Tables

Seasonal changes in the flux of planktonic foraminifera have to be understood to interpret corresponding proxy-based reconstructions. To study the seasonal cycle of planktonic foraminifera species we developed a numerical model of species concentration (PLAFOM). This model is forced with a global hydrographic dataset (e.g. tem5 perature, mixed layer depth) and with biological information taken from an ecosystem model (e.g. “food type”, zooplankton abundance) to predict monthly concentrations of the most common planktonic foraminifera species used for proxies: N. pachyderma (sinistral and dextral varieties), G. bulloides, G. ruber (white variety) and G. sacculifer. The sensitivity of each species with respect to temperature (optimal temperature and 10 range of tolerance) is derived from sediment-trap studies. Overall, the spatial distribution patterns of most of the species are comparable to core-top data. N. pachyderma (sin.) is limited to polar regions, N. pachyderma (dex.) and G. bulloides are the most common species in high productivity zones like upwelling areas, while G. ruber and G. sacculifer are more abundant in tropical and subtropical 15 oligotrophic waters. Modeled seasonal variation match well with sediment-trap records in most of the locations for N. pachyderma (sin), N. pachyderma (dex.) and G. bulloides. G. ruber and G. sacculifer show, in general, lower concentrations and less seasonal variability in all sites. The lower variability is reflected in the model output, but the small scale variations are not reproduced by the model in several locations. 20 Due to the fact that the model is forced by climatological data, it can not capture interannual variations. The sensitivity experiments we carried out show that, inside the temperature tolerance range, food availability is the main parameter which controls the abundance of some species. The here presented model represents a powerful tool to explore the response of 25


Introduction
Planktonic foraminifera are widely used for paleoceanographic reconstructions. They constitute a minor percentage of total zooplankton, but the deposition of their shells on the ocean floor contributes substantially to sediments. Individual species of planktonic foraminifera are essentially distributed geographically as well as vertically in the water 5 column, according to their requirements for physiology, feeding and reproduction within specific temperature ranges (e.g., Bé and Hamilton, 1967;Bé and Tolderlund, 1971). Shells of planktonic foraminifera extracted from marine sediments serve as an archive of chemical and physical signals that can be used to quantify past environmental conditions, such as sea-water temperature (e.g., Pflauman et al., 1996;Malmgren et al., 10 2001), stratification of the thermocline (e.g., Mulitza et al., 1997), atmospheric CO 2 concentration (Pearson and Palmer, 2000) and biological productivity (Kiefer, 1998). Past sea-surface temperatures can be estimated by either quantifying differences between modern and fossil species assemblages (e.g., CLIMAP, 1976;Pflauman et al., 1996;Malmgren et al., 2001); or by analyzing the isotopic or trace-element compo- 15 sition of calcite in the shell (e.g., Rohling and Cooke, 1999;Lea, 1999). In general, all estimation procedures are based on a correlation between modern conditions and assemblage composition or shell chemistry. This correlation is applied for the interpretation of the fossil assemblages by statistical techniques of varying complexity to yield an estimation of, e.g., past water temperatures (see, for example, Malmgren et could induce variations in the seasonal succession of the planktonic foraminifera and this variation needs to be quantified to understand corresponding proxy-based reconstructions. To study the seasonal variations of planktonic foraminifera species we have developed a numerical model for planktonic foraminifera at species level. Previously, 15Z arić et al. (2006) developed an empirical model based on hydrographic and productivity data. In contrast to the model ofZarić et al. (2006), we present a dynamic model for simulating the growth rate of the foraminifera population. This study shows model predictions for spatial and temporal distribution patterns of the five most important modern planktonic foraminifera used as proxies. We also show the comparison of the model silicate as nutrients.
The ecosystem model is driven by hydrographic data that is derived from an ocean circulation model and from climatologies. The forcing data include the local processes of turbulent mixing, vertical velocity at the base of the mixed layer, and seasonal mixedlayer entrainment/detrainment. Horizontal advection is not included; thus there is no lateral exchange. For the implementation, all vertical structure is ignored using the mean values of the surface mixed-layer.
Previously, this two-dimensional model has been validated against a diverse set of field observations from several JGOFS (Joint Global Ocean Flux Study) and historical time series locations (Moore et al., 2002a), satellite observations, and global nutrient 15 climatologies (Moore et al., 2002b). The full list of model terms, parametrizations, equations and behaviour in the global domain is described in detail in Moore et al. (2002a,b).

PLAFOM
The planktonic foraminifera model determines the distribution of the following 5 20 species: Neogloboquadrina pachyderma (sinistral and dextral coiling varieties), Globigerina bulloides, Globigerinoides ruber (white variety) and Globigerinoides sacculifer. These species have often been considered to be sensitive to sea-surface temperature, and therefore their assemblage can be used to estimate past sea-surface temperatures by means of transfer functions. Introduction EGU copepods, pteropods, and ostracods (Caron and Bé, 1984;Spindler et al., 1984;Hemleben et al., 1985Hemleben et al., , 1989Arnold and Parker, 1999). Non-spinose species are largely herbivorous, although in some specimens muscle tissue has been found in food vacuoles (Anderson et al., 1979). On a seasonal scale, food is assumed to be the predominant factor affecting the distribution of planktonic foraminifera. Planktonic foraminifers 5 appear to respond to the redistribution of nutrients and phytoplankton very quickly, increasing the number of individuals within several days (Schiebel et al., 1995). Information about food availability is obtained from the ecosystem model. In the model, the food sources may be either zooplankton, small phytoplankton, diatoms or organic detritus. 10 For each species the change in foraminifera concentration, expressed as biomass [mmolC/m 3 ], is calculated by the following equation: where F is the foraminifera carbon concentration, GGE (gross growth efficiency) is the portion of grazed matter that is incorporated into foraminifera biomass, which we 15 assume to be constant regardless of the food source (C). The food sources are diatoms (D), small phytoplankton (SP ), zooplankton (Z) or detritus (DR). Each species of foraminifera has a different food preference and therefore this is represented using different grazing proportions. The food supply predicted by the ecosystem model is calculated with reference to carbon weight, the same units as for the foraminifera in the 20 model.

Grazing in PLAFOM
The growth rates are determined by available food using a modified form of Michaelis-Menton kinetics (Eq. 2), of planktonic foraminifera consume a wide variety of zooplankton and phytoplankton prey, and they are capable of a reasonably flexible adaptation to varying trophic regimes. The food of N. pachyderma (sinistral and dextral varieties) comprises, almost exclusively, of phytoplankton, commonly diatoms (Hemleben et al., 1989). G. bulloides presents biological characteristics that place it on the borderline between spinose and 10 non-spinose species; while most spinose species possess algal symbionts G. bulloides does not (Hemleben et al., 1989;Murray, 1991). G. ruber and G. sacculifer are both spinose species hosting dinoflagellate endosymbionts. They feed mostly on zooplankton, although G. ruber has lower zooplankton dependence than other spinose species (Hemleben et al., 1989). Culture experiments of G. sacculifer confirm that it 15 depends on zooplankton food (Bé et al., 1981). It is adapted to low productivity areas (mainly the centers of the oceanic gyres), perhaps because it obtains nutrition from its symbionts (Watkins et al., 1996); but the seasonal maximum abundance occurs when the productivity in these regions is maximal. To include this contradictory information for the parametrization of G. ruber and G. sacculifer we use maximum 20 nutrient and chlorophyll concentration as a separate input variable to monthly nutrient and chlorophyll data. Maximum growth rates for the foraminifera (G max ) varies with the food source. Since zooplankton carbon concentration is most of the time much lower than phytoplankton carbon concentration, when zooplankton is the food source G max is set higher than when diatoms, small phytoplankton or detritus are the food source. Introduction EGU more than in any other physical variable. The experimental work of Bijma et al. (1990) shows evidence for direct control of temperature over vital processes. They showed a correlation between in vitro temperature tolerance limits and the known natural limits of the experimental species. The tolerance limits of most species, however, are not sharply defined. Instead, departure from optimal conditions causes a gradual reduc-5 tion of vital processes (Arnold and Parker, 1999). Zarić et al. (2005Zarić et al. ( , 2006) compiled a planktonic foraminiferal flux dataset from time-series sediment-trap observations from 42 sites across the world ocean, and analyzed species sensitivity to temperature relating fluxes and relative abundances of seven species to sea-surface temperature. Based on this work, we approximate the temperature relation to a normal distribution. 10 Therefore each species exhibits an optimal SST and an SST tolerance range. The growth rate is limited by temperature through the parameter α (Eq. 3). The relationship with temperature assumes that the foraminifera concentration at any site is normally distributed, with an optimum temperature where the relative abundance is highest. Away from this optimum temperature the relative abundance decreases until a criti- 15 cal temperature beyond which the species does not occur. This pattern, with a central peak and symmetrical tails, can be approximated by Gaussian distribution (Eq. 3). The value of α varies between 0 (out of limit of tolerance) and 1 (optimal temperature).
Here, n is a parameter to scale the values of α between 0 and 1. The normal distri-20 bution is characterized by two parameters: the optimal temperature, T opt ; and some variability expressed as standard deviation, σ, which represents the SST (T s ) tolerance range of a give species. Species with small σ are more sensitive to temperature. 1σ around T opt accounts for the temperature where about the 68% of the population is living. The values of all parameters for each species are summarized in Table 2. Of N. pachyderma (sin.) exhibit the narrowest SST tolerance range. N. pachyderma (sin.) is absent at SSTs above 23.7 • C (Zarić et al., 2005). N. pachyderma (sin.) is a polar species and survives even within the annual sea ice (Antartic), where it feeds on diatoms (Dieckman et al., 1991;Spindler, 1996). N. pachyderma (dex.) and G. bulloides are present almost throughout the whole oceanic SST range; however, N. pachyderma 5 (dex.) exhibits a clear preference for mid-temperatures. For G. bulloides temperature does not seem to be a controlling factor. It has the second largest temperature tolerance, after N. pachyderma (dex.), and does not show a unimodal distribution when flux is ploted versus temperature (Zarić et al., 2005). G. bulloides comprises at least six different genetic types and as a result exhibits a polymodal distribution pattern (Darling 10 et al., 1999(Darling 10 et al., , 2000(Darling 10 et al., , 2003Stewart et al., 2001;Kucera and Darling, 2002). Zarić et al. (2005) showed in their study that in the tropical Indian Ocean G. bulloides is present at higher temperatures than in the Atlantic and Pacific Ocean. In this region highest abundances of G. bulloides occur at SSTs at which Atlantic as well as Pacific samples show reduced fluxes. Since our study is applied at a global scale, temperature calibra- 15 tion is based on the preferred temperatures of G. bulloides in the Pacific and Atlantic Ocean; excluding the samples of the North Indian Ocean. Instead, we modified the normal distribution for N. pachyderma (dex.) and G. bulloides to accept wider limits under high food availability through the parameter k.

20
The mortality equation comprises of three terms representing losses due to respiration, predation by higher trophic levels and competition (Eqs. 4-8).
Since our model does not include lateral advection, a minimum threshold is needed to preserve the foraminifera population over the winter at high latitudes or during periods with insufficient food supply in regions with high seasonal variability. We set the Introduction EGU minimum level the mortality term is set to zero (Eq. 6).
Predators specialized on planktonic foraminifers are not known, therefore the mortality equation does not explicitly depend upon predator abundance. To represent predation we choose a quadratic form dependent on foraminifera biomass itself (Eq. 5). Although not specifically modeled, this may be interpreted either as cannibalism or as predation by a higher trophic level not being explicitly modeled (Steele and Henderson, 1992;Edwards and Yool, 2000). The variable pl represents the quadratic mortality-rate coefficient, which is used to scale mortality to grazing. From a bioenergetic perspective predation is also temperature dependent. Food consumption rates typically increase 15 with increasing temperature; therefore higher trophic levels will exert more predation pressure with increasing temperature (M. Peck, personal communication). b is the parameter used to scale the temperature function between 0 and 1. Note that T s represents the absolute SST, and the maximum SST (T m ) assumed in the model corresponds to 303.15 K (30 • C). The respiration loss is a linear term of 6% per day (rl ), the 20 same value as assumed by Moore et al. (2002a) for zooplankton. Competition occurs between different species of foraminifera inhabiting the same regions. cl i j represents the maximum competition pressure of the species i upon the 4332 equator), and a temporal resolution of one month. This corresponds to the resolution of the underlying ecosystem model (Moore et al., 2002a,b) and to the underlying ocean model (Gent et al., 1998). Lateral advection is not included, therefore there is no connection between the gridpoints. We used the same forcing as Moore et al. (2002a). Mixed-layer temperatures are 10 from World Ocean Atlas 1998 (Levitus, 1982) modified for the model's grid (NODC-WOA98 data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from the web site at http://www.cdc.noaa.gov/), surface shortwave radiation from the ISCCP cloud-cover-corrected dataset (Bishop and Rossow, 1991;Rossow and Schiffer, 1991), mixed-layer depth is based on the density criterion (Monterey and Levitus, 1997). The 15 minimum mixed-layer depth is set at 25 m. The vertical velocity at the base of mixed layer is derived from the NCAR-3D ocean model. Since lateral advection is not explicitly included these data are smoothed and modified to account for the effect of lateral advection in upwelling areas. The turbulent exchange rate at the base of the mixed layer is set at a constant value of 0.15 m/day. Sea-ice coverage was obtained from 20 the EOSDIS NSIDC satellite data (Cavalieri et al., 1990). Atmospheric iron flux was obtained from the dust deposition model study of Fung (1994, 1995). More details about the forcing can be obtained from Moore et al. (2002a). Bottom boundary conditions are the same as for the zooplankton component of the ecosystem model. For all foraminifera species we assumed a uniform distribution inside 25 the mixed layer, whereas below the mixed layer the concentration was calculated as a function of the surface concentration and the mixed-layer depth. When the mixedlayer is shallow (a minimum of 25 m) the concentration below the mixed-layer will be 4333 Introduction EGU the 75% of the surface concentration. With the increasing of the mixed-layer depth, the concentration below it decreases linearly, until reaching the value 0 at a mixed layer depth of 100 m. This is a realistic limit, as the maximum production of the species in questions occurs within this depth range (Bé, 1977;Duplessy et al., 1981;Murray, 1991;Watkins and Mix, 1998 (Prell et al., 1999) to compare our model results with sedimentary faunal assemblages. This database contains coretop planktonic foraminifera counts from 1264 cores throughout the world ocean. We extended this database with the dataset by Pflauman et al. (1996), which contains 10 planktonic foraminifera counts for 738 surface samples from the North and South Atlantic; and with another 57 core-top samples from the eastern Indian Ocean (Martinez et al., 1998). For comparison with the model output the faunal data were transformed to relative abundances of the 5 foraminifera species simulated with our model. Another adaptation of this dataset has been done to take into account the size differences of 15 each species. Then, the number of individuals were transformed to biomass (mgC/m 3 ).
For this transformation, we calculated the volume occupied by the cytoplasm approximating the shape of all the species to a sphere and assuming that all the volume is occupied by the cytoplasm. For the mean size of each species we used sediment-trap data from Peeters et al. (1999). We assumed the carbon content of the cytoplasm is 20 the same in all species, 0.089 pgC/µm 3 (Michaels et al., 1995).
To asses the deviation between observed and modeled species distribution we calculated the root mean squared error (RMSE). For this, each core-top data was compared to the nearest model grid point. No averaging was applied to the core-top data. This is justified because the observational data base is identical for all species and our interest 25 is only in comparing model performance for the five species. For the inter-hemispheric comparison coretop data were regridded to the model grid; and the average value was calculated between core-top data corresponding to the same grid point. China Sea (Tian et al., 2005) and from the Arabian Sea (Schulz et al., 2002). Table 2 summarizes locations, details and references of the sediment-trap studies used in this study. Figure 1 ilustrates locations of the sediment-trap stations. Sediment traps show a high temporal resolution and record the flux continuously over several months or years. Because of the sinking speed of foraminiferal shells 15 (150-1300 m/day depending on their weight and size, Takahashi and Be, 1984), the sediment-trap samples are not significantly affected by dissolution, lateral advection or bioturbation and therefore can be directly related to modern surface hydrography (e.g., Tedesco and Thunell, 2003;Marchant et al., 2004;Mohiuddin et al., 2004;Zarić et al., 2005). However, due to the short duration of the collecting periods (most of Introduction

Spatial distribution patterns
Modeled global distribution patterns of the five foraminifera species are compared to core-top data ( Fig. 2-6). The results shown here correspond to the relative abundances (in biomass) of individual species considering only the five species included in the 5 model. The global output of N. pachyderma (sin.) shows the lowest mean squared error, around 10%, while for the rest of the species it varied between 20% and 25%. N. pachyderma (sin.) is a cold-water species. Previous work showed that it can survive and grow within Antartic sea ice, where diatom concentration is higher than in the water column (Dieckman et al., 1991;Spindler, 1996;Schnack-Schiel et al., 2001). 10 Core-top as well as modeled assemblages show the highest relative abundances (up to 100%) in polar waters (Fig. 2).
N. pachyderma (dex.) is typical of subpolar to transitional water masses. In the sediment surface samples N. pachyderma (dex.) shows a very high relative abundance in North Atlantic, in the Benguella upwelling system, parts of the Southern Ocean and 15 in the equatorial Pacific upwelling. It is also present in lower abundance in the upwelling systems of northwest of Africa. The model output shows very high concentrations in the Peru-Chile current and the eastern boundary upwelling systems, as well as south of Island, and moderate abundances at mid latitudes ( Fig. 3).
Like N. pachyderma (dex.), G. bulloides is typical of subpolar (Bé, 1977;Bradshaw, 20 1959) and transitional water masses (Tolderlund and Be, 1971), and is also found in upwelling areas (Duplessy et al., 1981;Hemleben et al., 1989;Thunell and Reynolds, 1984). Temperature does not seem to be a controlling factor in the distribution of this species, althought the exact relationships between environmental parameters and distribution in G. bulloides may be masked by the fact that this species group com- 25 prises several distinct genotypes (Darling et al., 1999(Darling et al., , 2000(Darling et al., , 2003Stewart et al., 2001;Kucera and Darling, 2002). Generally, the abundance of G. bulloides is related to high productivity areas (Prell and Curry, 1981;Bé et al., 1985;Hemleben et al., 1989 raudeau, 1993;Watkins and Mix, 1998;Zarić et al., 2005). G. bulloides shows a high relative abundance in the surface sediment samples of the North Atlantic Ocean, the upwelling systems of northwest of Africa and Benguela, the Southern Ocean, the North Indian Ocean and with to a lesser extent the upwelling region of Baja California. The model results show high concentrations of G. bulloides in the subpolar waters of both 5 hemispheres, in the eastern boundary currents of the southern hemisphere and in some locations of the Arabian Sea (Fig. 4). G. ruber exhibits two varieties; a pink and a white form. The pink variety is limited to the Atlantic Ocean, therefore we have only modeled the white variety. G. ruber (white) is a spinose species generally found in tropical to subtropical water masses. It hosts dinoflagellate endosymbionts (Hemleben et al., 1989), and is omnivorous. These characteristics of bearing spines, utilization of zooplankton prey and symbiotic association are typical of foraminifers adapted to oligotrophic waters. Watkins et al. (1996) suggested that the adaptation to oligotrophic waters could be because they obtain nutrition from their symbionts. The seafloor record shows a high relative abundance in the 15 central North and South Atlantic as well as the South Pacific and a less pronounced relative abundance in the South Indian Ocean up to 40 • S. The model output shows high relative abundances in tropical and subtropical waters of the Atlantic, Pacific and Indian Oceans, while it is absent, or in very low relative abundances, in the upwelling areas (Fig. 5). 20 G. sacculifer shows a clear preference for the upper temperature band and is absent at low temperatures (Zarić et al., 2005). Core-top data show it is limited to tropical waters and reflect its narrow temperature tolerance (Fig. 6). The relative abundance is lower compared to the other four species considered in the model. The highest abundances observed in surface sediments appears in the equatorial Pacific and central 25 Indian Ocean. It has very low concentrations in most of the core-top data of the Arabian Sea. The annual distribution pattern reflected in the model is limited to tropical waters, with very high concentrations in the equatorial Pacific. It is absent in eastern upwelling systems as well as in the upwelling of the Arabian Sea. Most of the shells of foraminifera in juvenile stage are not included in this size fraction.

EGU
Nevertheless, the objective of our study is to detect relative changes in the seasonal distribution, therefore the absolute values are not very important. In most of the cases the interannual variability is very high, for example in the Wedell Sea where the annual ice expansion plays an important role in the spring bloom, or in upwelling areas, where the intensification of the upwelling and the entrainment of nutrients is related to the 15 wind intensity. It should be noted that the model is forced with climatologies (i.e., long term mean), and therefore is unable to reproduce interannual variabilities. Figure 7 show examples of N. pachyderma (sin.) in northern North Atlantic (a), Northwest Pacific (b) and Wedell Sea (c). In all of these locations the interannual variability is very high, but the timing of the signal matches well between observed and 20 predicted data. hand, predicts an annual cycle with a double peak; one at the beginning of summer, matching with sediment-trap data; and a second one during the autumn (may and june) not reflected in the trap data. sediment-trap data from Japan Trench, in the Kuroshio current system domain (34.16 • N 141.98 • E), shows the highest fluxes during the spring (April-May) for two consecutive years. At this location, the model output 20 acts in accordance with observational data. Figure 9c represents another station north of the Kuroshio current, in the Northwest Pacific. The model predicts in this case an annual cycle with two peaks, a small one during the winter (December-January) and the main peak during early summer (May-June). The sediment-trap data shows quite an irregular pattern, but in both sampled years (1997/1998 and 1998/1999) there is 25 an increased flux during winter months, matching with the small peak predicted by the model. During the second sampling year, in 1999, we can observe a second peak in the summer, which was absent in the previous year. The last time series (Fig. 9d) represents the Walvis Ridge, located close to the Benguela upwelling, in the coastal- unimodal pattern with highest fluxes in September-October. In the Bay of Bengal, both model output and trap data show less variability than the other sites. At the Northwest Pacific, South of Japan, sediment-trap data revealed an irregular pattern. However, the data shows overall low flux from December to April with a high peak in June-July. The model predicts high concentrations from November to July with a decrease during the 20 summer months (July-September).
G. sacculifer shows low fluxes in all sediment-trap data used for model validation. In Fig. 11 we present some comparisons of model output and experimental data. In the rest of the stations the model predicts very low concentrations. When the model reaches the threshold value, setted as the minimum population size (0.01 mmolC/m 3 ), 25 the hydrographic component of the model starts to dominate over the population dynamic itself. Therefore, in most of the locations the comparison is not viable. The upper two panels (a and b in Fig. 11) represent two stations in the northern Indian Ocean, located on a north-south transect of Bay of Bengal. In the northern trap both 4340 the data and the model output show little variability throughout the year. The central trap shows a prominent increase of foraminiferal flux in July. The model prediction is in good agreement with the data in both cases, showing less variability in the northern site and increases of the G. sacculifer population during summer in the central site. In the western equatorial Atlantic (northern Brasil Basin) the fluxes were low during the entire sampling period, with maximum values in December and June. The model output does not match the data, with maximum concentrations during April-May and minimum during fall. The last example comes from the central Arabian Sea. In this site, the model predicts concentrations below the threshold value. At such low concentrations, the model is unable to reproduce variations due to the population dynamics 10 component. Therefore, for a better comparison we choose adjacent locations. The trap data shows a bimodal pattern, with a prominent increase increase of flux during the both monsoons (June-August and November-April). Instead of that, the model simulates high concentrations from July to November and a decrease during winter months. 15

Spatio-temporal distribution pattern
We analyzed the model prediction for the temporal variation of G. bulloides in the North Atlantic. Figure 12 shows the model output of G. bulloides concentrations throughout the year in the North Atlantic. The maximum concentrations occur around 40 • N during spring (March-April) and around 60 • N during summer (June-July), following the typical 20 phytoplankton bloom of North Atlantic (Colebrook, 1982;Litchman et al., 2006).

Sensitivity analysis: spatio-temporal distribution patterns with constant temperature
We run the foraminifera module with a constant temperature of 12 • C to test the sensitivity of G. bulloides to other environmental parameters (mainly food availability). This 25 temperature corresponds to the optimal temperature of this species assumed in the EGU model. Figure 13 shows the spatio-temporal distribution of G. bulloides in the North Atlantic for this experimental run. In general, absolute concentrations of G. bulloides are higher than in the standard run. The model shows again the highest concentrations first in the southern part (around 40 • ) and during summer the bloom is shifting to higher latitudes.

Comparison with core-top data
In general, the global distribution patterns of foraminifera species predicted by the model are similar to those expected from core-top data. The core-top data reflect the integrated flux through the water column, whilst our 10 model reflects the situation in the mixed layer. Accordingly, we can expect some discrepancies between model and core-top distributions due to the different depth habitats of the species. But the five species of foraminifera simulated with our model live in most of the cases in the upper part of the water column, thus at a global scale we can expect a similar distribution. In addition, sedimentary assemblages may be altered by 15 selective dissolution (Berger, 1968;Thunell and Honjo, 1981;Le and Thunell, 1996;Ditter and Henrich, 1999), by displacement through subsurface currents or by bioturbation processes (Bé, 1977;Bé and Hutson, 1977;Boltovskoy, 1994). Since we did not take into account any of the factors we also can expect some differences due to these processes. 20 N. pachyderma (sin.) is the only foraminifera species that growth in polar waters and also within sea-ice (Dieckmann et al., 1991;Spindler, 2006). The model output is in very good agreement with core-top data, showing very high concentrations at the poles and abscence or very low concentrations in the rest of the oceans. Darling et al. (2006) have shown that N. pachyderma (sin.) has many different cryptic species. As a 25 consequence, the temperature tolerance of the southern populations is larger than that 4342 in the southern hemisphere. This confirms that our parametrization is based on temperature tolerance of the southern population. However, this difference is not large, and therefore we conclude that using this particular model and its parametrization, the different genotypes of N. pachyderma (sin.) can be grouped as a single species.
Model results for the distribution of N. pachyderma (dex.) were, in general, in agree-10 ment with data from sediment surface samples. The model simulates very high relative abundances (up to 90%) in the three eastern boundary upwelling systems. These high relative abundances in the equatorial Pacific and the Benguella upwelling are also reflected in the sediment samples. In comparison with core-top data, in the Atlantic Ocean results at mid latitudes were overestimated, while those at higher latitudes were 15 underestimated. The pattern in the Pacific Ocean looks different. In the equatorial Pacific, north of the eastern boundary upwelling region, the model underestimates the relative abundance of N. pachyderma (dex.). At this region, the surface temperature is higher than the typical for N. pachyderma (dex.) (minimum temperatures above 22 • C). It is possible that N. pachyderma (dex.) individuals present in these core-tops repre-20 sent a population living below the mixed layer, as has been described in previous works (Murray, 1991;Pujol and Grazzini, 1995;Kuroyanagi and Kawahata, 2004), or that they are expatriated specimens from the upwelling region.
G. bulloides is common at mid-latitudes and subpolar waters, but it is also present in the subtropical waters of the Indian Ocean. It use to be related to high productivity ar-25 eas (Hemleben et al., 1989;Sautter and Thunell, 1989;Guptha and Mohan, 1996) EGU lloides genotypes. The two warm water types are found mainly in tropical/subtropical regions, whereas cold water types are found in transitional to subpolar waters (Kucera and Darling, 2002). Zarić et al. (2005) studied the sensitivity of several planktonic foraminifera species to sea-surface temperature and concluded that the population of G. bulloides present in the tropical Indian Ocean comprises mainly the warm-water 5 genotype. The parametrization of the model is done for a global scale without including the warm water type, and therefore the increased relative abundance of warm water G. bulloides in the tropical Indian Ocean is not captured by the model. The high concentrations simulated in the Arabian Sea are due to high phytoplankton concentration predicted by the ecosystem model. In the model, the temperature relation is approximated to a modified normal distribution. The modification of the normal distribution by the introduction of the food dependent relation (through the parameter CT ) allows G. bulloides to grow in tropical waters. The model underestimated G. bulloides relative abundances in northwest of Africa as a result of its overestimation of N. pachyderma (dex.) relative abundance. 15 The model simulated global distribution pattern of G. ruber (white) shows a good agreement with the core-top data. The relative abundances in the Atlantic Ocean are comparable to core-top data, while in the Indian Ocean abundances are overestimated. There are few core-top data from the North Pacific and although their low number question the accuracy of the model validation, they do appear to support the model.
Both core-top and model output show that the distribution of G. sacculifer is limited to tropical waters. The model captures well the distribution patterns in the Atlantic Ocean and the East Pacific. The predicted relative abundances of G. sacculifer in the Indian and West Pacific Oceans are underestimated in the model; the distribution seen in the sediment data shows a wider spatial coverage than in the simulated results. This could 25 be due to the competition exerted by G. ruber. G. ruber is overestimated in the Indian Ocean and therefore the competition exerting upon the other species is overstated. In general, the seasonal signal of species concentrations are similar to sediment-trap records of most of the species. Figure 7 shows some examples of N. pachyderma (sin.) in the northern North Atlantic, Northwest Pacific and Wedell Sea. In the sediment-trap data from the northern 5 North Atlantic (Fig. 7a) during the last two years of the sampling period the timing of the seasonal peak coincides with the model prediction, whereas in the first sampled year is out of phase. In the Northwest Pacific (station PAPA, Fig. 7b), there is a good agreement between data and model output, showing maximum seasonal values between March and July. During the winter of 1985 there is second bloom not predicted by the 10 model. We do not know if this phenomenon is a reoccurring feature since there are data gaps for this period in other years. In the last location, the Wedell Sea (Fig. 7c), the data do not show evidence of any bloom, probably due to the variability of ice dynamic in the area. Despite the interannanual variabilities, the intraannual variability matches well with the observed data. 15 The seasonal pattern of predicted concentration for N. pachyderma (dex.) matches fairly well with the trap record in most of the locations. In the Northwest Pacific (Fig. 8b) sediment-trap data was only available for one year. These data gaps make the model validation complicated, since we can not confirm patterns as typical for this region. Nevertheless, the model predicts a similar seasonal variation to that seen in the limited 20 sediment-trap data available. The observational data of Walvis Ridge (Fig. 8c) does no show a clear seasonal pattern during the first year of the sampling (1989). However, during the last two years there is an increase of the population during autumn (from September to December). This location is very close to Benguella upwelling system, and therefore is strongly influenced by the variability of the upwelling intensity and wind 25 system. The model output reflects a smoother situation, but the timing of the bloom matches well with those reflected by the trap data. In the Subantartic Zone (46.76 • S 142.07 • E, Fig. 8d EGU population seen in the trap-data during the first austral summer. There is no evidence of increase of N. pachyderma (dex.) flux during the second summer, but the trap data are no longe enough to confirm the absence of the summer peak. The model prediction capability of G. bulloides varies between locations. At some locations, for example in the Japan Trench (34.16 • N 141.98 • E), the predictions corre-5 spond to the observational data, with a single annual peak during spring. At the station located at the Walvis ridge (20.05 • S 9.16 • E), close to the coastal upwelling zone, the model predicts successfully the bimodal pattern, with the main bloom in the austral spring (October-November) and a secondary bloom in fall (May-June). Contrarily, in some other locations the result differed significantly from sediment-trap data (Fig. 9). 10 For example, the in Peru Chile current the model predicts two annual peaks, while the sediment-trap data clearly show a single peak during the austral summer (from December to end of January). This station is situated on the eastern boundary upwelling system of south America and therefore the environmental conditions and ecological successions are more complicated than in open ocean. In this region the model is not 15 able to predict the pattern shown by the sediment-trap data. It should be noted that the sediment-trap data exist only for a single year. It is, therefore, possible that this data does not reflect the true typical annual pattern. In the subarctic station of the Northwest Pacific (39.01 • N 147.00 • E), along the boundary of the Kuroshio and Oyashio currents, sediment-trap data showed the variability in a sampling period of two years. During the 20 second sampling year (1999) the data and the model output fit well with two maxima, one during the winter months (December-January) and a second one in early summer. During the first year (1998) the summer peak was absent. This station is located in the subarctic front, with a complex hydrography and mixing of several water masses (Mohiuddin et al., 2002). The absence of the summer peak could be related to interan-25 nual changes in water sources and water column conditions due to interaction between subarctic water masses (Oyashio Current) and subtropical water masses at this site.
In general, the sediment-trap data of G. ruber (white) show less variability than N. pachyderma and G. bulloides (Fig. 10). The seasonal pattern of G. ruber (white) at Cape 4346 EGU Blanc (21.15 • N 20.69 • W), located to the upwelling zone of the cold Canary Current, shows dominant flux-maxima in fall (September-December), with variable amplitudes (Fig. 10a). The model predicts a wider peak, from summer to winter, matching well with the timing of sediment-trap data (except for the last year, when the fall bloom is missing). In the western equatorial Atlantic (7.51 • S 28.03 • W), and Northwest Pacific 5 (25.00 • N 136.99 • E) sediment-trap data show quite an irregular pattern. Despite this, it is possible to distinguish a bimodal pattern, with one peak in June-July and a second in November-December. During the second year the winter peak in the western equatorial Atlantic is absent. At these stations the model does not predict the expected seasonal cycle. Because the sediment-trap data do not show a clear annual cycle, we can not confirm this pattern as the typical annual pattern, and therefore the validation at this site is not highly reliable. The site at the Bay of Bengal, in the northern Indian Ocean, shows less variability both in sediment-trap data and in the model output. The Bay of Bengal is affected by a monsoonal climatic regime, with the southwest monsoon (June to August) creating a small upwelling cell along the eastern coast of India; and 15 the northeast monsoon (November to April) forming a moderately developed Cyclonic Gyre (Guptha et al., 1997). Sediment-trap data does not show prominent maxima associated with monsoonal seasons, but even if the variability is quite low, it shows three flux maxima: during December to January, May to June and August to September. Nevertheless, these oscillations are very small compared to the other sites, and the 20 low variability is reflected in the model output.
Most of the sites from which sediment-trap data are available for, the model underpredicts the concentrations of G. sacculifer. In tropical waters, most of the sedimenttrap data are located in high productivity areas or close to upwelling zones. G. sacculifer is parametrized to be more abundant in low productivity regions, and therefore EGU for a proper comparison (Fig. 11). For example, in the Bay of Bengal, the model is in good agreement with the sediment-trap data. In the northern Bay of Bengal both sedimnent-trap data and model result show little variability throughout the year indicating that the monsoonal system does not substantially affect the G. sacculifer population. In the central Bay of Bengal the population increases during summer months, 5 probably due to the southwest monsoon (June to August), which dominates the northern Indian Ocean (Cullen, 1981). During the southwest monsoon surface circulation is dominated by a wind-driven anticyclonic gyre, generating a strong boundary current and a small upwelling cell along the eastern coast of India. Unlike in the northern Bay of Bengal, the foraminiferal population in the central Bay of Bengal is affected by 10 hydrographic changes associated to the monsoon seasons Guptha et al. (1997). The Brasil Basin (western equatorial Atlantic, Fig. 11c) is characterized by representing typical oligotrophic open-ocean conditions (Fischer and Wefer, 1996). This is reflected in sediment-trap data, which indicates very low fluxes of G. sacculifer. The model output also shows very low concentrations, but the small variations seem to be out of phase 15 with respect to the trap data, possibly because the values are close to the threshold value and at this point the physical processes start to dominate over the ecology. In the Arabian Sea, planktonic foraminifera reach highest production rates at monsoon seasons responding to hydrographical changes related to the monsoons: southwest monsoon (June-August), resulting in a upwelling cell in the coastline of Somalia and Oman; 20 and northeast monsoon (November-April) which suppresses the upwelling (Curry et al., 1992). During both monsoonal periods, the depth of the mixed layer increases and nutrient-rich water is injected into the euphotic zone (Monterey and Levitus, 1997). G. sacculifer has been identified as typical for northeast monsoon rather than the southwest monsoon seasons (Curry et al., 1992;Ishikawa and Oda, 2007), although the 25 sediment-trap data we used show a bimodal pattern related to both monsoon seasons, with a prominent increase during the southwest monsoon of 1987. The model cannot predict a clear response to the monsoonal system. This could be due to the fact that the ecosystem model is unable to predict a bimodal pattern, related to both monsoon 4348 during spring and maximum concentrations at higher latitudes in June, linked to the phytoplankton bloom succession in the the North Atlantic. This indicates that the temperature is not the only controlling factor, but the food supply plays an important role in the distribution pattern of this species. The experiment confirms the results of Ganssen and Kroon (2000), who from isotopic studies on North Atlantic surface sediments concluded that G. bulloides reflects temperatures of a northward migrating spring bloom. The model can reflect its preference for high productivity environments, as has been suggested by different authors (e.g., Bé and Tolderlund, 1971;Bé, 1977;Duplessy et al., 1981;Prell and Curry, 1981;Guptha and Mohan, 1996). 15 A foraminifera model at a global scale has been developed and its results compared with core-top and sediment-trap data. The model presented here globally predicts monthly planktonic foraminifera concentrations for 5 species: N. pachyderma (sin.), N. pachyderma (dex.), G. bulloides, G. ruber (white) and G. sacculifer. It is a nonlinear dynamic model simulating growth rate of the foraminifera populations coupled 20 to a ecosystem model (Moore et al., 2002a), which calculates food availability for foraminifera.

Summary and conclusions
Overall, the global distribution patterns of the predicted species are similar to coretop data. In the case of G. bulloides, several genotypes included in this group appear to be adapted to different environmental conditions, complicating the modeling at a global Introduction

EGU
scale. Incorporation of more species, hence, a better representation of competition, could further improve the quality of the model. Modeled seasonal variation agrees well with sediment-trap records in most of the locations for N. pachyderma (sin), N. pachyderma (dex.) and G. bulloides. G. ruber and G. sacculifer show, in general, lower concentrations and little variability at all sites.

5
The lower variability is reflected in the model, but it fails to adequately describe the the small scale variations in several locations. Due to the forcing, interannual variations can not be elucidated by our model.
The experimental run with constant temperature of 12 • C shows that temperature is not the only controlling factor in the distribution pattern of some species of planktonic foraminifera, but food availability (primary production in case of G. bulloides) is also an important factor controlling the habitat of some species.
Our model provides a tool to predict seasonal variations of five foraminifera species at different environments. With this approach the model has the potential to identify seasonal flux signals and to quantify the seasonal bias in foraminifera-based proxy 15 records.