Quantifying the effect of seasonal and vertical habitat tracking on planktonic foraminifera proxies

The composition of planktonic foraminiferal (PF) calcite is routinely used to reconstruct climate variability. However, PF ecology leaves a large imprint on the proxy signal: seasonal and vertical habitats of PF species vary spatially, causing variable offsets from annual mean surface conditions recorded by sedimentary assemblages. PF seasonality changes with temperature in a way that minimises the environmental change that individual species experience and it is not unlikely that changes in depth habitat also result from such habitat tracking. While this behaviour could lead to an underestimation of spatial or temporal trends as well as of variability in proxy records, most palaeoceanographic studies are (implicitly) based on the assumption of a constant habitat. Up to now, the effect of habitat tracking on foraminifera proxy records has not yet been formally quantified on a global scale. Here we attempt to characterise this effect on the amplitude of environmental change recorded in sedimentary PF using core top δ18O data from six species. We find that the offset from mean annual near-surface δ18O values varies with temperature, with PF δ18O indicating warmer than mean conditions in colder waters (on average by −0.1 ‰ (equivalent to 0.4 C) per C), thus providing a firstorder quantification of the degree of underestimation due to habitat tracking. We use an empirical model to estimate the contribution of seasonality to the observed difference between PF and annual mean δ18O and use the residual 1δ18O to assess trends in calcification depth. Our analysis indicates that given an observation-based model parametrisation calcification depth increases with temperature in all species and sensitivity analysis suggests that a temperature-related seasonal habitat adjustment is essential to explain the observed isotope signal. Habitat tracking can thus lead to a significant reduction in the amplitude of recorded environmental change. However, we show that this behaviour is predictable. This allows accounting for habitat tracking, enabling more meaningful reconstructions and improved data–model comparison.


Discussion
Five out of six species analysed show a temperature dependency of the offset between δ 18 O of the foraminiferal shells and the annual mean δ 18 O of the upper water column (Fig. 2). In addition, these 245 species show a positive relation between apparent calcification depth and temperature (Fig. 5).
Together, these observations provide a strong indication that temperature, either directly or by acting on other temperature-related variables, causes changes in the habitat of foraminifera. Such an important role for temperature in determining the vertical and seasonal habitat is not unexpected given that temperature appears to be dominant in controlling the spatial distribution of species 250 (Morey et al., 2005;Bé and Hutson, 1977), their flux (Zaric et al., 2005) and seasonality (Jonkers and Kučera, 2015) and appears important for test growth (Lombard et al., 2009).
Several studies have shown that formation of secondary calcite layers (e.g. gametogenic calcite or a crust) at the end of the life of a specimen, presumably deep in the water column could be responsible for higher δ 18 O of sedimentary foraminifera compared to those collected in the upper water column 255 Clim. Past Discuss., doi:10.5194/cp-2016Discuss., doi:10.5194/cp- -125, 2016 Manuscript under review for journal Clim. Past Published: 5 December 2016 c Author(s) 2016. CC-BY 3.0 License. (Duplessy et al., 1981;Bé, 1980). To the best of our knowledge there is no evidence that such secondary calcite is formed with a different isotopic (dis)equilibrium than the lamellar calcite. We therefore assume that our inferences are not affected by differences in calcification during ontogeny.
Nevertheless, the addition of such a crust in deeper (colder) waters could in principle lead to the observed increase in apparent calcification depth with temperature because of steeper vertical 260 temperature gradients in the tropics. However, foraminifera grow their tests exponentially and the last chambers that make up most of the test mass are formed in the last few days of their life, presumably close to the time of the secondary calcite formation (Bé, 1980). The compositional contrast between the bulk of the lamellar calcite and the crust calcite is thus likely to be smaller than estimated from the comparison of surface tows and sediment (cf. Jonkers et al., 2016). Consequently, 265 the apparent calcification depth we infer here likely incorporates this effect and the increase in apparent calcification depth that we observe most likely reflects homeostatic habitat adjustment.  . 2). Some of these values are unrealistically large and stem from observations in the 280 northern North Atlantic south of 50° N, thus outside the general distribution range of the species. This suggests that these observations may reflect expatriated specimens that calcified in colder regions or may point to inaccuracies in the chronological control and reflect (partly) shells of glacial age.
Alternatively, these samples could be affected by admixture of sinistrally coiled N. incompta (Darling Clim. Past Discuss., doi:10.5194/cp-2016-125, 2016 Manuscript under review for journal Clim. Past Published: 5 December 2016 c Author(s) 2016. CC-BY 3.0 License. clear latitudinal shift in the timing of the peak flux (Jonkers et al., 2010;Jonkers et al., 2013;Jensen, 1998;Wolfteich, 1994;Kohfeld et al., 1996). However, the species is also known to inhabit a broad, but generally deeper, zone of the upper water column (Carstens et al., 1997;Pados and Spielhagen, 2014) where seasonal temperature is smaller than in the near surface layer, possibly rendering a seasonality effect difficult to detect. 290 At face value, the absence of a Δδ 18 O annual.mean -temperature trend in G. bulloides may suggest that this species holds the best promise of providing reconstructions of mean annual near surface conditions ( Fig. 2). However, the distribution of Δδ 18 O annual.mean is noisy, suggesting that caution is required to interpret the species proxy signal. Similar to N. pachyderma this species also shows clear latitudinal changes in seasonality (Jonkers and Kučera, 2015;Tolderlund and Bé, 1971). However, G. bulloides is 295 characterised by considerable cryptic diversity (Darling and Wade, 2008). Possible genotypic ecological differences could therefore obscure ecological patterns at the morphospecies level.
Alternatively, being an opportunistic species, depth and seasonal habitat variability of G. bulloides may be driven by other parameters than temperature. Indeed, previous studies have shown that the distribution of this species is driven by food availability (Schiebel et al., 1997;Jonkers and Kučera, 300 2015). Whether or not the species shows habitat tracking and how this would affect its fossil record remains unclear, but we caution that the result of our study cannot be taken to indicate that proxy records from this species record the actual magnitude of environmental change.
Since planktonic foraminifera seasonality and calcification depth appear to be affected by habitat 305 tracking, our ability to accurately reconstruct past ocean properties would benefit from improved understanding of the drivers of their habitat variability. In particular, the controls on depth (and calcification) habitat remain poorly constrained. Due to logistical challenges, very few studies exist that have attempted to systematically investigate depth habitat variability. On the other hand, the realisation of the importance of habitat homeostasy in planktonic foraminifera could help to 310 formulate more realistic, mechanistic models of planktonic foraminiferal distribution in time and space (e.g. Lombard et al., 2011) and further improve our capabilities of forward proxy modelling. At any rate, the observations and the simple conceptual modelling exercise shown here serve as Clim. Past Discuss., doi:10.5194/cp-2016-125, 2016 Manuscript under review for journal Clim. Past Published: 5 December 2016 c Author(s) 2016. CC-BY 3.0 License. reminder that assumptions of constant seasonality and depth habitat underlying many paleoceanographic studies are not valid and the implications thereof are likely to be substantial. Our 315 analysis indicates that an observed change in a proxy value reflects a change in the climate state as well as a change in the species habitat.

Implications
Habitat tracking behaviour of planktonic foraminifera has important implications for paleoceanographic reconstructions. It may suggest that the temperature niche of planktonic 320 foraminifera inferred from their abundance in the sediment (e.g. Kucera, 2007) may be overestimated since their occurrence is not driven by mean annual sea surface temperature, but rather by whether their temperature niche is realised at any depth or season. It should thus be possible to define planktonic foraminifera temperature ranges (sensitivity) more precisely, which may help to improve transfer functions and is important for understanding of their ecology. 325 Another consequence of habitat tracking is that spatial and temporal differences reflected in the sedimentary foraminifera represent an underestimation of the actual gradients in the mean conditions, because temperature change forces the foraminifera to live in a seasonal or vertical 'window' where conditions are closest to optimal (cf. Jonkers and Kučera, 2015). We observe considerable variability in the slope of the Δδ 18 O annual.mean -temperature relationships, but the average 330 for the four species that show the clearest signal (G. ruber (pink and white), T. sacculifer and N. incompta) is 0.1 ‰ °C -1 (Fig. 2). This is equivalent to a 40 % (0.4 °C °C -1 ) underestimation of reconstructed temperature change. The

Conclusions
Through comparison of observed and predicted δ 18 O data of six common planktonic foraminifera we 375 have demonstrated that the average geochemical signal preserved in a population of fossil shells shows a temperature-dependent offset from mean annual sea surface conditions. This most likely reflects shifts in the seasonal and depth habitat in response to temperature, or temperature-related environmental, changes (Fig. 9). As a consequence of this homeostatic behaviour, the fossil record of these species, and likely also of others, does not reflect the full range of climate variability. Our 380 analysis indicates that spatial and temporal gradients in temperature may be underestimated by 40 %, clearly highlighting the need to account for climate-dependent habitat variability in the interpretation of paleoceanographic records based on planktonic foraminifera. Using a simple empirical model we attempted to assess the relative influence of seasonality and depth habitat variability. While improvements to this empirical approach are possible, we observe species-specific 385 partitioning of depth habitat versus seasonality that appears consistent with oceanographic conditions within their areal distribution. In the tropical species G. ruber (pink) we find that habitat tracking is primarily due to adjustments in the calcification depth. This is in agreement with the larger vertical than seasonal temperature gradients in the tropics. The offsets from annual mean surface conditions in N. incompta, on the other hand, appear dominantly driven by changes in the 390 seasonality, consistent with the dominance of seasonal over vertical temperature variability in the regions where it occurs. Our data underscore the importance of ecology in setting the climate signal preserved in fossil foraminifera. The recognition of predictable habitat tracking will help to improve the accuracy of paleoceanographic reconstructions and aid model-data comparison.          (2003). Note that the basic patterns indicative of habitat tracking remain, but that the general 450 calcification depth appears greater, also at lower temperatures.