Pyrodinium bahamense var. bahamense cysts as a dinoflagellate population and. depositional environment proxy in Puerto Mosquito, Vieques, Puerto Rico

Transcription

1 Pyrodinium bahamense var. bahamense cysts as a dinoflagellate population and depositional environment proxy in Puerto Mosquito, Vieques, Puerto Rico Kelly Hereid Senior Integrative Exercise March 9, 2007 Submitted in partial fulfillment of the requirements for the Bachelor of Arts degree in Geology, Carleton College, Northfield, Minnesota.

2 TABLE OF CONTENTS Abstract Introduction...1 Study aims 1 Pyrodinium bahamense var. bahamense and relatives 1 Population controls 4 Dinoflagellate cysts and blooms 4 Mixing 5 Competition 9 Analyzing carbon sources 10 Methods...11 Field work 11 Sample selection 14 Laboratory work 14 Limitations on methods 16 Results...18 Sediment analysis 18 Cysts 24 Carbon isotope data 24 Discussion...28 Sediment analysis 28 Cysts 32 Carbon isotope data 33

3 Directions for further research 39 Acknowledgements...40 References...41

4 Pyrodinium bahamense var. bahamense cysts as a dinoflagellate population and deposition proxy in Puerto Mosquito, Vieques, Puerto Rico Kelly Hereid Carleton College Senior Integrative Exercise March 9, 2007 Advisor: Clinton Cowan, Carleton College ABSTRACT Puerto Mosquito is a bioluminescent bay in Vieques, Puerto Rico occupied by the dinoflagellate Pyrodinium bahamense var. bahamense. We took a sediment core from this bay to analyze dinoflagellate cyst populations. The cysts of this dinoflagellate are an accurate proxy for population and depositional information, as determined by assessing relative contributions of carbon sources for the study bay using a δ 13 C analysis. The cyst counts show substantial variability that correlates positively with variations in the µm sediment grain size fraction, indicating that changes in cyst distribution are based on hydrological changes that alter the deposition of sediments rather than population changes. Therefore, there are no substantial changes in the dinoflagellate population over the ~34 year time period recorded in the core, so the population is at its maximum capacity for the bay, and that capacity has not changed significantly in recent times. Keywords: dinoflagellates, cysts, carbon-13, sedimentary environment, mangrove swamps, ecology

5 1 INTRODUCTION Study Aims This study is intended to analyze the validity of using cysts as population proxies for a blooming population of the dinoflagellate Pyrodinium bahamense var. bahamense in Puerto Mosquito, a bioluminescent bay located in Vieques, Puerto Rico. Cyst populations preserved in sediment from the bay are compared with modern dinoflagellate populations and sediment δ 13 C data to see how accurately cysts record changes in dinoflagellate populations through time. Another goal of this project is to develop a method to study dinoflagellate cysts in a way that uses fewer of the dangerous chemicals required by typical palynological technique. This involved substituting a combination of HCl and sieving to minimize the carbonate and silicate in the samples to remove sample preparations involving HF. Pyrodinium bahamense var. bahamense and Relatives Pyrodinium bahamense is a photoautotrophic cyst-forming dinoflagellate found in shallow tropical marine and estuarine systems throughout the world where the water is warmer than 25 C (Fig. 1; Badylak and Phlips, 2004). P. bahamense exists in two varieties: the highly toxic P. bahamense var. compressum in the tropical Pacific and the fairly benign bioluminescent P. bahamense var. bahamense in the tropical Atlantic (Steidinger et al., 1980; Balech, 1985; Badylak et al., 2004). P. bahamense var. compressum produces a potent neurotoxin known as satitoxin, and is responsible for the greatest number of paralytic shellfish poisoning cases in the world (Azanza and Taylor, 2001).

6 2 B A Figure 1. Pyrodinium bahamense var. bahamense (A) vegetative stage and (B) cyst (Steidinger et al., 1980; Matsuoka and Fukuyo, 2000).

7 3 However, there are some difficulties inherent in the study of P. bahamense var. compressum. It typically occurs only in relatively transient blooms that are not necessarily in the same location from year to year, as opposed to the continuous bloom of P. bahamense var. bahamense found in the study location. This limits available data sets for the species, hinders any attempts to understand long-term population trends. Also, the parts of the world that experience P. bahamense var. compressum blooms are typically highly disturbed by human activity, particularly waste-related eutrophication and disturbance of sediments caused by trawling (Usup and Yu, 1991; Sombrito et al., 2004). Therefore, studies of P. bahamense var. compressum have had to try to compensate for these human environmental disturbances, which significantly complicate bloom analysis. The study location, Puerto Mosquito in Vieques, Puerto Rico, alleviates many of the problems associated with study in the locations where P. bahamense var. compressum blooms. The study bay is rather unusual in that it contains a standing crop of phytoplankton rather than a resuspension-driven bloom; the lack of resuspension improves the accuracy of cysts as a population proxy. Also, the bloom of P. bahamense var. bahamense in this study occurs in a bay that is nearly undisturbed by human activity. Removing the variables associated with human disturbance significantly simplifies defining a link between an environmental factor and its effect on a bloom. Moreover, the study of P. bahamense var. bahamense took on new urgency with reports that it can produce the same neurotoxin as P. bahamense var. compressum, which is cause for further study on this poorly understood organism (Landsberg, 2002; Landsberg et al., 2006).

8 4 The two varieties of P. bahamense are closely related to Alexandrium spp., if they are even a separate genus (Usup et al., 2002; Leaw et al., 2005). The extensive studies conducted on the geographically widespread Alexandrium spp. can offer some useful analogues for information that is not known about Pyrodinium. Furthermore, a continuous Pyrodinium bloom offers opportunities for study that may inform the understanding of more intermittent Alexandrium blooms, which cause substantial harmful algal blooms in both tropical and temperate regions of the world. Population Controls Dinoflagellate Cysts and Blooms P. bahamense is one of many dinoflagellate species that forms a resistant and protective casing, called a cyst. Dinoflagellates usually multiply by binary fission, but sometimes haploid cells in the water column combine to make a diploid zygote, which then undergoes encystment, or cyst formation (Olney, 2002). While some dinoflagellate species rely on cysts for reproduction only under environmental stress, P. bahamense cysts are a necessary dormant stage in the life cycle of the species (Anderson and Wall, 1978). Particularly dense patches of cysts settle out of the water column and into the sediment in what are known as cyst beds. The germination of cysts from these concentrated beds is what drives bloom initiation (Nehring, 1996; Sombrito et al., 2004). However, not all of the cysts will germinate. Once buried by sediment, a cyst can remain viable for years, but it needs some mechanism to leave the sediment column or it will never germinate (Nehring, 1996; Anderson et al., 2005). In the absence of mixing, only a slow trickle of cysts in the surface sediments will germinate, and many will be

9 5 permanently buried, but a sediment disturbance can be a powerful bloom-initiating event (Nehring, 1996; Kirn et al., 2005). Mixing Vertical mixing of the water column is a somewhat complicated control on the growth of cyst-forming dinoflagellates (Fig. 2). After cysts settle out of the water column and down into the sediment, they generally need to be resuspended to germinate. Cysts typically germinate at very low rates from surface sediments, and not at all once they are buried. However, in the water column, cysts germinate at ten times the rate of those at the top of the sediment column (Nehring, 1996). As such, an event that disturbs sediments and resuspends cysts can be a powerful trigger for a dinoflagellate bloom (Kirn et al., 2005). Sediments can be disturbed in a variety of ways, but water that has been highly mixed is particularly tied to triggering bloom conditions for cyst-forming species (Usup and Yu, 1991; Azanza et al., 2004). The problem with blooms triggered by resuspension is that germinated cysts are not well-preserved, so each bloom reduces the validity of using cysts as a population proxy in the area. This species of dinoflagellate is a type of phytoplankton, so it needs to stay in the photic zone to photosynthesize and reach bloom populations, which is best achieved with low mixing rates (Monbet, 1992; Koseff et al., 1993; Huisman et al., 1999b, a; Ebert et al., 2001; Lim et al., 2006). The way to deal with these contradictory mixing requirements in cyst-forming dinoflagellates is to incorporate temporal variation in mixing rates into the model. If a period of high mixing is followed by a period of low mixing, then the resuspended and germinated cysts would be able to meet their light requirements and initiate a bloom (Farida et al., 1996; Villanoy et al., 1996). A practical

10 6 A C B Figure 2 D

11 7 E Figure 2 (continued). Potential water column mixing scenarios. (A) High mixing, low depth - light penetrates to bottom, bloom. (B) High depth, low mixing - phytoplankton growth rate exceeds rate of transport to bottom, bloom. (C) High depth, high mixing, vertical stratification - stratification limits depth, light penetrates past stratified layer, bloom. (D) High depth, high mixing - phytoplankton transport rate away from light faster than growth rate, light limitation, no bloom. (E) High mixing disturbs sediment - resuspends cysts, increases turbidity, light limitation, no bloom. (F) High mixing disturbs sediment, delayed vertical stratification - monsoon winds mix water column, disturbing sediments and suspending cysts in water column, no bloom. (G) Monsoon rain input is delayed by several days traveling down rivers. Fresh water causes haline stratification, which keeps phytoplankton germinated from cysts in photic zone, bloom. F G

12 8 example of this situation is a tropical monsoon, in which a period of high wind causes vertical mixing, and the delayed pulse in fresh water (rainfall) delivered by rivers sets up haline stratification that decreases the depth of vertical mixing shortly afterwards (Byun et al., 2005; Moore et al., 2006). Indeed, many blooms are associated with the high winds or rains typical of a monsoon both in and around the tropics (Usup and Yu, 1991; Farida et al., 1996; Azanza et al., 1998; Babaran et al., 1998; Azanza and Miranda, 2001; Yin, 2003; Azanza et al., 2004). The mixing status of Puerto Mosquito therefore makes it a prime candidate for phytoplankton growth. The bay has a typically fairly stable water mass, as indicated by its long water residence time (Keck, 2007). As a general rule, wind and tides increase mixing, and temperature and salinity gradients create vertical stratification that decreases mixing (Monbet, 1992; Koseff et al., 1993; Yin, 2003; Moore et al., 2006). Puerto Mosquito has extremely low tidal variation, so tidal mixing of the water column is not a concern. Wind is also somewhat controlled by the surrounding mangrove swamps, so even though there are no sharp gradients in temperature or salinity to generate vertical stratification, mixing in this bay is quite low, with only the occasional powerful wind from storm events stirring the water column enough to disturb the sediments (Keck, 2007). The blooming population is now so large that mixing events are not necessary. The sediment cyst supply could probably drive the bloom on its own, making high vertical mixing events that typically are the triggers for P. bahamense blooms less necessary. This explains why Puerto Mosquito has the unique feature of a standing crop of P. bahamense rather than a resuspension-initiated bloom.

13 9 Competition While the mixing regime of Puerto Mosquito makes it well-suited to support a phytoplankton bloom, there must be some mechanism by which P. bahamense became the dominant phytoplankton species in the bay. This is the role of competition and resource availability. The competitive conditions have to be correct to initiate a phytoplankton bloom. Plankton ecosystems are rather unique in that low species diversity leads to high biomass and vice versa (Irigoien et al., 2004; Aktan et al., 2005; Buyukates and Roelke, 2005; Duarte et al., 2006; Passy and Legendre, 2006). Usually, more productive communities are composed of a greater number of species utilizing a broader range of resources and operating in mutually beneficial relationships (Bruno et al., 2005; Downing, 2005; Cardinale et al., 2006). The wide variety of ecological niches in most ecosystems contributes substantially to species diversity. However, in a eutrophic system such as Puerto Mosquito, the amount of available nutrients sets the maximum possible phytoplankton population, and the species that can best obtain light and shade out any other species will become dominant (Huisman et al., 1999b; Irigoien et al., 2004). As such, it is unsurprising that one species of phytoplankton came to dominate the study location, which is what led to its extremely high biomass. The quantity and balance of nutrient resources helps to determine the composition of a bloom. Eutrophication often initiates a phytoplankton bloom, due to the control that nutrient limitation places on a phytoplankton populations (Cloern et al., 1995; Smith, 2006). This condition is typically met in an environment where the water has a long residence time. Such a system hydrologically concentrates nutrients and the

14 10 phytoplankton populations, instead of allowing these to be dispersed by currents or tides (Knoppers et al., 1991; Phlips et al., 2002; Badylak and Phlips, 2004). Then, the N:P ratio needs to be high to limit heterotrophs, and make the plankton bloom consist of phytoplankton (Saetre et al., 1997; Azanza et al., 2004; Caroppo et al., 2006; Smith, 2006). Additionally, Si needs to be limiting to prevent diatoms from being dominant, instead allowing the bloom to be primarily composed of dinoflagellates (Moncheva et al., 2001; Yin, 2003). Nearly all of the studies of optimal growth are based on current or fairly recent blooms. This is a fairly limited data set that is complicated by unknown variations in environments, as the study blooms are typically variable in size and transient in both timing and location. Therefore, the bloom in Vieques offers a unique opportunity. By holding the location constant, it minimizes the impact of unknown factors, thus isolating environmental factors for study. Furthermore, by including the past population record, this substantially increases the size of the data set, offering a clearer picture of population controls. Analyzing Carbon Sources Carbon stable isotope ratios are used to analyze the sources of carbon inputs for an ecosystem (Fry and Sherr, 1984; Peterson and Fry, 1987; de Boer, 2000; Dehairs et al., 2000; Yamamuro, 2000; Dittmar et al., 2001; Bouillon et al., 2003; Jones et al., 2003; Bouillon et al., 2004; McCallister et al., 2004; Kaldy et al., 2005; Goni et al., 2006; Hu et al., 2006; Zong et al., 2006). The source serves as a baseline ratio of carbon isotopes, which can then be altered by fractionation and mixing of multiple sources (Peterson and Fry, 1987). The main type of fractionation expected in the study bay after the inputs

15 11 from different primary producers is a small enrichment of 13 C in the sediment related to decomposition (Bouillon et al., 2004). Therefore, carbon isotopes can be a useful measure of the relative contributions from various carbon sources in the study site. The three main carbon contributors to Puerto Mosquito are mangroves, macrophytes, and phytoplankton (Fig. 3). Changes in the proportion of the carbon contribution each makes to the sediment reflect depositional changes, caused either by population variation or changing hydrologic conditions. METHODS Field Work We collected cores from three bays in Vieques, Puerto Rico (Fig. 4). The main study bay is Puerto Mosquito (highly bioluminescent). The bay s very narrow inlet contributes to its long water residence time. This location is also marked by low tidal variation, a primarily silty bottom, and it is surrounded by mangrove swamps. The control bay is Puerto Ferro (no observable bioluminescence). This location has a much wider inlet and associated shorter water residence times, along with a sandy bottom and a surrounding mangrove swamp (Keck, 2007). The comparative bay is Bahía Tapón, which is slightly bioluminescent today, but was much more so in the past (O'Connell, 2006). Bahía Tapón is geographically and geomorphologically similar to Puerto Mosquito, so we hope to learn what distinguishes the bays from one another. The cores were taken with short-length PVC pipes in shallow water and using a gravity corer in deep water in multiple locations within each bay, both of which are preferable for accurate cyst counting at depth in the sediment column (Anderson et al.,

16 12 A B C Figure 3. Major carbon sources in Puerto Mosquito. (A) Mangroves (photo by Suzanne O Connell). (B) Macrophytes (photo by Andrew Nelson). (C) Pyrodinium bahamense var. bahamense, dark circles, scale bar is 50µm.

17 Figure 4. Map of study area showing core locations (Erin Tainer). Puerto Mosquito Puerto Ferro Bahía Tapón 13

18 ). These cores were extruded and sliced open for sedimentary facies characterization, and dinoflagellate cyst samples were taken and preserved in glutaraldehyde. Each sample was taken to represent an average over either a whole facies interval or a 1cm increment. This study comprises one core sampled at 1cm increments (Fig. 5). The study core is located in Puerto Mosquito in water ~1-1.5m deep. This area is silty, and covered by a bed of seagrass. The core itself was mostly silt and clay near the top with coarse sand sized shell fragments, with larger shell pieces and whole shells deeper in the core. The samples are composed of carbonate mud and shell fragments, a small amount of quartz sand, biogenic silica (mostly diatoms), and organic carbon. The carbonate fraction becomes more significant deeper in the cores, as most are marked by a shell-dominated storm facies at depth. Sample Selection The study core was chosen based on a variety of factors. Since it was sampled at 1cm intervals, the core is more useful for correlating cyst densities to a particular period of time (rather than a particular facies). This core also has a substantial amount of sediment in the size fraction of silt and clay, which is the size fraction that contains cysts. This section lacks 210 Pb data, so sedimentation rate estimates come from a core from a geographically similar part of the bay (which is subdivided by facies rather than 1cm increments). Laboratory Work Each sample was mixed with water and allowed to settle out for several days to obtain a consistent amount of pore water between samples. Excess liquid was decanted

19 Figure 5. Section of core sampled for cysts. Sampled at 1cm increments until beginning of facies change to storm facies at 14cm (photo by Suzanne O Connell). 15

20 16 over a 20µm sieve to catch floating material, the sediment was thoroughly mixed to get an even distribution of grain sizes throughout, and a ~1g portion of each sample was removed for analysis. An additional portion was dried and weighed to get the average water content for each sample. Wet samples were weighed, and each sample was washed through a 105µm sieve and a 20µm sieve. The material on both sieves was weighed, with the remainder of the weight assumed to be the clay fraction that passed through the 20µm sieve. The material on the 20µm sieve was used to prepare the slides. The size fraction between 105µm and 20µm was dissolved overnight in muriatic acid (31.45% HCl), and weighed again. The remaining sample was then spread among as many gridded slides as was necessary to disperse the sediment for accurate counts, and the cysts were counted on 4 transects on each slide with a light microscope (at 100x and 400x) and multiplied to get a total for each sample (methods modified from (Matsuoka and Fukuyo, 2000; Sangiorgi et al., 2006). The age and sedimentation rate of the samples were determined by 210 Pb dating (Keck, 2007). Limitations on Methods Cyst counts are normalized for sample size and composition by controlling for the wet weight of the sediment sample in the grain size that contains the cysts. However, varying water content in the sediment may introduce error in the sediment weight. This error has been limited in a variety of ways. The samples settled out of an excess quantity of liquid (the glutaraldehyde preservative and water) over a period of several days, so they have a stable amount of liquid in their pore spaces. Also, the size fraction of the

21 17 sediment samples is limited to µm, which limits the range of variation in pore spaces. The primary issue with wet sediment weights is that water content can vary substantially by depth within a core due to compaction. This error is minimized by the fact that each sample was thoroughly mixed when the water was added. Since the samples are small (~5-10 ml volume), the weight of the overlying sediments within each sample is minimal, so the effect of this weight on the amount of water in pore spaces is correspondingly reduced. Additionally, the average water content of each sample (obtained from measuring the wet and dry weights of a representative portion of each sample) allows for the calculation of the dry weight of the sample, which gives an absolute ratio between the number cysts and a given weight of sediment. However, some potential sources of error in normalizing cyst counts by sediment weight still exist. In particular, the sediment samples get transferred back and forth between sieves, Petri dishes, and slides a number of times in the process of preparing them to be counted, and some sediment will inevitably be lost at each of these transfers. Since any weight that is unaccounted for is assumed to be part of the <20µm size fraction, the weight of that size fraction is systematically overestimated. Nevertheless, the error is a small one, and should be on a similar order of magnitude for each sample, so the effects of this error on normalizing cyst counts by sediment weights should be minimal. Furthermore, the 20µm sieve drains water less well than the 105µm sieve, so it is more difficult to remove all excess water and leave only that water located in the sediment pore spaces to obtain a consistent amount of water for accurate wet weights. Therefore,

22 18 particular care was taken to clear extra water from the sample before it was weighed after sieving. The state of the samples when they were counted also affects the overall counts. The samples were kept in the preservative for several months, and the acid used to dissolve the carbonate is stronger than that used in most other methods. Consequently, the spines that are the cysts most distinguishable feature were lost through dissolution. This made distinguishing the cysts far more difficult, and may alter the final count. Therefore, stable carbon isotope data is included as a check on the validity of the cyst count. RESULTS Sediment analysis The sample grain size distribution is summarized in Figure 6. There is a decrease in the µm size fraction and increase in the >105µm size fraction with depth, as the samples approach the depth of the facies change seen in Figure 5. This change is matched by a decrease in pore water content with depth (Fig. 7). The amount of carbonate in the samples increases further down in the core, corresponding to a decrease in the organic carbon and silica fraction (Figs. 8, 9). Removing the HF preparation means there is no way to remove silica from the samples, so no organic carbon fraction is available. The ages of two cores from the study bay at various depths are shown in Figure 10. These ages are calculated from 210 Pb data obtained from core samples (Keck, 2007). These show a slight increase in sedimentation rate over time, but this is likely due to

23 <20µm (calculated) µm (pre-acid) >105µm % 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Depth in sediment column (cm) Figure 6. Study core sediment grain size distribution.

24 Top water 0-1cm 1-2cm 2-3cm 3-4cm 4-5cm 5-6cm 6-7cm 7-8cm 8-9cm 9-10cm 10-11cm 11-12cm 12-14cm Depth in sediment column (cm) Figure 7. Percent water weight per sample.

25 Top water 0-1cm 1-2cm 2-3cm 3-4cm 4-5cm 5-6cm 6-7cm 7-8cm 8-9cm 9-10cm 10-11cm 11-12cm 12-14cm Depth in sediment column (cm) Figure 8. Percentage of each sediment sample in the µm grain size composed of carbonate.

26 Top water 0-1cm 1-2cm 2-3cm 3-4cm 4-5cm 5-6cm 6-7cm 7-8cm 8-9cm 9-10cm 10-11cm 11-12cm 12-14cm Depth in sediment column (cm) Figure 9. Combined percentage of sample composed of organic C and silica.

27 Figure 10. Age of two cores with depth from Pb-210 data (Keck, 2007). PM4 is shallow core used to get sedimentation rate of 0.407cm/yr. PM4 PMD Age (yrs) Depth in sediment column (cm)

28 24 compaction with depth. PM4 is the core most similar in location to the study core, as PMD1 is from the middle of the bay, so the sedimentation rate for PM4 of 0.407cm/yr is used for calculations involving the study core. Cysts The cyst counts for 1g of wet sediment are shown in Figure 11. The calculation of the number of cysts in 1g of dry sediment (Fig. 12) is based on data from the wet sediment counts and the average percent water composition of each sample (Fig. 7). The average number of cysts per gram of wet sediment is 4697 ± 1318 (1 standard deviation), and the average number of cysts per gram of dry sediment is ± 8775 (1 standard deviation). The broad ranges reflect the variability in the cyst counts, most likely due to sectioning samples for counts and multiplying to get totals. The overall cyst trend shows a gradually decreasing number of cysts with depth. Carbon isotope data The δ 13 C data for cores from Puerto Mosquito are located in Figure 13. The samples are typically somewhat depleted in 13 C with depth in the core, with the exception of one core composed of mangrove peat that is highly depleted in the top 1cm of sediment, with a δ 13 C value of However, the isotopic ratios stay fairly constant for the top 50cm of the sediment column, which based on the sedimentation rate corresponds to the past 123 years. Therefore, for the timescale of this study the depletion at depth can be ignored.

29 Depth in sediment column (cm) Figure 11. Cyst counts per gram of wet sediment with depth in study core.

30 Depth in sediment column (cm) Figure 12. Cyst counts per gram of dry sediment with depth in study core.

31 Figure 13. Stable carbon isotope ratios with depth in cores throughout Puerto Mosquito. Red line shows sediment average isotopic composition of phytoplankton Depth in sediment column (cm)

32 28 DISCUSSION Sediment analysis The changes in sediment sample characteristics can be attributed to the transition to a shell-filled facies past 14cm. The facies change explains the increase in carbonate, the larger grain size, and the decrease in pore water space, since the shells are made of carbonate, are larger than the fragments found in the sample section, and do not have significant amounts of pore space. Therefore, these changes in sediment characteristics are not caused by changes in cyst population. The sediment size distribution can be used as a check to see what size fraction dinoflagellate cysts sort with hydrologically. The correlations between cyst counts and the three size fractions used in this study can be found in Figures 14, 15, and 16. There is no correlation between cysts and the <20µm size fraction, a strong positive correlation between cysts and the µm size fraction, and a negative correlation between cysts and the >105µm size fraction. Cysts are usually ~45µm in diameter, so even though they are less dense than inorganic sediment they sort with the silt size fraction rather than the lighter <20µm fraction (Williams and Bujak, 1985). There is no correlation with the smallest grain size fraction (rather than a negative correlation) because clay-sized particles can be deposited in the same areas as silt due to flocculation, meaning the rates of deposition of the two smaller size fraction are unrelated to each other. The >105µm size fraction and the µm size fraction are negatively correlated since their depositional environments are in opposition to one another. The more sediment that is deposited in the large size fraction, the more likely it is that smaller sediment particles will remain suspended in the water column.

33 Number of cysts Figure 14. Number of cysts shows no correlation with <20µm grain size fraction.

34 Number of cysts Figure 15. Number of cysts shows postive correlation with µm grain size.

35 Number of cysts Figure 16. Negative correlation between number of cysts and the >105µm grain size fraction.

36 32 The negative correlation with the large size fraction demonstrates where cyst beds are likely to form. Even though cysts may be formed across a variety of depositional regimes, they are hydrodynamically unstable where larger grain sizes are deposited, so they will stay suspended in the water column. Therefore, even though a bloom can spread to a shallow sandy area given enough light, new blooms are unlikely to initiate from that location, since the cysts would have been carried elsewhere. Cysts The study bay has very low mixing, with a tidal range on the order of 40cm and minimal currents (Keck, 2007). Therefore, many cysts get buried in the sediment, improving the accuracy of population estimates. This accuracy improvement comes because empty or germinated cysts of Alexandrium, one of the closest relatives to Pyrodinium, do not hold their shape, so are less likely to be preserved than living or ungerminated cysts (Mizushima and Matsuoka, 2004; Leaw et al., 2005). This preservation bias should be taken into account in cyst counts, but the nature of the study site minimizes the magnitude of the bias. Cysts provide the primary source of information about past populations of dinoflagellates, and P. bahamense is a particularly good species to use for this type of analysis. Some dinoflagellates (such as species in Protoceratium or Scrippsiella) form cysts only when they are stressed by environmental conditions, so the correlation between dinoflagellates in the water column and cysts being deposited is low (Godhe et al., 2001; Peperzak, 2006). However, since P. bahamense form resting cysts as a part of their life cycle, there is a continuous stream of cysts deposited into the sediment, so cyst

37 33 densities can be used as accurate and continuous proxies for population size (Anderson and Wall, 1978). Different sediments are affected by decay processes to a different extent, often based on oxygen content. If the cysts are prone to aerobic degradation, the population signal in the sediment record can quickly be obscured by varying decay rates (Versteegh and Zonneveld, 2002). However, the cysts of this species are known to preserve particularly well. P. bahamense cysts are typically known to paleontologists as Polysphaeridium zoharyi, Hemicystodinium zoharyi, Hystrichosphaeridium zoharyi, or Polysphaeridium subtile (Wall and Dale, 1969; Williams and Bujak, 1985). These cysts are found continuously in the fossil record from the lower Eocene (~50 Ma) to the present (Williams and Bujak, 1985). They are also known for being particularly resistant to oxidation, despite having an organic wall (Zonneveld et al., 2001). This excellent preservation potential means that population counts will not be significantly affected by the age of the cysts or small-scale sediment conditions given the age of sediment in the study cores, so compensation for decay with depth is unnecessary. The cyst preservation potential is confirmed by the data from this study. There is a gradual decrease in the number of cysts in the sediment with depth (Fig. 12), but there is no exponential drop that would be expected if the cysts were heavily affected by decay. Carbon isotope data The primary potential sources of this organic matter include the mangroves surrounding the bay, macrophytes growing on the bottom of the bay (mostly seagrasses), and phytoplankton (Fig. 4). Since this bay is minimally flushed, autochthonous sources of carbon such as phytoplankton are much more important than any allochthonous input

38 34 (Matson and Brinson, 1990; Dehairs et al., 2000). The mangroves from Puerto Mosquito have an average isotopic composition of ± 0.6, which falls within the range described in the literature (Bouillon et al., 2003; Corbisier et al., 2006; Muzuka and Shunula, 2006; Keck, 2007). Sediments whose organic carbon is almost entirely derived from mangroves have an average δ 13 C value from around -26 to -28, which is consistent with the observation that sediments and soils are typically slightly enriched in 13 C by ~2 relative to their main input (Ehleringer et al., 2000; Bouillon et al., 2003; Bouillon et al., 2004). However, the higher the organic carbon content of the sediments, the closer the sediment isotopic composition will be to that of mangroves, which explains the very depleted sample from the core composed of mangrove peat (Lallier-Verges et al., 1998; Bouillon et al., 2003). Seagrass, which is the dominant macrophyte in this system, has a δ 13 C signature that ranges from to -18.6, and the overall macrophyte average is -12 in the sediments (Fry and Sherr, 1984; Bouillon et al., 2004). Marine phytoplankton has an average δ 13 C value of -21, giving an average sediment δ 13 C signature of (Fry and Sherr, 1984; Gu et al., 2006). The δ 13 C values of the samples from Puerto Mosquito decreases gradually with depth in the core, but stays fairly constant for the top 50cm of sediment at ± 1.8 (Fig. 13). Since the incubated core samples were enriched in 13 C by 0.8 to 1.0, there is probably a microbial component in the sediment that increases the δ 13 C values over time (Keck, 2007). However, this change over time will be ignored as it deals with sediments that are outside the scope of this study. The average isotopic composition of sediments in Puerto Mosquito indicates mixed carbon inputs. While this could be caused just by the mixing of seagrass- and

39 35 mangrove-derived organic matter, it makes more sense to include a contribution from the biomass of the large dinoflagellate bloom, particularly because seagrass decomposes more quickly than mangrove leaves, minimizing its isotopic contribution to the sediment (de Boer, 2000; Gonneea et al., 2004). There is no way to exactly quantify relative inputs from each source in a three-input system without another parameter to constrain source type, so Figure 17 describes possible mixing scenarios that would create the isotopic profile seen in the study bay. Assuming inputs from mangroves and macrophytes stay the same over time, any variation in δ 13 C should come from changes in the dinoflagellate population. This variation will be on a small scale since the amount of biomass contributed to the sediment by one dinoflagellate is substantially less than the contribution of one seagrass or mangrove leaf. This makes it particularly useful that the study bay has not been affected by human activities, as anthropogenic inputs can cause variations in the δ 13 C signature of the sediments (Owen and Lee, 2004; Hu et al., 2006). Nevertheless, phytoplankton can contribute a substantial portion of the organic carbon in the sediment despite relatively low primary production (Fry and Sherr, 1984; Gonneea et al., 2004). Carbon isotope data thus serve as a check to verify the cyst population data. The cyst data show good correlation with stable carbon isotope values from other cores in the study bay. The δ 13 C values for the top 20cm of other cores in Puerto Mosquito (with incubated cores and a mangrove peat core removed) are shown in Figure 18. A δ 13 C value of shows a phytoplankton-dominated carbon source (Gu et al., 2006). This data is confirmed for the study bay by the δ 13 C value of for the top 1cm of sediment from a core taken in the middle of the bay (Keck, 2007). Since terrestrial mangrove organic carbon does not extend beyond the fresh and brackish parts

40 δ13c ( ) 36 % of C source from dinoflagellates Dinoflagellates Seagrass Mangrove A 100% 0% B Figure 17. Three source mixing model for stable carbon isotopes. (A) Shows contribution made by dinoflagellates, seagrass, and mangrove to total δ13c. Red box indicates average isotopic composition of Puerto Mosquito sediments ± 1 standard deviation. (B) Bottom axis corresponds to A. Shows percent contribution of dinoflagellates (black), seagrass (red) and mangrove (blue) to sediment organic C.

41 Figure 18. Stable carbon isotope ratios in top 20cm of cores throughout Puerto Mosquito. Red line shows sediment average isotopic composition of phytoplankton Depth in sediment column (cm)

42 38 of the bay, and this core location is too deep for light to penetrate for seagrass beds, this value should reflect the phytoplankton carbon input for the sediments (Fry and Sherr, 1984). Therefore, deviations from this value indicate a decreased influence of phytoplankton as a carbon source, which should correspond to a drop in the cyst counts. There are three deviations in the δ 13 C value from the typical sediment value of around over the depth interval with cyst counts: at 2cm (~ -2 ), at 5cm (~ -3 ), and at 15cm ( ~ +3 ). These variations indicate that fewer cysts and phytoplankton were being deposited in parts of the bay, since a carbon source with a more positive (seagrass) or negative (mangrove) δ 13 C signature than phytoplankton contributed relatively more to the sediment organic carbon at the time the sediment was deposited. The isotope deviations correlate with decreases in the cyst counts, which drop at 3cm, 5cm, and 14cm. Therefore, the carbon isotope data confirms the accuracy of the cyst counts in the sediment, and variations in the rate of deposition of phytoplankton. A more detailed cyst count could be used as a proxy for the phytoplankton population in the bay. However, this study illustrates some of the complications to this approach. This bay is likely at its carrying capacity for phytoplankton, and since it is relatively undisturbed this carrying capacity has probably not changed substantially in the time frame offered by the study core (~34 years based on 210 Pb-derived sedimentation rate). Therefore, the variations in cyst numbers in the sediment likely reflect hydrological variation, which is corroborated by the close correlation between cyst counts and grain size fractions of the sediment samples. Changing water flow conditions altered the distribution of sediment in the particular grain size that carries cysts, which changed the relative cyst density and thus the δ 13 C signature of that particular sediment layer. In

43 39 Puerto Mosquito, changes in cyst composition with depth in the sediment likely reflect spatial population and depositional variations rather than drastic changes in the size of the bloom. This demonstrates the need for careful grain size control in any cyst-based dinoflagellate population studies. Directions for further research The δ 13 C isotopic signature is obscured somewhat since most of the isotope samples represent averages over 5cm sediment depth intervals, which is why the depths with deviations also have cores that show normal isotopic values. Therefore, finer-scale isotope samples could provide a more detailed check on cyst populations, which could then allow sediment carbon isotope data to predict declines in cyst densities at depth. This research is most useful if conditions in the sediment can be correlated with what is actually happening in the modern environment. This study collected data in the field on seawater salinity, temperature, and chemistry, as well as pore water chemistry and current bloom population densities. Once this data has been processed, it can be correlated with environmental data from the study core, to put absolute numbers on past environmental parameters that could be influencing population dynamics. Of particular interest is the relationship between population and salinity changes. Given a reasonable proxy for past salinities (which can be correlated to modern seawater data), it should be possible to see changing cyst densities associated with a rise or drop in inferred salinity. The close relatives of Pyrodinium, Alexandrium, have been shown to alter their growth rate and toxin production with variations in salinity (Jensen and Moestrup, 1997; Hamasaki et al., 2001; Grzebyk et al., 2003; Lim and Ogata, 2005). In light of evidence that P. bahamense var. bahamense may in fact produce the same neurotoxin as its Pacific

44 40 cousin, its reaction to salinity could be particularly useful information (Landsberg, 2002; Landsberg et al., 2006). ACKNOWLEDGEMENTS Thank you to the Keck Geology Consortium for organizing and funding this project. Also invaluable was the advice I received on methods from Phil Camill and on style and substance from Clint Cowan. Lastly, my appreciation goes to all the kids in Mudd who dealt with my piles of junk and periodic rants this term, particularly the amazing and ridiculous class of 2007.

45 41 REFERENCES Aktan, Y., Tufekci, V., Tufekci, H., and Aykulu, G., 2005, Distribution patterns, biomass estimates and diversity of phytoplankton in Izmit Bay (Turkey): Estuarine Coastal and Shelf Science, v. 64, no. 2-3, p Anderson, D. M., Aubrey, D. G., Tyler, M. A., and Coats, D. W., 1982, Vertical and horizontal distributions of dinoflagellate cysts in sediments Gonyaulax spp: Limnology and oceanography., vol., v. 27, no. 4, p Anderson, D. M., Stock, C. A., Keafer, B. A., Nelson, A. B., Thompson, B., McGillicuddy, D. J., Keller, M., Matrai, P. A., and Martin, J., 2005, Alexandrium fundyense cyst dynamics in the Gulf of Maine: Deep-Sea Research Part Ii-Topical Studies in Oceanography, v. 52, no , p Anderson, D. M., and Wall, D., 1978, Potential importance of benthic cysts of Gonyaulax tamarensis and Gonyaulax excavata in initiating toxic dinoflagellate blooms: Journal of phycology, v. 2, p Azanza, R. V., and Miranda, L. N., 2001, Phytoplankton composition and Pyrodinium bahamense toxic blooms in Manila Bay, Philippines: Journal of Shellfish Research, v. 20, no. 3, p Azanza, R. V., Roman, R. O., and Miranda, L. N., 1998, Shellfish toxicity and Pyrodinium cell density in Bataan, Philippines ( ): Journal of Shellfish Research, v. 17, no. 5, p Azanza, R. V., Siringan, F. P., San Diego-Mcglone, M. L., Yniguez, A. T., Macalalad, N. H., Zamora, P. B., Agustin, M. B., and Matsuoka, K., 2004, Horizontal dinoflagellate cyst distribution, sediment characteristics and benthic flux in Manila Bay, Philippines: Phycological Research, v. 52, no. 4, p Azanza, R. V., and Taylor, F., 2001, Are Pyrodinium Blooms in the Southeast Asian Region Recurring and Spreading? A View at the End of the Millennium: AMBIO, v. 30, no. 6, p Babaran, R. P., Espinosa, R. A., and Abalos, T. U., 1998, Initiating and triggering mechanisms causing harmful algal blooms: Journal of Shellfish Research, v. 17, no. 5, p Badylak, S., Kelley, K., and Phlips, E. J., 2004, A description of Pyrodinium bahamense (Dinophyceae) from the Indian river lagoon, florida, USA: Phycologia, v. 43, no. 6, p Badylak, S., and Phlips, E. J., 2004, Spatial and temporal patterns of phytoplankton composition in subtropical coastal lagoon, the Indian River Lagoon, Florida, USA: Journal of Plankton Research, v. 26, no. 10, p Balech, E., 1985, A Revision of Pyrodinium-Bahamense Plate (Dinoflagellata): Review of Palaeobotany and Palynology, v. 45, no. 1-2, p Bouillon, S., Dahdouh-Guebas, F., Rao, A., Koedam, N., and Dehairs, F., 2003, Sources of organic carbon in mangrove sediments: variability and possible ecological implications: Hydrobiologia, v. 495, no. 1-3, p Bouillon, S., Moens, T., and Dehairs, F., 2004, Carbon sources supporting benthic mineralization in mangrove and adjacent seagrass sediments (Gazi Bay, Kenya): Biogeosciences, v. 1, no. 1, p

46 Bruno, J. F., Boyer, K. E., Duffy, J. E., Lee, S. C., and Kertesz, J. S., 2005, Effects of macroalgal species identity and richness on primary production in benthic marine communities: Ecology Letters, v. 8, no. 11, p Buyukates, Y., and Roelke, D., 2005, Influence of pulsed inflows and nutrient loading on zooplankton and phytoplankton community structure and biomass in microcosm experiments using estuarine assemblages: Hydrobiologia, v. 548, p Byun, D. S., Wang, X. H., Hart, D. E., and Cho, Y. K., 2005, Modeling the effect of freshwater inflows on the development of spring blooms in an estuarine embayment: Estuarine Coastal and Shelf Science, v. 65, no. 1-2, p Cardinale, B. J., Srivastava, D. S., Duffy, J. E., Wright, J. P., Downing, A. L., Sankaran, M., and Jouseau, C., 2006, Effects of biodiversity on the functioning of trophic groups and ecosystems: Nature, v. 443, no. 7114, p Caroppo, C., Turicchia, S., and Margheri, M. C., 2006, Phytoplankton assemblages in coastal waters of the northern Ionian Sea (eastern Mediterranean), with special reference to cyanobacteria: Journal of the Marine Biological Association of the United Kingdom, v. 86, no. 5, p Cloern, J. E., Grenz, C., and VidergarLucas, L., 1995, An empirical model of the phytoplankton chlorophyll:carbon ratio - The conversion factor between productivity and growth rate: Limnology and Oceanography, v. 40, no. 7, p Corbisier, T. N., Soares, L. S. H., Petti, M. A. V., Muto, E. Y., Silva, M. H. C., McClelland, J., and Valiela, I., 2006, Use of isotopic signatures to assess the food web in a tropical shallow marine ecosystem of Southeastern Brazil: Aquatic Ecology, v. 40, no. 3, p de Boer, W. F., 2000, Biomass dynamics of seagrasses and the role of mangrove and seagrass vegetation as different nutrient sources for an intertidal ecosystem: Aquatic Botany, v. 66, no. 3, p Dehairs, F., Rao, R. G., Mohan, P. C., Raman, A. V., Marguillier, S., and Hellings, L., 2000, Tracing mangrove carbon in suspended matter and aquatic fauna of the Gautami-Godavari Delta, Bay of Bengal (India): Hydrobiologia, v. 431, no. 2-3, p Dittmar, T., Lara, R., and Kattner, G., 2001, River or mangrove? Tracing major organic matter sources in tropical Brazilian coastal waters: Marine Chemistry, v. 73, no. 3-4, p Downing, A. L., 2005, Relative effects of species composition and richness on ecosystem properties in ponds: Ecology, v. 86, no. 3, p Duarte, P., Macedo, M. F., and da Fonseca, L. C., 2006, The relationship between phytoplankton diversity and community function in a coastal lagoon: Hydrobiologia, v. 555, p Ebert, U., Arrayas, M., Temme, N., Sommeijer, B., and Huisman, J., 2001, Critical conditions for phytoplankton blooms: Bulletin of Mathematical Biology, v. 63, no. 6, p Ehleringer, J. R., Buchmann, N., and Flanagan, L. B., 2000, Carbon isotope ratios in belowground carbon cycle processes: Ecological Applications, v. 10, no. 2, p

47 Farida, F., Bajarias, A., and Relox, J., 1996, Hydrological and climatological parameters associated with the Pyrodinium blooms in Manila Bay, Philippines: Paris (France), UNESCO. Fry, B., and Sherr, E. B., 1984, Delta-C-13 Measurements as Indicators of Carbon Flow in Marine and Fresh-Water Ecosystems: Contributions in Marine Science, v. 27, no. SEP, p Godhe, A., Noren, F., Kuylenstierna, M., Ekberg, C., and Karlson, B., 2001, Relationship between planktonic dinoflagellate abundance, cysts recovered in sediment traps and environmental factors in the Gullmar Fjord, Sweden: Journal of Plankton Research, v. 23, no. 9, p Goni, M. A., Monacci, N., Gisewhite, R., Ogston, A., Crockett, J., and Nittrouer, C., 2006, Distribution and sources of particulate organic matter in the water column and sediments of the Fly River Delta, Gulf of Papua (Papua New Guinea): Estuarine Coastal and Shelf Science, v. 69, no. 1-2, p Gonneea, M. E., Paytan, A., and Herrera-Silveira, J. A., 2004, Tracing organic matter sources and carbon burial in mangrove sediments over the past 160 years: Estuarine Coastal and Shelf Science, v. 61, no. 2, p Grzebyk, D., Bechemin, C., Ward, C. J., Verite, C., Codd, G. A., and Maestrini, S. Y., 2003, Effects of salinity and two coastal waters on the growth and toxin content of the dinoflagellate Alexandrium minutum: Journal of Plankton Research, v. 25, no. 10, p Gu, B. H., Chapman, A. D., and Schelske, C. L., 2006, Factors controlling seasonal variations in stable isotope composition of particulate organic matter in a soft water eutrophic lake: Limnology and Oceanography, v. 51, no. 6, p Hamasaki, K., Horie, M., Tokimitsu, S., Toda, T., and Taguchi, S., 2001, Variability in toxicity of the dinoflagellate Alexandrium tamarense isolated from Hiroshima Bay, western Japan, as a reflection of changing environmental conditions: Journal of Plankton Research, v. 23, no. 3, p Hu, J. F., Peng, P. A., Ha, G. D., Mai, B. X., and Zhang, G., 2006, Distribution and sources of organic carbon, nitrogen and their isotopes in sediments of the subtropical Pearl River estuary and adjacent shelf, Southern China: Marine Chemistry, v. 98, no. 2-4, p Huisman, J., van Oostveen, P., and Weissing, F. J., 1999a, Critical depth and critical turbulence: Two different mechanisms for the development of phytoplankton blooms: Limnology and Oceanography, v. 44, no. 7, p , 1999b, Species dynamics in phytoplankton blooms: Incomplete mixing and competition for light: American Naturalist, v. 154, no. 1, p Irigoien, X., Huisman, J., and Harris, R. P., 2004, Global biodiversity patterns of marine phytoplankton and zooplankton: Nature, v. 429, no. 6994, p Jensen, M. O., and Moestrup, O., 1997, Autecology of the toxic dinoflagellate Alexandrium ostenfeldii: Life history and growth at different temperatures and salinities: European Journal of Phycology, v. 32, no. 1, p Jones, W. B., Cifuentes, L. A., and Kaldy, J. E., 2003, Stable carbon isotope evidence for coupling between sedimentary bacteria and seagrasses in a sub-tropical lagoon: Marine Ecology-Progress Series, v. 255, p