positive relationship observed between the period of coastal water inflow (TPLEinflow) and the consequent increment of chlorophyll a in oligo- and mesohaline outflowing water, indicates that the saline coastal water penetrates further upstream the estuary allowing blooms of euryhaline phytoplankton species to develop, particularly when Tinflow exceeded 2e3 days. Regarding the relationship between the phytoplankton doubling time and the water residence time, it is noteworthy that 84% of the occasions with coastal water inflow were relatively short, i.e. less than four days (Fig. 5B). The coastal water inflow lasted longer (4e6 days) in just four out of 25 occasions. The comparison of the PLE phytoplankton doubling time (1e9 days) with the periods of saltwater intrusion (TPLEinflow) indicates that the biomass accumulation would be more likely to occur during times when the phytoplankton requirements of nutrients and light are attended, i.e., mainly during austral spring and summer. This assertion is further supported by the fact that the relatively higher chlorophyll a measured during PLE outflows (Chlamaxdev) were significantly larger in austral spring and summer in comparison to austral autumn and winter. Despite the rather shallow depths in the PLE, light limitation of phytoplankton growth is likely to occur between May and August, when light intensity in the water column is usually less than the critical value (Abreu et al., 1994a). In addition to the seasonal light limitation, sporadic high seston concentrations from wind-induced sediment resuspension, may reduce phytoplankton growth even in austral spring and summer. Considering that larger part of the PLE is shallow (<2 m), light limitation would be alleviated with improved water transparency during saltwater inflows (Fig. 7F). Increasing retention time during saltwater inflows probably enhances sedimentation and reduces resuspension of particulate matter (Fig. 7E). Highest chlorophyll a values measured in oligo- and mesohaline waters in the PLE are in general dominated by euryhaline species originating from the coastal region such as the diatoms Skeletonema costatum, Chaetoceros, Thalassiosira, Ceratulina, Rhizosolenia, Pseudosolenia, which were seeded in the estuarine region with saltwater inflow (Abreu et al., 1994a; Fujita and Odebrecht, 2007; Bergesch et al., 2009; Haraguchi et al., in this issue). These fastgrowing species experience favourable growth conditions in the estuary, exceeding possible losses caused by grazing and sedimentation. In addition to seeding euryhaline species, saltwater inflow blocks the freshwater outflow from the Patos Lagoon, which increases retention time in the PLE and leads to higher phytoplankton biomass. Our results show that the retention time in estuarine systems can be an important factor influencing phytoplankton biomass; however, the importance of this mechanism is modulated by light conditions, as mentioned above, as well as the excess freshwater outflow, which depends on the rainfall and evaporation rates in the entire Patos Lagoon watershed. In general, annual mean values of chlorophyll a in the PLE were positively related to rainfall, up to a threshold of 1500 mm per year (Abreu et al., 2010). Higher amounts ~ o years, would wash out most of of rain, typically observed in El Nin the phytoplankton biomass production from the PLE to the coastal region, reducing the likelihood of phytoplankton blooms developing in the estuarine region. This observation is in accordance

lines for the inorganic nutrients indicate levels (Fisher et al., 1992) of potential phytoplankton growth limitation. Note the log-scale on all plots except Secchi depth.

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004

C. Odebrecht et al. / Estuarine, Coastal and Shelf Science xxx (2015) 1e10

9

(mainly N and P) into the systems and allow a faster phytoplankton growth. The results of our study, on the other hand, demonstrate that in the PLE the saltwater input is as important as freshwater contribution to the system, since the duration of coastal water intrusion into the estuary determines the potential for phytoplankton growth and biomass accumulation.

Acknowledgments

Fig. 8. Conceptual diagram of the retention time mechanism: A) Southerly winds induce a barotropic gradient from the ocean to the estuary (line) pushing the coastal water through the narrow communication channel; B) This leads to an increase of the estuarine water retention time (RT) that surpass the phytoplankton (black dots) doubling time (DT), allowing biomass accumulation inside the estuary. C) Change from southerly to northerly winds releases the water towards the coast, with export of accumulated biomass (black dots).

with the water retention time mechanism depicted in this study, since high rainfall and freshwater discharge would diminish the retention time. In this situation, only strong southerly winds would block larger freshwater outflow from the Patos Lagoon to the € ller et al., 2001, 2007). coastal region (Mo According to the retention time mechanism, one might expect higher chlorophyll a values due to the higher retention time in years with low rainfall and freshwater outflow, e.g. found during ~ a events (1998e2000), when Southern Brazil experistrong La Nin enced a long drought. However, this was not observed (Abreu et al., 2010; Haraguchi et al., in this issue), indicating that other processes are also important for the overall chlorophyll a level. Abreu et al. (2010) suggested that nutrient exhaustion in the estuary during long retention time could cause a decline of phytoplankton biomass. However, nutrient concentrations near the inlet (this study) and halfway up the estuary (Haraguchi et al., in this issue) suggests that phytoplankton growth is rarely limited by nutrient availability. On the other hand, Odebrecht et al. (2005) showed that phosphate was potentially limiting in the southern part of Patos Lagoon during winter 1988 when the freshwater discharge was relatively low. Inter-annual variations in freshwater discharge, controlling nutrient inputs as well as phytoplankton sedimentation rates in the Patos Lagoon, may consequently change the freshwater end-point for chlorophyll a of the conservative mixing curve for the PLE, such that the mixing baseline in dry years is generally lower than in wet years. These observations indicate that processes acting on the short time scale, like phytoplankton growth and biomass accumulation due to increased water retention time, influence the longer time scales of phytoplankton variability in this ecosystem, and that climate signals acting on a multiannual scale also affect short-term phytoplankton variability. As a final remark, it is important to stress that water retention time is recognized to be a controlling factor in several estuaries and lagoons in Europe (Lemaire et al., 2002; Borsuk et al., 2004; Ferreira et al., 2005), USA (Badylak amd Phlips, 2004; Lane et al., 2011; Phlips et al., 2012; Hall et al., 2013; Putland et al., 2014), Asia (Su et al., 2004; Kasai et al., 2010) and South America (Knoppers et al., 1991; Bonilla et al., 2005; Abreu et al., 2010). However, most of these studies mainly considered the residence time of freshwater in the estuary, since besides stabilization of the water column, freshwater input is essential to bring new nutrients

This study is part of the Brazilian Long Term Ecological Program (PELD) financed by CNPq, Brazilian Ministry of Sciences and FAPERGS, the COCOA project under the BONUS program funded by EU (7th framework program) and the Danish Research Council, and the EXCHANGE project funded by the Danish Research Council. We thank Paulo Mattos and Ricardo Costa for operational help with the sensors maintenance.

References Abreu, P.C., Odebrecht, C., Gonzalez, A., 1994a. Particulate and dissolved phytoplankton production of the Patos Lagoon Estuary, southern Brazil. J. Plankton Res. 16, 737e753. Abreu, P.C., Graneli, E., Odebrecht, C., Kitzmann, D., Proença, L.A., Resgalla, C., 1994b. Effect of fish and mesozooplankton manipulation on the phytoplankton community in the Patos Lagoon Estuary, southern Brazil. Estuaries 17, 575e584. Abreu, P.C., Hartmann, C., Odebrecht, C., 1995. Nutrient-rich saltwater and its influence on the phytoplankton of the Patos Lagoon Estuary, southern Brazil. Estuar. Coast. Shelf Sci. 40, 219e229. Abreu, P.C., Bergesch, M., Proença, L.A., Garcia, C.A.E., Odebrecht, C., 2010. Short- and long-term chlorophyll a variability in the shallow microtidal Patos Lagoon Estuary, Southern Brazil. Estuar. Coasts 33, 554e569. Badylak, S., Phlips, E.J., 2004. Spatial and temporal patterns of phytoplankton composition in a subtropical coastal lagoon, the Indian River Lagoon, Florida, USA. J. Plankton Res. 26, 1229e1247. Bergesch, M., Odebrecht, C., Abreu, P.C., 1995. Microalgas do estu ario da Lagoa dos ~o entre o sedimento e a coluna de  Patos: interaça agua. Nerítica 1, 273e289. Bergesch, M., Garcia, M., Odebrecht, C., 2009. Diversity and morphology of Skeletonema species in southern Brazil, Southwestern Atlantic Ocean. J. Phycol. 45, 1348e1352. rez, M.C., 2005. Influence of hydrology on Bonilla, S., Conde, D., Aubriot, L., Pe phytoplankton species composition and life strategies in a subtropical coastal lagoon periodically connected with the Atlantic Ocean. Estuaries 28, 884e895. Borsuk, M.E., Stow, C.A., Reckhow, K.H., 2004. Confounding effect of flow on estuarine response to nitrogen loading. J. Environ. Eng. 130, 605e614. Chen, Z., Hu, C., Muller-Karger, F.E., Luther, M.E., 2010. Short-term variability of suspended sediment and phytoplankton in Tampa Bay, Florida: observations from a coastal oceanographic tower and ocean color satellites. Estuar. Coast. Shelf Sci. 89, 62e72. Cloern, J.E., 2001. Our evolving conceptual model of the coastal eutrophication problem. Mar. Ecol. Prog. Ser. 210, 223e253. Cloern, J.E., Jassby, A.D., 2010. Patterns and scales of phytoplankton variability in estuarineecoastal ecosystems. Estuar. Coasts 33, 230e241. €ller, O.O., Niencheski, L.F.H., 2002. The Patos Lagoon Fernandes, E.H.L., Dyer, K.R., Mo ~ o event (1998). Cont. Shelf Res. 22, 1699e1713. hydrodynamics during an El Nin € ller Jr., O.O., 2005. Spatial gradients in the flow of Fernandes, E.H.L., Dyer, K.R., Mo southern Patos Lagoon. J. Coast. Res. 21, 759e769. Ferreira, J.G., Wolff, W.J., Simas, T.C., Bricker, S.B., 2005. Does biodiversity of estuarine phytoplankton depend on hydrology? Ecol. Model. 187, 513e523. Fisher, T.R., Peele, E.R., Ammerman, J.W., Harding Jr., L.W., 1992. Nutrient limitation of phytoplankton in Chesapeake Bay. Mar. Ecol. Prog. Ser. 82, 51e63. Fujita, C.C.Y., Odebrecht, C., 2007. Short term variability of phytoplankton composition and biomass in a shallow  area of the Patos Lagoon Estuary (Southern ^ntica Rio Gd. 29, 93e107. Brazil). Atla Gallegos, C.L., Neale, P.J., 2015. Long-term variations in primary production in a eutrophic sub-estuary: contribution of short-term events. Estuar. Coast. Shelf Sci. (in this issue). Hall, N.S., Paerl, H.W., Peierls, B.L., Whipple, A.C., Rossignol, K., 2013. Effects of climatic variability on phytoplankton community structure and bloom development in the eutrophic, microtidal, New River Estuary, North Carolina, USA. Estuar. Coast. Shelf Sci. 117, 70e82. Hall, N.S., Whipple, A.C., Paerl, H.W., 2015. Vertical spatio-temporal patterns of phytoplankton due to migration behaviors in two microtidal estuarine ecosystems: influence on phytoplankton function and structure. Estuar. Coast. Shelf Sci. (in this issue). Haraguchi, L., Carstensen, J., Abreu, P.C., Odebrecht, C., 2015. Long-term changes of the phytoplankton community in a subtropical highly flushed turbid ecosystem, the Patos Lagoon Estuary, Brazil. Estuar. Coast. Shelf Sci. (in this issue).

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004

10

C. Odebrecht et al. / Estuarine, Coastal and Shelf Science xxx (2015) 1e10

Kasai, A., Kurikawa, Y., Ueno, M., Robert, D., Yamashita, Y., 2010. Salt-wedge intrusion of seawater and its implication for phytoplankton dynamics in the Yura Estuary, Japan. Estuar. Coast. Shelf Sci. 86, 408e414. Kjerfve, B., 1986. Comparative oceanography of coastal lagoons. In: Wolfe, D.A. (Ed.), Estuarine Variability. Academic Press, New York, pp. 63e81. Kjerfve, B., 1994. Coastal lagoons. In: Kjerfve, B. (Ed.), Coastal Lagoon Processes. Elsevier Science Publishers B.V., Amsterdam, pp. 1e8. Klein, A.H., 1997. Regional climate. In: Seeliger, U., Odebrecht, C., Castello, J.P. (Eds.), Subtropical Convergence Environments: the Coast and Sea in the Southwestern Atlantic. Springer Verlag, Berlin, pp. 5e7. Knoppers, B., Kjerfve, B., Carmuze, J.P., 1991. Trophic state and water turn-over time in six choked coastal lagoons in Brazil. Biogeochemistry 14, 149e166. Lane, R.R., Madden, C.J., Day Jr., J.W., Solet, D.J., 2011. Hydrologic and nutrient dynamics of a coastal bay and wetland receiving discharge from the Atchafalaya River. Hydrobiologia 658, 55e66. Lemaire, E., Abril, G., De Wit, R., Etcheber, H., 2002. Distribution of phytoplankton pigments in nine European estuaries and implications for an estuarine typology. Biogeochemistry 59, 5e23. Lucas, L.V., Thompson, J.K., Brown, L.R., 2009. Why are diverse relationships observed between phytoplankton biomass and transport time? Oceanogr. Limnol. 54, 381e390. €ller Jr., O.O., Castaing, P., 1999. Hydrographical characteristics of the estuarine Mo area of Patos Lagoon (30
S, Brazil). In: Perillo, G.M.E., Piccolo, M.C., PinoQuivira, M. (Eds.), Estuaries of South America: Their Geomorphology and Dynamics. Springer Verlag, Berlin, pp. 83e100. €ller Jr., O.O., Castaing, P., Salomon, J.C., Lazure, P., 2001. The influence of local and Mo non-local forcing effects on the subtidal circulation of Patos Lagoon. Estuaries 24, 275e289. €ller Jr., O.O., Castaing, P., Fernandes, E.H., Lazure, P., 2007. Tidal frequency dyMo namics of a Southern Brazil coastal lagoon: choking and short period forced oscillations. Estuar. Coasts 30, 311e320. rrez, F.J., Díez-Minguito, M., Losada, M.A., Ruiz, J., 2011. Temporal Navarro, G., Gutie and spatial variability in the Guadalquivir estuary: a challenge for real-time telemetry. Ocean. Dyn. 61, 753e765. €ller, O.O., Niencheski, L.F., Proença, L.A., Torgan, L.C., Odebrecht, C., Abreu, P.C., Mo 2005. Drought effects on pelagic properties in the shallow and turbid Patos Lagoon, Brazil. Estuaries 28, 675e685.

Odebrecht, C., Abreu, P.C., Bemvenuti, C.E., Coppertino, M., Muelbert, J.H., Vieira, J.P., Seeliger, U., 2010. The Patos Lagoon Estuary: biotic responses to natural and anthropogenic impacts in the last decades (1979-2008). In: Kennisch, M., Paerl, H. (Eds.), Coastal Lagoons: Systems of Natural and Anthropogenic Change. Taylor & Francis/CRC Press, Boca Raton, pp. 437e459. Paerl, H.W., Rossignol, K.L., Hall, N.S., Peierls, B.L., Wetz, M.S., 2010. Phytoplankton community indicators of short- and long-term ecological change in the anthropogenically and climatically impacted Neuse River Estuary, North Carolina, USA. Estuar. Coasts 33, 485e497. Peierls, B., Hall, H., Paerl, H., 2012. Non-monotonic responses od phytoplankton biomass accumulation to hydrologic variability: a comparison of two coastal plain North Carolina estuaries. Estuar. Coasts 35, 1376e1392.  Phlips, E.J., Badylak, S., Hart, J., Haunert, D., Lockwood, J., ODonnell, K., Sun, D., Viveros, P., Yilmaz, M., 2012. Climatic influences on autochthonous and allochthonous phytoplankton blooms in a subtropical estuary, St. Lucie Estuary, Florida, USA. Estuar. Coasts 35, 335e352. Putland, J.N., Mortazavi, B., Iverson, R.L., Wise, S.W., 2014. Phytoplankton biomass and composition in a river-dominated estuary during two summers of contrasting river discharge. Estuar. Coasts 37, 664e679. ~ edo-Argüelles, M., Costa Goela, P., Cristina, S., Rieradevall, M., D0 Roselli, L., Can Adamo, R., Newton, A., 2013. Do physiography and hydrology determine the physico-chemical properties and trophic status of coastal lagoons? A comparative approach. Estuar. Coast. Shelf Sci. 117, 29e36. Seeliger, U., Costa, C.S.B., Abreu, P.C., 1997. Energy flow and habitats in the Patos Lagoon Estuary: primary production cycles. In: Seeliger, U., Odebrecht, C., Castello, J.P. (Eds.), Subtropical Convergence Environments: the Coast and Sea in the Southwestern Atlantic. Springer Verlag, Berlin, pp. 65e69. Su, H.M., Lin, H.-J., Hung, J.-J., 2004. Effects of tidal flushing on phytoplankton in a eutrophic tropical lagoon in Taiwan. Estuar. Coast. Shelf Sci. 61, 739e750. Thompson, P.A., O'Brien, T.D., Paerl, H.W., Peierls, B.L., Harrison, P.J., Robb, M., 2015. Precipitation as a driver of phytoplankton ecology: a climatic perspective. Estuar. Coast. Shelf Sci. (in this issue). Welschmeyer, N.A., 1994. Fluorometric analysis of chlorophyll a in the presence of chlorophyll b and phaeopigments. Limnol. Oceanogr. 39, 1985e1992.

Please cite this article in press as: Odebrecht, C., et al., Retention time generates short-term phytoplankton blooms in a shallow microtidal subtropical estuary, Estuarine, Coastal and Shelf Science (2015), http://dx.doi.org/10.1016/j.ecss.2015.03.004