Phytoplankton account for roughly half of global primary production . They form the base of the oceanic food web e.g., and drive the ocean's biological pump, i.e., one of the Earth's largest biogeochemical cycles, with major implications for atmospheric CO2 and climate . Yet, our understanding of the processes controlling their spatio-temporal variations is limited, particularly at the oceanic submesoscale to mesoscale, that is at scales of the order of 0.1 to 100 km e.g., reviews by. In this study we focus on mesoscale eddies, that is vortices with diameters of 50 km or more, and thus leave out the submesoscale variations. This choice is largely driven by the spatial resolution of the data we employ, but it is also motivated by the fact that mesoscale eddies have been shown to dominate the ocean's kinetic energy spectrum , and affect phytoplankton in a major way . In comparison, the contribution of submesoscale processes to the variance in kinetic energy is smaller, and its role for phytoplankton variability, although potentially large is not well characterized. In contrast, mesoscale eddies have been recognized to be among the most important drivers for the spatio-temporal variance of phytoplankton e.g.,, as has been noted already from the analyses of some of the very first ocean color satellite images of chlorophyll (Chl), a proxy for phytoplankton biomass . Despite decades of work since this discovery, the mechanisms governing the interaction of phytoplankton with mesoscale eddies remain poorly understood, even though there is a broad consensus that different sets of mechanisms dominate in different regions and at different times, and that the different polarity of the mesoscale eddies tends to induce signals of opposite sign .

Lateral advection arising from local stirring of eddies has been argued to be a major driver globally. The argument is based on the observed correlation of the magnitude of eddy-associated Chl anomalies (δChl) and the larger-scale Chl gradient ambient to eddies . Further, it has been suggested that advection of Chl by eddies via trapping, i.e., the enclosing and dragging along of water masses, causes δChl , particularly in boundary current regions characterized by steep zonal Chl gradients. Numerous other potential mechanisms through which eddies affect phytoplankton have been identified e.g.,, including vertical and lateral advection of nutrients, restratification and vertical mixing, and providing spatial niches through isolation of water parcels. These mechanisms modulate the phytoplankton's light exposure, their nutrient availability or their grazing pressure, i.e., they affect their net balance between growth and decay. Thus, in contrast to the physical mechanisms of stirring and trapping where phytoplankton is merely passively being advected, these mechanisms create eddy-associated phytoplankton biomass anomalies by altering their biogeochemical rates.

In the Southern Ocean, an area of light and iron limitation of phytoplankton , with distinct Chl heterogeneity , and abundant with mesoscale eddies e.g.,, individual eddies haven been found to modulate Chl through many of the processes described above e.g.,. Here, we aim (i) to provide a reference estimate of the average seasonal Chl anomalies associated with mesoscale eddies in the different regions of the Southern Ocean, distinguishing anticyclones and cyclones, and (ii) to discuss the mechanisms likely causing the observed average imprint. Previous studies have used eddy kinetic energy as a proxy for eddy activity rather than sea level anomalies (SLA), which does not allow a distinction by polarity of eddies , did not focus on the Southern Ocean , or lacked a discussion of the seasonality of the relationship. In this study we show that the imprint of cyclones and anticyclones on Chl is mostly of opposite sign, largely zonal, and features a substantial seasonality along the ACC. Our results indicate that most of this mesoscale imprint can be explained by advection of Chl by mesoscale eddies.

Our approach is to identify individual eddies based on satellite estimates of SLA and combine those with satellite estimates of Chl . We discuss possible mechanisms playing a role based on large-scale Chl gradients and the local shape of the average imprint of eddies on Chl .

We first introduce our analysis framework before describing the methods and data sources. This allows us to explain the approaches we use to assess the potential mechanisms explaining the δChl associated with Southern Ocean mesoscale eddies.

The zonal pattern of the δChl identified here for the Southern Ocean is similar to that described by across the world's oceans. Analogous to the results of our analyses, they also found spatial variations in the sign of δChl associated with either cyclonic and anticyclonic eddies. Yet, there are also substantial differences, especially along the ACC, where, e.g., the δChl is more widespread and more intense than previously acknowledged. Further, the seasonal variations in the Southern Ocean appear to be stronger than elsewhere , except, perhaps, for the eastern Indian Ocean and the South China Sea . Possibly, the underappreciated δChl along the ACC is due to previous conflation of seasonal anomalies of opposite sign, resulting in a much weaker annual signal. To our knowledge, such seasonal changes in the sign of δChl in a particular region have not been reported before. Hence, the strong seasonality accompanied by a change in the sign of δChl along the ACC and south of the ACC appears to be rather specific to the Southern Ocean.

The spatio-temporal variability of δChl may not be that surprising in hindsight, given that the same mechanism, e.g., advection can lead to either positive or negative signs for the same polarity depending on the sign of the lateral gradient. In addition, several mechanisms may be involved simultaneously, so that small differences in their relative importance can lead to substantial differences in the net sign of the response . Nevertheless, we have demonstrated that most of the eddy induced signatures of δChl in the Southern Ocean are likely due to stirring, a mechanism that has been shown to control δChl in the low to mid-latitude ocean as well . Stirring is an effective mechanism for eddies to cause δChl as eddy rotation is omnidirectional and thus necessarily perpendicular to the ambient Chl isolines. This fact, combined with the steep meridional Chl gradients in the Southern Ocean, favor stirring as the driving mechanism for δChl. Stirring in such an environment of meridional Chl gradients supports Chl anomalies of a banded, zonal structure, similar in pattern and magnitude to the actual observed δChl, in most regions and seasons. Stirring is also favored compared to trapping by the fact that the majority of the eddies are relatively short lived and also have low translational speeds, such that the average eddy does not get far during its lifetime. This means that the eddy is much more likely to efficiently stir the environmental gradient due to its rotation than to move great distances up or down the gradient.

Next to stirring, our work elucidated also the importance of the other processes, namely trapping and changes in biogeochemical rates, in certain regions and at certain times. This leads to a relatively complex mosaic of dominance across space and time in the Southern Ocean (Fig. ). This synthesis figure reveals that stirring as the sole mechanism is limited to the subtropical waters outside of regions with deep winter mixed layers, and for anticyclones along the northern ACC from summer to autumn (Fig. , yellow). Our results suggest that trapping contributes to δChl for anticyclones along the southern ACC from summer to autumn and in Antarctic waters in autumn and spring. It also adds to the δChl of cyclones in most regions and seasons, except for parts of the subtropical waters (see also Fig. a, south and southwest of Australia). Yet, the magnitude of the potential of trapping is generally small, with the exception, perhaps, of a few specific regions, such as the eastern boundary currents, and those to the southeast of the Kerguelen Islands, and in the Drake Passage (Fig. S3 in the Supplement, see also ). In these regions, eddies tend to move down intense zonal Chl gradients (Fig. S2 in the Supplement, right column), carrying their high initial Chl with them. This tends to result in positive δChl year round for both anticyclones and cyclones (Fig. ). An additional possible explanation for the positive δChl is the offshore advection of iron trapped in the nearshore region by eddies that fuels extra growth in the offshore waters, as suggested e.g., for Haida eddies in the North Pacific , or for eddies passing the Kerguelen Plateau .

The weaker role of trapping relative to stirring can be explained by the inherently westward propagation of mesoscale eddies, meaning a propagation largely along Chl isolines, as zonal Chl gradients typically are much smaller than meridional Chl gradients. An additional reason is the aforementioned short propagation distance of an average eddy. Moreover, the efficiency of trapping is often also reduced owing to the trapped waters from the eddies' origins being diluted along their pathways . This dilution effect may help to explain also the puzzling observation that despite stirring being the dominant process overall, the spatial structure of the δChl within the eddies is much more monopole than dipole (Figs. , ). This can be resolved by hypothesizing that the lateral entrainment weakens the dipole component of the δChl generated by stirring, while strengthening the monopole component (see illustration in Fig. a).

The clearest case for a substantial contribution of changes in biogeochemical rates on δChl was found for the northern ACC region during winter and spring, when the mixed layers are deep (Fig. , blue), and correlations of Chl and SLA are negative. The associated negative δChl of anticyclones is consistent with the mechanism of an amplification of the prevailing light limitation in the deep mixed layers . As a result of their suppressing of the thermoclines, anticyclones tend to deepen the mixed layer depths by several tens of meters, especially in winter . Hence, phytoplankton within the mixed layer will be exposed to reduced mean light levels in anticyclones as compared to ambient waters, leading to lower phytoplankton growth. The opposite is the case for cyclones.

In the same region from summer to autumn, the weak trapping potential, the pronounced monopole-shape of δChl and the closed Chl contours suggest that the δChl is at least partly caused by the effects of eddies on the local biogeochemical rates. Here, the positive correlations of SLA and δChl could arise due to eddy-induced modifications of the prevailing iron limitation. Anticyclones could reduce the iron limitation and lead to positive δChl owing to their being more weakly stratified, leading to intensified vertical mixing in the high wind conditions of the Southern Ocean, bringing more iron from below to the surface. Vice-versa, the iron limitation could be enhanced by cyclones owing to their weaker vertical mixing . Hypothetically, an alleviation of grazing pressure due to reduced predator–prey encounter rates in deepened mixed layers in anticyclones could further favor positive δChl, and more shallow mixed layers could increase grazing pressure for cyclones. Thus, we argue that along the northern ACC, the seasonal sign switch of δChl could be explained by varying degrees of light and iron limitation and grazing pressure over the course of the year .

Finally, along the southern ACC and in Antarctic waters in autumn to spring, the potential of stirring and trapping oftentimes are of the same sign. However, δChl associated with eddies is insignificant in these waters in many places (dark gray regions, Fig. ). Presumably, these situations arise because the eddy effects on the local biogeochemical rates may almost perfectly cancel the advective effects.

Our analysis is constrained to the surface ocean, hence three caveats need to be kept in mind: (i) one potential issue is non-homogeneous vertical Chl profiles, e.g., the presence of unrecognized subsurface Chl maxima, but subsurface Chl maxima are presumably not prominent in our focus area , as wind speeds are high and mixed layers deep, promoting well-mixed Chl levels over the upper ocean; further, surface and mixed layer depth integrated analyses provide similar results in terms of SLA-Chl correlations (based on model simulations, Hajoon Song, personal communication, 2017), supporting the assumption that an analysis of surface Chl is representative for the total Chl in the water column. (ii) Modification of mixed layer depths by eddies may result in a surface Chl concentration modification due to a dilution effect. Especially in winter to spring, when the mixed layers are deep, we cannot exclude that this effect contributes to the observed δChl. Yet as noted in (i), surface and mixed layer depth integrated analyses provide similar results in a model simulation. (iii) Potential effects of eddies on phytoplankton growth presumably occur mostly in the lower euphotic zone and may thus be expressed more weakly at the surface . Therefore, we note that because our study is based on ocean surface data it may underestimate the total effect of mesoscale eddies on biogeochemical rates.

We may further underestimate the overall effect of Southern Ocean eddies on Chl, because of additional effects of mesoscale eddies that are not considered in our analysis. Such effects include the impact of smaller mesoscale features, and of submesoscale processes near the edges of eddies , e.g., eddy-jet interactions and associated horizontal shear-induced patches of up- and downwelling. Such features are included in our analysis only insofar as they have rectified effects on the larger mesoscale Chl patterns resolved by the data we use. Another effect we do not consider is non-local stirring , the contribution of eddies to lateral dispersion outside the eddies' cores in interaction with the ambient flow. This effect, for instance, shapes iron plumes downstream of shelves along the ACC, thus preconditioning Chl blooms . Therefore, we note that the overall effect of mesoscale eddies on biogeochemical rates may be larger than suggested by our analysis of the mesoscale, local, imprint of eddies on Chl. Finally, we note that our analysis does not include the effect of submesoscale processes outside eddies as well as any unstructured turbulence in general.

The prevalent and strong correlations between anomalies in surface Chl and mesoscale variability have triggered substantial research, but many unresolved issues remain, particularly regarding their causes . With this study, we aim to provide an observational reference for the seasonal climatological δChl associated with mesoscale eddies across the Southern Ocean, a region where the detailed regional and seasonal relationship of eddies and Chl previously had not been discussed. To this end, we combined satellite estimates of Chl with ocean mesoscale eddies (diameters larger than ∼50 km) identified based on satellite estimates of SLA. The very large number of collocations of eddies and Chl allowed us to retrieve statistically robust results despite the frequent data gaps and the high spatio-temporal variability of Chl.

We found a relatively complex pattern of Chl anomalies (δChl) associated with mesoscale eddies, with many anomalies exceeding ±10 % of their mean value over wide areas of the Southern Ocean. The δChl for cyclones is positive in subtropical waters, but negative along the ACC; anticyclones show a similar pattern, but of opposite sign. A pronounced seasonality of the imprint is apparent especially along the ACC and in Antarctic waters, featuring a sign switch of the anomaly over the course of the year.

While multiple mechanisms may be at play at the same time to cause the observed δChl , our analyses suggest that lateral advection due to stirring by eddies and associated lateral entrainment and permeable trapping explain a large fraction of the observed δChl. This conclusion is based on our analysis of the climatological Chl gradients, eddy rotation and propagation pathways, and the local shape of the δChl of eddies.

A prominent region and season where eddy-induced advection is insufficient to explain δChl are the northern ACC characterized by deep mixed layers during winter and spring and a seasonal sign switch of δChl. Here, winter to spring negative and positive δChl of anticyclones and cyclones, respectively, are consistent with changes in mixed layer depth and the ensuing light regime. The opposite signs of δChl from summer to autumn are consistent with an abatement of iron limitation by anticyclones via a relatively weak stratification facilitating vertical mixing, and, possibly, with an abatement of grazing pressure caused by anticyclones through deepened mixed layers; and vice versa for cyclones. In other regions and seasons our analysis does not exclude a modulation of δChl by effects of eddies on biogeochemical rates, even though our results suggest that lateral advection is likely the dominant mechanism.

Future work may include the investigation of where and when Southern Ocean eddies substantially affect biogeochemical rates, such as through modulation of alternating roles of iron and light limitation as well as grazing pressure along the ACC . The growing number of sub-surface biogeochemical measurements across eddies may be of help here, such as those collected by the increasing number of biogeochemical floats (http://biogeochemical-argo.org, last access: 2 August 2018). In addition, targeted experiments with numerical ocean-biogeochemical models with the option to alternately switch on and off Chl sources and sinks could be employed to shed light on the questions of what the role of eddy-effects is on Chl sources and sinks relative to advection, for higher trophic levels , or for the magnitude and structure of export . Furthermore, such models could be used to assess if these effects of eddies on phytoplankton substantially affect Southern Ocean biogeochemistry. Of particular interest are their modifications of the mode waters that originate from the Southern Ocean region with deep winter mixed layers. This is crucial, as these mode waters supply the low latitude ocean with nutrients and sequester a substantial amount of anthropogenic carbon . The final thread is the expansion of this work to smaller scales, and perhaps also more ephemeral turbulent structures, such as fronts.