Hydrobiologia (2011) 666:181–196 DOI 10.1007/s10750-010-0421-6
ZOOPLANKTON ECOLOGY
Copepod feeding in the ocean: scaling patterns, composition of their diet and the bias of estimates due to microzooplankton grazing during incubations Enric Saiz • Albert Calbet
Published online: 4 September 2010 Ó Springer Science+Business Media B.V. 2010
Abstract Here, we report insights from the compilation and analysis of data on marine calanoid copepod feeding rates in the ocean. Our study shows that food availability and body weight are major factors shaping copepod feeding rates in the field, with a relatively minor role of temperature. Although the maximal feeding rates of copepods that are observed in the field agree with the well-known 3/4 of body size scaling rule for animals, copepod feeding in the oceans is typically limited and departs from this rule. Ciliates and dinoflagellates appear to be highly relevant in the composition of copepod diets, and this represents an indirect increase in the flux of primary production that is likely to reach the upper trophic levels; this contribution is higher in the less productive systems and may help to explain accounts of proportionally higher standing stocks of copepods supported per unit of primary producer biomass in oligotrophic environments. Contrary to common belief, diatoms emerge from our dataset as
Electronic supplementary material The online version of this article (doi:10.1007/s10750-010-0421-6) contains supplementary material, which is available to authorized users. Guest editors: J.-S. Hwang and K. Martens / Zooplankton Behavior and Ecology E. Saiz (&) A. Calbet Institut de Cie`ncies del Mar, CSIC, Ps. Marı´tim de la Barceloneta, 37–49, 08003 Barcelona, Catalunya, Spain e-mail:
[email protected]
small contributors to the diet of copepods, except in some very productive ecosystems. We have also evaluated the bias in the estimation of copepod grazing rates due to within-bottle trophic cascade effects caused by the removal of microheterotrophs by copepods. This release of microzooplankton grazing pressure accounts for a relevant, but moderate, increase in copepod grazing estimates (ca. 20–30%); this bias has an effect on both the carbon flux budgets through copepods and on our view of their diet composition. However, caution is recommended against the indiscriminate use of corrections because they may turn out to be overestimates of the bias. We advise that both uncorrected and corrected grazing rates should be provided in future studies, as they probably correspond to the lower and upper boundaries of the true grazing rates. Keywords Copepod Zooplankton Grazing Ingestion Microzooplankton Body size Temperature Food Diatom Dinoflagellate Ciliate
Introduction The pivotal role of marine copepods in the transfer of primary production to upper trophic levels (fish) has motivated a major effort to study their feeding and trophic impact on phytoplankton. Relative to the large number of studies dealing with copepod feeding in the field, few efforts have been conducted to
123
182
compile existing data to detect general patterns in marine copepod feeding. Hansen et al. (1997) discussed the subject of body size scaling of feeding and growth in zooplankton, reviewing laboratorydetermined maximum rates for these processes from the literature. However, their study was not focussed on copepods, and the actual dataset for copepods was rather small. Peters & Downing (1984) specifically addressed the topic of body size scaling for copepods, but the feeding rates reported in their study were probably biased because they were largely based on older literature that viewed copepods as essentially being herbivorous and, thus, omitted other food sources. In addition, Calbet (2001) presented a comparative analysis of the importance of mesozooplankton (mostly copepods) grazing impact on phytoplanktonic primary production in different marine ecosystems. His study focussed on the direct trophic impacts of mesozooplankton on phytoplankton, not considering grazing on other prey, and it did not address issues of scaling. Finally, Calbet & Saiz (2005) compiled the data that were available at the time to highlight the contribution of ciliates to the diet of marine copepods and the biogeochemical implications of this in the functioning of pelagic food webs. This review of the available literature was further extended in Saiz & Calbet (2007) to summarise the current knowledge on marine copepod feeding rates and to search for general patterns and limiting factors of ecological relevance, e.g. body size (weight), temperature and food concentration. Although only a few years have passed since Saiz & Calbet (2007), we feel that the large number of field studies dealing with copepod feeding that have been published in the last 5 years and the new insights that might be derived from them justify this revisited version of Calbet & Saiz (2005) and Saiz & Calbet (2007). The number of field observations in Saiz & Calbet (2007) was rather limited (n = 90 for calanoid copepodite and adult stages). In this study, we increased the dataset by 70%, which was expected to allow for more robust conclusions, especially regarding the topic of the effects of temperature in driving field feeding rates. More importantly, in the past several years the appreciation by zooplanktologists of the complexity of the microbial component in planktonic food webs has resulted in more detailed accounts of the composition of marine copepod diets, which should be
123
Hydrobiologia (2011) 666:181–196
examined for their implications on the carbon fluxes mediated by copepods in the oceans. For instance, Nejstgaard et al. (1997, 2001) raised the point that, in copepod feeding incubations, the high clearance rates of copepods on microzooplankton release microzooplankton grazing pressure on phytoplankton compared to control bottles without added copepods, resulting in the underestimation of copepod grazing rates on phytoplankton. This artefact could explain many (often unpublished) observations of negative clearance rates of copepods on phytoplankton in field experiments. Originally, Nejstgaard et al. (1997, 2001) proposed several equations to correct for this bias by using estimates of microzooplankton grazing rates from concurrent dilution experiments (Landry & Hassett, 1982). In recent years, the application of such corrections—usually using microzooplankton grazing rates from the literature (e.g. Huo et al., 2008) or other proposed alternatives (Vargas & Gonzalez, 2004; Klaas et al., 2008), as opposed to from concurrent dilution experiments—are more frequently found in the copepod feeding literature. Surprisingly, no comprehensive studies have yet addressed the adequacy and consequences of such corrections on carbon fluxes mediated by copepod feeding. In this study, we have compiled field data on marine copepod feeding rates in the ocean and addressed the following three topics: (a) scaling patterns related to body weight, temperature and food concentration; (b) the composition of copepod diets in the field and (c) the bias in the estimation of copepod-mediated carbon fluxes due to the release of grazing pressure by microzooplankton in incubation experiments.
Materials and methods Field data collection Following the procedure in Saiz & Calbet (2007), we limited our dataset to field studies reporting data on copepod ingestion rates of the broadest range of potential prey, simultaneously considering both autotrophic and heterotrophic prey (always including at least ciliates, which are a major heterotrophic contributor to copepod diet; Calbet & Saiz, 2005). We restricted the compilation of our dataset to
Hydrobiologia (2011) 666:181–196
experiments dealing with the copepodite and adult stages of calanoid copepods, which are the most abundant in the literature. The copepod grazing rates were derived from incubations in natural seawater using either a single species or, in some cases, congeneric species that the researcher was not able to sort alive. Typically, only incubations over ca. 24 h periods (in exceptional cases up to 48 h) were considered to take into account daily feeding rhythms and to avoid biased estimates of daily rations. We omitted from consideration copepod grazing rates determined by the pigment gut content method because it is constrained to phytoplankton prey, excluding any heterotrophic components of the diet; in addition, determinations with this method typically do not take into account the daily variation that a 24-h incubation does and, therefore, represent only a snapshot of the feeding status of copepods. In most cases, the prey concentration in the experiments was determined under the microscope (except for a few studies where phytoplankton was quantified by chlorophyll concentration). For the analysis of copepod diets, four prey groups were considered where possible: ciliates, dinoflagellates, diatoms and a pooled group of other phytoplankters. In a few cases, only ciliates (or ciliates and dinoflagellates) were distinguished as prey in the original papers, and all of the other prey had been pooled as phytoplankton. Following the recommendations of Calbet & Saiz (2005), the data on ciliates (both biomass and ingestion rates upon) were increased by 30% to correct for losses in cell numbers during preservation (unless they were already corrected in the original paper). We searched the literature for copepod grazing rates in marine ecosystems in the Aquatic Sciences and Fisheries Abstract electronic database, using the keywords marine, copepod, and either feeding or grazing or ingestion. The database prior to 2004 was based on the previous review on copepod feeding scaling by Saiz & Calbet (2007), which was updated in this study with the new literature that had been produced up to summer of 2009. The references used in the previous study (Saiz & Calbet, 2007) were re-examined to gather new data regarding diet composition and to allow the application of trophic cascade corrections in grazing calculations. Data were obtained from tables or scanned figures from original papers or kindly provided by authors upon request. We used the
183
compiled average estimated feeding rate for each experiment, not the replicated values, to avoid inflating the sample size. In many cases, the authors were contacted to provide essential variables for our study that were not reported in the original paper such as copepod body weight or size, temperature, or initial prey biomass in the experiments, or any additional data or clarification that we required. The resulting dataset for the feeding rates of calanoid copepods in the field (153 data points, Online Resource 1) includes reports from oceanic and coastal waters extending from polar to tropical regions. All data are expressed on a carbon basis because most of the included papers and data provided by the authors expressed their results as carbon. In the few studies in which body mass data were not available, they were derived from the literature.
Results Figure 1 shows the distribution of the temperature, individual body weight and copepod daily ration (% of body carbon ingested daily) of the compiled data as a function of the food concentration at the study site (availability of sestonic carbon). The data collected from the literature cover a wide range of temperatures, from below 0°C (Disko Bay, Greenland) up to 30°C (Florida, USA), and with food availability reflecting very different situations, from poor environments (Gulf of Aqaba, Red Sea) to phytoplankton bloom situations in very productive areas (Mejillones Coast, Chile; Greenland, and Norway). The body sizes of copepods ranged from small calanoids (Acartia clausi and Acartia tonsa, Paracalanus parvus, Ctenocalanus vanus, several Clausocalanus species) to the large calanoids (e.g. Neocalanus cristatus, Calanus hyperboreus). Daily rations were clearly different according to the food concentration at the studied sites (Fig. 1), with median values of 2.3, 11.7 and 45.4% (day-1) for study areas characterised by sestonic carbon concentrations of \50, 50–250 and [250 lg C l-1, respectively. Effects of body size, food concentration and temperature on ingestion rates The scatterplots of the calanoid copepod ingestion rates (lg C ind-1 day-1) as a function of temperature
123
184
Hydrobiologia (2011) 666:181–196 b Fig. 1 Distribution of temperature, copepod body weight and
(a) 35
copepod daily rations (% body carbon ingested day-1) in the dataset as a function of food concentration in the study site
Temperature (°C)
30 25 20 15 10 5 0 -5 250
Food concentration (µ g C L-1)
Body weight (µ g C ind -1)
(b)
104
103
102
101
100 250
Food concentration (µ g C L-1) Daily ration (as % body C ingested d-1)
(c) 103 102 101 100
(°C), copepod body weight (lg C ind-1) and food availability (lg C l-1) are shown in Fig. 2. Ingestion rates are presented per capita to avoid spurious correlations of weight-specific rates with body weight. For all statistical analyses, the variables ingestion rate, food concentration and body weight were logarithmically transformed to allow for nonlinear relationships between the variables; temperature was not transformed because the logarithm of a biological rate is usually regarded as a linear function of temperature (Arrhenius law). Ingestion rate and food concentration showed a strong, significant relationship (r2 = 0.57, P \ 0.001), whereas the relationships between ingestion rate and both body weight and temperature were significant, but they explained little of the variance (respectively, r2 = 0.13 and r2 = 0.05; Fig. 2). As, in nature, the three predictive variables (food concentration, body weight and temperature) act on calanoid feeding rates at the same time, a multiple regression model should better estimate the dependence of feeding rates on these variables. A preliminary analysis of the pairwise correlation between the predicting variables (Fig. 3) showed that only body weight and temperature were significantly correlated (r = -0.68, P \ 0.001, n = 153); all other possible pairwise combinations showed no significant correlation (P [ 0.05). Although the inverse correlation between temperature and body size was not high enough to explain a masking effect due to collinearity (Zar, 1999), we opted to derive a new body mass variable free of temperature effects by calculating the residuals (Wres) of the linear regression analysis between temperature and log body mass (Fig. 3) in an attempt to evaluate this possibility. Therefore, we fitted the data to two alternative multiple regression models: logðI Þ ¼ a þ b logðCÞ þ c logðW Þ þ dT;
10
-1
10-2 250
Food concentration (µ g C L-1)
123
ð1Þ
where I is the ingestion rate (lg C ind-1 day-1); W is the body weight (lg C ind-1); C is the food concentration (lg C l-1); T is the temperature (°C); a is the intercept; b, c, and d are the corresponding regression coefficients; and
Hydrobiologia (2011) 666:181–196
(a) Ingestion rate (µg C ind-1 d-1)
10
185 bFig. 2 Bivariate plots of ingestion rates of calanoid copepods
log ING = 0.67 – 0.02 TEMP r2=0.05, p
