the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 26 Jun 2025
On the Role of Light and Mixing in Shaping Southwestern Atlantic Shelf Blooms
Abstract. The influence of light availability and mixed layer depth (MLD) on phytoplankton bloom dynamics was examined across the Argentine Continental Shelf in the Southwest Atlantic Ocean (SWAO). Using satellite-derived chlorophyll-a concentration, photosynthetically available radiation, and euphotic depth (Zeu) data, combined with reanalysis products for MLD and wind fields, the spatial and temporal variability of key phenological parameters was analyzed, including bloom initiation, peak timing, and bloom intensity, over the 1998–2019 period. Distinct geographic trends in bloom dynamics were observed. In the Central Shelf (CS), blooms typically initiate (May–August) and peak (September–November) relatively early which correlated with shallow MLDs and increasing light, while coastal areas showed even earlier initiation (April) due to highly variable environmental conditions. In turn, the Patagonian Shelf (PS) experienced delayed initiation (September onwards) and peaks (December–January) due to deeper MLDs and colder Subantarctic waters. Bloom intensity also exhibited spatial variability, with the highest values observed in the southern PS and regions influenced by frontal systems, where nutrient-rich upwelling and favorable light conditions enhanced phytoplankton growth. Statistical modeling revealed that light penetration (Zeu) and its interplay with mixing (Zeu:MLD ratio) were the strongest predictors of bloom anomalies at most sites. However, the predictive power of these relationships varied in regions influenced by local processes, like tidal mixing or frontal zones. Predictive models need to be integrated with regional oceanographic features to improve assessments of bloom phenology and primary production in such highly variable shelf ecosystems.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(2201 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2201 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-2033', Emmanuel Boss, 07 Jul 2025
-
AC1: 'Reply on RC1', Robert Frouin, 15 Jul 2025
We thank the reviewer for the thoughtful and constructive review, which we believe has helped clarify the scope, strengths, and limitations of our study. We respond below to each major point, incorporating supporting evidence and contextual rationale for our methodological choices.
Reviewer Comment:
"Chlorophyll is not phytoplankton biomass but a proxy (e.g. Cullen, 1985). There are other proxies accessible from remote-sensing that could allow one to evaluate the degree to which the observed chlorophyll is affected by photo-acclimation. The latter is a phenomena, well known from house plants, in which cells exposed to low light produce more chlorophyll per biomass compared to cells exposed to high light. Given the topic of the analysis done in this paper, looking at the influence of light on chlorophyll, accounting to this phenomenon is necessary (for example by looking also to C_phyto derived from particulate backscattering a product of remote sensing). In phytoplankton culture Carbon/chl ratios vary from 30-300, an order of magnitude."
Response:
We agree with the reviewer that chlorophyll-a concentration is not a direct measure of phytoplankton carbon biomass and is subject to variability due to photoacclimation, species composition, and environmental conditions, as originally outlined by Cullen (1985) and in subsequent studies (e.g., Geider, 1987; Sathyendranath et al., 2009). Indeed, photoacclimation leads to significant changes in intracellular chlorophyll content, particularly in response to varying light regimes, which is highly relevant when studying the influence of light on phytoplankton dynamics.
While we fully recognize these limitations, our choice to use surface chlorophyll-a (Chl-a) as the primary indicator in this study was motivated primarily by its robustness and consistency as a satellite-based proxy available for long-term, large-scale studies of phytoplankton dynamics. This is particularly true for studies aiming to investigate phenological patterns over two decades across an extensive and heterogeneous shelf system such as the Southwestern Atlantic.
Alternative remote sensing products, such as particulate backscattering-derived phytoplankton carbon (C_phyto), have limited validation in our study region and insufficient temporal coverage for the 1998–2019 period. Furthermore, previous tests using the carbon-based productivity model (CbPM; Behrenfeld et al., 2005) in this region (Dogliotti et al., 2014) showed poor performance, further constraining the utility of satellite-derived C_phyto in our analysis.
Observed regional variability in the C:Chl ratio supports the reviewer’s point, and some of us have indeed investigated this in other studies. For example, C:Chl ratios ranged from 4.8 to 187.7 at the EPEA station on the northern Argentine coast (Silva et al., 2009) and were as low as 3–12 in the Magellan Strait (Lutz et al., 2016). These findings reflect the influence of both environmental gradients and species composition across this broad shelf ecosystem.
Given these constraints, our study does not attempt to quantify absolute biomass or productivity but rather focuses on analyzing the spatiotemporal variability of bloom intensity and timing based on Chl-a patterns. The phenological metrics we compute (e.g., bloom initiation, peak timing, bloom anomalies) are relative in nature and rely on internal thresholds rather than absolute biomass values, thereby minimizing the impact of photoacclimation-driven variability on our conclusions.
Nevertheless, we acknowledge that interpreting Chl-a variability solely as a reflection of biomass changes can be misleading without accounting for photoacclimation. We will explicitly clarify this limitation in the revised manuscript and acknowledge that Chl-a-based interpretations may, in some cases, reflect changes in intracellular pigment concentration rather than actual phytoplankton accumulation.
In future work, we agree that incorporating regionally validated C_phyto products or in situ C:Chl ratio observations would significantly enhance our understanding of the physiological and ecological controls on bloom development. For this study, however, Chl-a remains the most appropriate and practical parameter for long-term spatially explicit phenological analysis.
Reviewer Comment:
"Phytoplankton biomass in the ocean is not well described by its surface concentration, as phytoplankton occupy, throughout the year and in most environment the ML and depths below having sufficient light to grow in. When one looks only at the surface (and does not even integrate over the MLD), one will confuse a deep mixing event (which could either increase surface chl concentration if the chlorophyll max is entrained or decrease surface chl if the waters below are depleted of chlorophyll) that do not affect the integrated phytoplankton concentration as events that are ecologically important in terms of biomass."
Response:
We fully agree with the reviewer that surface chlorophyll concentration may not accurately represent the vertically integrated phytoplankton biomass, as well reviewed by Cullen (2015), and that this limitation can lead to potential misinterpretations. Some of us have reported differences in the vertical distribution of phytoplankton in the water column in the Argentine Sea (Lutz et al., 2010; Segura et al., 2021). However, we believe that our approach remains justified for several reasons, which we clarify and expand upon below.
Surface Chl-a remains the only consistently available satellite-based metric with sufficient spatiotemporal coverage to conduct multi-decadal phenological analysis at the regional scale addressed in our study. Integrated biomass from ocean color observations is not currently operationally available, and attempts to derive such products require assumptions (e.g., Chl vertical profiles, bio-optical models) that are not yet robust or validated across the Argentine Shelf.
While surface Chl-a does not directly represent total biomass, we do not assume it does in our interpretation. Rather, we use it as a practical index of surface-layer bloom dynamics and phenology, consistent with other remote sensing-based studies (e.g., Siegel et al., 2002; Racault et al., 2011; Ferreira et al., 2021). Our analysis is framed in terms of surface bloom anomalies and timing, not total water column productivity or standing stock.
That said, we acknowledge that surface Chl-a can sometimes be decoupled from vertical biomass structure, particularly during deep mixing events. To partially address this issue, our study includes not only instantaneous surface conditions (e.g., MLD, PAR, Zeu) at the time of the bloom peak, but also lagged environmental variables, specifically, 2- and 4-week averages of MLD and wind forcing. These lagged variables were included precisely to capture the influence of prior mixing history, which may affect whether subsurface biomass is entrained or diluted in the surface layer.
Regionally, field data from the Argentine Shelf show that surface Chl often tracks biomass variability reasonably well, especially during bloom periods. For instance, in the same region, it has been shown that primary production estimates based on uniform biomass profiles perform reliably despite the lack of vertical integration (Dogliotti et al., 2014), largely because the depth of significant production is limited by light and frequently shallower than the physical mixed layer.
Moreover, our phenology metrics (e.g., timing of bloom peak, anomalies) are calculated relative to long-term local baselines and are used to evaluate variability within site rather than across depths. This minimizes the impact of potential systematic offsets between surface and integrated biomass, especially since we do not attempt to interpret absolute production or biomass changes.
In conclusion, while we agree with the reviewer that vertically integrated biomass would be a more ecologically meaningful measure, the lack of reliable, long-term, vertically resolved datasets over the study domain precludes such an approach for the current analysis. However, by incorporating both contemporaneous and lagged physical predictors, we attempt to capture some of the vertical dynamics indirectly. We will strengthen the manuscript by clearly stating this limitation and by emphasizing that our conclusions apply to surface bloom behavior, with recognition that deeper biomass dynamics may diverge in certain regimes.
Reviewer Comment:
“Which bring us to the problem of the euphotic depth. Phytoplankton care about the absolute photon flux and recent evidence from the field suggest they can photosynthesize near their theoretical limit, which is more than three orders of magnitude of the value suggested by Sverdrup, 1953 (see Hoppe et al, 2024, Nature Communications, Rendelhoff et al., 2020 Science Advances, and discussion in Behrenfeld and Boss, 2017, which include lab studies). The 1% light level cannot act as a substitute to Zeu for this reason. While Zeu will be affected by factors affecting Kd (e.g. absorption by water constituents), it will not be affected by seasons and latitude, both of which clearly affect the amount of light in the upper ocean.”
Response:
We appreciate this insightful comment concerning the use of Zeu, traditionally defined as the depth where photosynthetically available radiation (PAR) is reduced to 1% of its surface value. We fully agree with the reviewer that phytoplankton can photosynthesize at irradiance levels well below this conventional threshold, as demonstrated in both laboratory and in situ studies. These results highlight that the 1% light level does not represent a true physiological limit for productivity, particularly for light-adapted phytoplankton populations.
In previous modeling work for this region (Dogliotti et al., 2014), we accounted for photosynthesis occurring below the 1% light level by integrating production over the full water column, acknowledging that productivity decreases gradually rather than ceasing abruptly at Zeu(1%). Although the present study does not address primary production, that experience informs our current use of Zeu(1%) as a conservative and pragmatic proxy for the depth of significant light penetration. Our goal here is to examine bloom phenology and its relationship with light–mixing dynamics, for which Zeu(1%) provides a consistent and interpretable satellite-derived index. While we recognize this threshold does not represent a physiological cutoff, it remains a widely used metric, particularly when in situ irradiance profiles or full light budgets are unavailable.
Importantly, we use Zeu in the context of the Zeu:MLD ratio, a diagnostic meant to represent the balance between light availability and vertical mixing. This ratio has been shown to be a useful indicator of bloom-conducive conditions in multiple studies (e.g., Siegel et al., 2002; Krug et al., 2018; Ferreira et al., 2021), despite the fact that photosynthesis continues below the 1% light level. From this perspective, our use of Zeu(1%) is not meant to imply a hard physiological cutoff, but rather to provide a practical, satellite-derived measure of how far light penetrates into the water column relative to the mixed layer, which in our study is derived from ocean reanalysis products.
Moreover, the 1% PAR level offers some accounting for vertical light limitation in a region such as the Southwestern Atlantic, where phytoplankton biomass is relatively high and the water column exhibits strong seasonal variability in both optical and physical properties. It is therefore reasonable to assume that the zone above Zeu captures the bulk of the biomass contributing to bloom events and that the Zeu:MLD ratio remains informative for interpreting seasonal changes in light–mixing dynamics.
In summary, we agree that Zeu(1%) does not reflect the minimum irradiance at which photosynthesis is possible, and we will clarify this point more explicitly in the revised manuscript. Nonetheless, it remains a consistent, conservative, and interpretable satellite-based index that serves our goal of characterizing relative bloom drivers over a 22-year time series.
Reviewer Comment:
"Phytoplankton in the ocean surface divide on a scale of a day (e.g. review by Ed Laws in ARMS) even in the most nutrient deplete regions of the ocean. However, most of this production is consumed on the same time scale. The small accumulation in time we designate as ‘the bloom’, occur on time scales of months and is due to a slight deviation from steady state as growth conditions of upper ocean phytoplankton keep improving (termed quasi steady state by Evans and Parslow). The ‘end’ of a bloom coincides with maximal growth rate rather than a collapse of the bloom. It is simply that loss rate catching up with the phytoplankton growth rate resulting in the stopping of accumulation. To constrain loss process from space, one can look at NPP (another satellite product). NPP/biomass=growth rate. Contrasting it with the accumulation rate (change of biomass in time) one can get an estimate of the loss rate (=growth rate-accumulation rate). Doing this you can get a sense of how ecological processes (controlling loss) may affect your observations."
Response:
We appreciate the reviewer’s insightful discussion of the dynamic balance between phytoplankton growth and loss processes in shaping the temporal evolution of surface blooms. We agree that the accumulation of biomass observed in surface chlorophyll time series reflects not just enhanced growth, but also a temporary mismatch between growth and loss rates, as described in quasi–steady state frameworks (e.g., Evans and Parslow, 1985). Furthermore, we acknowledge the conceptual value of contrasting net primary production (NPP) and biomass to infer ecological losses, as NPP/biomass provides an estimate of division rate, and the difference between division and accumulation can, in principle, yield a loss rate.
That said, implementing such an approach with the current satellite data products presents substantial challenges in our study region. Available global NPP products, such as those based on the Vertically Generalized Production Model (VGPM; Behrenfeld and Falkowski, 1997) and the Carbon-based Production Model (CbPM; Behrenfeld et al., 2005), rely on assumptions about photophysiology, nutrient limitation, and temperature dependencies that may not be appropriate for the highly dynamic and optically complex waters of the Southwestern Atlantic shelf. In particular, factors such as frontal systems, episodic nutrient inputs, and non-temperature-driven regulation of photosynthesis complicate the application of these global models in our region. Previous work (Dogliotti et al., 2014) showed that these satellite-based models do not perform well here when compared to in situ productivity measurements.
Moreover, deriving robust biomass estimates from space, especially in terms of phytoplankton carbon (C_phyto), also remains problematic due to the limited validation of backscattering-based algorithms (e.g., Graff et al., 2015), the scarcity of local in situ carbon data, and the difficulty of disaggregating contributions from detritus or other particles. This makes the denominator in the proposed NPP/biomass diagnostic potentially unreliable, compounding the uncertainty in any inferred growth or loss rate estimates.
Even if suitable NPP and biomass estimates were available, implementing a temporal differencing method to estimate accumulation (d[biomass]/dt) and infer loss rates would require high-quality, temporally smooth biomass time series, a level of fidelity not always achievable with currently available satellite products in shelf environments. While we recognize that this approach could yield deeper insight into the underlying ecological processes governing bloom evolution, we feel it falls beyond the scope of the present study.
Our aim was to explore the associations between bloom phenology (i.e., timing and magnitude anomalies) and physical drivers related to light and mixing. We do not attempt to close the biomass budget or disentangle physiological growth from ecological loss terms. However, we agree this is a promising avenue for future work, particularly as satellite algorithms and regional biogeochemical models continue to improve in accuracy and specificity for coastal and frontal systems. We will clarify in the revised manuscript that our analysis focuses on bloom variability observable at the surface and does not resolve underlying production or loss dynamics explicitly.
Reviewer Comment:
"The use of the 5% threshold above the median for bloom initiation is not well explained in the paper cited or here.How sensitive are your results to this value and how do your uncertainties in Chlorophyll affect the actual date of there 55thpercentile?"
Response::
We thank the reviewer for raising this important point. The use of a relative threshold, such as the 5% increase above the median chlorophyll-a (Chl-a) concentration, is indeed a methodological choice that warrants justification and transparency. Our approach follows a number of prior studies that have used similar percentile-based or threshold-based definitions to characterize bloom initiation and termination (e.g., Siegel et al., 2002; Racault et al., 2012; Ferreira et al., 2021; Delgado et al., 2023). The rationale is to detect significant increases in Chl-a relative to each pixel’s own climatological baseline, allowing the detection of phenological events that are locally meaningful and robust to regional gradients.
The choice of a 5% increase above the median was made to capture moderate but ecologically meaningful bloom events while limiting sensitivity to noise. This threshold has precedent in the literature, including Racault et al. (2012), and aligns with methods that define bloom onset based on anomalies relative to a climatological baseline. To test the robustness of this choice, we varied the threshold across a plausible range (3%, 5%, 7%) and found that the resulting spatial and interannual patterns in bloom initiation, peak, and termination remained consistent. The 3% and 7% values were selected to bracket the 5% threshold, representing slightly weaker and stronger bloom signals while avoiding spurious detection due to noise or transient variability. These findings support the plausibility of the selected threshold from a data-driven perspective.
Regarding uncertainty in the Chl-a product itself, we note that we use weekly composites from the CCI-v5.0 dataset, which blends observations from multiple satellite sensors using a consistent and validated processing chain. The use of weekly means reduces short-term retrieval noise and minimizes the impact of transient atmospheric artifacts. While instantaneous or daily satellite-derived Chl-a retrievals typically exhibit uncertainties in the range of ±30–40%, weekly-averaged products such as CCI-v5.0 generally show lower uncertainties, on the order of 15–25%, due to temporal averaging and multi-sensor blending (e.g., Sathyendranath et al., 2019).
In terms of their effect on the timing of bloom initiation, these residual uncertainties could, in principle, influence the exact timing at which the Chl-a time series first exceeds our threshold (median + 5%). However, because we work with weekly (8-day) composite data, the temporal resolution of our phenology metrics is inherently limited to ±4 days. In practice, bloom development in the Southwestern Atlantic shelf tends to occur gradually over several weeks, with Chl-a increasing steadily rather than fluctuating erratically near the threshold. This smooth progression means that threshold crossing typically occurs in a clear upward trend, making the estimated initiation date relatively insensitive to the expected level of retrieval uncertainty. In addition, because our bloom metrics are based on relative anomalies with respect to the local climatological median, rather than on absolute concentrations, systematic biases and residual inter-sensor differences in the blended product are minimized, enhancing the robustness of bloom onset detection. This approach helps ensure that bloom detection is driven by meaningful ecological departures rather than noise or retrieval error.
We will clarify in the revised manuscript both the rationale for the 5% threshold and the results of our sensitivity tests, and we will explicitly note that the threshold defines a relative increase in Chl-a above a site-specific baseline, not a universal bloom criterion. We agree with the reviewer that this kind of methodological clarity is important for reproducibility and interpretation of results.
Our responses above aim to explain the rationale for our methodological choices and acknowledge the trade-offs inherent in using satellite-derived surface chlorophyll to study large-scale phytoplankton dynamics. We recognize the ecological complexity involved, from vertical structure and photoacclimation to the balance of growth and loss, and have addressed each of the reviewer’s thoughtful concerns within the limits of the available data and tools.
While our approach simplifies some biological processes, it is grounded in precedent and shaped by the constraints of regional-to-decadal satellite and reanalysis data. Our goal is not to resolve every mechanistic detail but to identify robust patterns in bloom phenology and their links to key physical drivers, in line with established satellite-based studies.
Based on the reviewer’s constructive feedback, we will revise the manuscript to better articulate the limitations of our approach, strengthen the justification for our methods, and situate our findings more clearly within the broader context of global phytoplankton monitoring. We hope these clarifications convey the relevance of our work to understanding bloom dynamics in the complex and understudied Southwestern Atlantic shelf.
Citation: https://doi.org/10.5194/egusphere-2025-2033-AC1
-
AC1: 'Reply on RC1', Robert Frouin, 15 Jul 2025
-
RC2: 'Comment on egusphere-2025-2033', Anonymous Referee #2, 29 Jul 2025
Review of ‘On the Role of Light and Mixing in Shaping Southwestern Atlantic Shelf Blooms’ by Dogliotti et al.
This paper presents a study of phytoplankton phenology on the shelf of South America in the Southwestern Atlantic Ocean. Its primary focus is on the relationship between light penetration depth, mixed layer depth, and chlorophyll-a concentration, which serves as a proxy for phytoplankton biomass, building on Sverdrup's Critical Depth Hypothesis.
The study is well-structured, easy to follow, and a pleasure to read. I appreciate the extensive literature background and the solid introduction. While I do not see any significant flaws, I believe the paper could benefit from clarifying and explaining a few points.
1. As someone who lives and works in a very different part of the world, I would appreciate an additional figure illustrating the bathymetry, the main currents, and the extents of the Central and Patagonian Shelf. This would enhance the reader's understanding of the regional specifics, which was emphasized in the text.
2. I understand the challenges related to phenology metric estimation, as I have studied phenology in a different region. Defining the year is crucial, and this is challenging in your study area, as blooms begin at nearly opposite times of the year in the southern and northern ends. Did you observe any clear boundary in the locations where you transitioned from the conventional definition of the year to the June-May definition? Did this require any statistical “blending or merging” of the data?
3. In the modelling section, you mentioned that you analyzed values of the abiotic factors two and four weeks before bloom initiation to account for pre-bloom conditions. Did you explore time lag analysis, or were these time ranges adopted from previous studies? The rationale for the two- and four-week ranges was unclear.
4. I found it striking that there was only a minor reference to river outflow in the Río de la Plata area. Although shallow (<20 m) areas of extremely high turbidity were masked out, river runoff affects the surrounding waters in many ways. For example, the additional nutrients fuel primary productivity, also far from the river mouth; introduced detritus and chromophoric dissolved organic matter (CDOM) can limit the euphotic zone, and water stratification may be enhanced due to significant salinity differences. Interannual and seasonal changes in river discharge were not included in the modelling section, which may have been limited by data availability. However, I would expect more direct references in the discussion section, not just minor references to frontal zones.
5. One last comment is, why in line 44 did you refer to PAR as the photosynthetically available region, while even later in your manuscript (line 201) you defined PAR as the Photosynthetically Available Radiation. The latter, seems to me to be a more commonly used definition.
-
AC2: 'Reply on RC2', Robert Frouin, 04 Aug 2025
1. As someone who lives and works in a very different part of the world, I would appreciate an additional figure illustrating the bathymetry, the main currents, and the extents of the Central and Patagonian Shelf. This would enhance the reader's understanding of the regional specifics, which was emphasized in the text.
We appreciate the reviewer’s suggestion that would help improve the manuscript. We will include a map with the bathymetry, main currents and shelf regions in figure 1.
2. I understand the challenges related to phenology metric estimation, as I have studied phenology in a different region. Defining the year is crucial, and this is challenging in your study area, as blooms begin at nearly opposite times of the year in the southern and northern ends. Did you observe any clear boundary in the locations where you transitioned from the conventional definition of the year to the June-May definition? Did this require any statistical “blending or merging” of the data?
We agree with the reviewer that defining the year is important and challenging given the large latitudinal range the study area encompasses. We used a different definition of the year only for the analysis at the specific sites, but not for the climatological analysis over the whole study area given that it is not easy to set an objective way of defining which year to use on a pixel by pixel basis. We found that the conventional Jan-Dec year was more suitable to analyze the phytoplankton growth cycle in the northern part of the study area (north of 39°), like EP and C4 sites, while for the other five sites the Jun-May year suited better. Given that most of the study area is located in the southern region, the Jun-May definition was used for the climatological analysis. We will clarify this point more explicitly in the revised manuscript.
3. In the modelling section, you mentioned that you analyzed values of the abiotic factors two and four weeks before bloom initiation to account for pre-bloom conditions. Did you explore time lag analysis, or were these time ranges adopted from previous studies? The rationale for the two- and four-week ranges was unclear.
We thank the reviewer for raising this point. There have been prior studies that analyzed different physical forcings preconditioning the water column prior to surface bloom initiation using different time ranges in different regions of the world with different results but with no clear consensus on a given time frame. For example, Henson et al. (2006) suggested that wind conditions in the previous winter influenced the starting date of the spring bloom in the North Atlantic (Henson et al. 2006), whereas Krug et al. (2018) found regionally variable responses to seasonal wind and MLD anomalies in the SW Iberia region. In our study we selected 2 and 4-week averages of MLD and wind-speed previous to the bloom peak as pragmatic time windows for two main reasons. First, phytoplankton growth responses to physical drivers such as MLD shoaling or wind-driven nutrient input typically unfold over timescales of one to several weeks, given division rates and the time required for nutrient replenishment or turbulence-induced cell damage or export out of the euphotic zone. The chosen windows capture these short-term preconditioning effects without being too close to the peak or too long to dilute the influence of immediate pre-bloom forcings. Second, given the 8-day temporal resolution of both the satellite and reanalysis datasets, the shortest meaningful lag that can be calculated corresponds to two consecutive time steps (about 2 weeks). This is the finest temporal scale available to represent pre-bloom conditions while remaining sufficiently separated from contemporaneous measurements at the bloom peak, thereby avoiding redundancy among predictors in the regression model. The 4-week lag (four time steps) extends this window to capture slightly longer-term preconditioning effects. These fixed intervals ensure that lagged averages are defined consistently across all sites and years and provide a balance between capturing relevant pre-bloom forcing and retaining enough independent observations relative to the number of predictors in the analysis.
4. I found it striking that there was only a minor reference to river outflow in the Río de la Plata area. Although shallow (<20 m) areas of extremely high turbidity were masked out, river runoff affects the surrounding waters in many ways. For example, the additional nutrients fuel primary productivity, also far from the river mouth; introduced detritus and chromophoric dissolved organic matter (CDOM) can limit the euphotic zone, and water stratification may be enhanced due to significant salinity differences. Interannual and seasonal changes in river discharge were not included in the modelling section, which may have been limited by data availability. However, I would expect more direct references in the discussion section, not just minor references to frontal zones.
We agree that the Río de la Plata outflow has an important influence in its surrounding waters. However, this was not considered in the modeling section given that six out of seven of the selected sites are far from and are not clearly influenced by the Río de la Plata or any other big river discharge. Only at the EP (EPEA) site the influence of the Río de la Plata waters has been previously reported. On particular occasions, not every year, lower salinity values have been measured in summer when the river outflow reaches its maximum southern extension (Carreto et al., 1995). The present study focuses on the main bloom peak, and at EPEA it usually occurs in autumn-winter, therefore a strong influence of the river waters is not expected at this site. In the context of the climatological analysis, the influence of the Río de la Plata has not been addressed given that, even though the more turbid waters have been masked, it has been shown that satellite-derived Chl-a is overestimated in the area influenced by the Río de la Plata (Armstrong et al., 2004; Garcia et al., 2005, 2006) probably due to high detritus (non-algal particles) and yellow substances (CDOM). We will mention this in the revised manuscript.
5. One last comment is, why in line 44 did you refer to PAR as the photosynthetically available region, while even later in your manuscript (line 201) you defined PAR as the Photosynthetically Available Radiation. The latter, seems to me to be a more commonly used definition.
We thank the reviewer for pointing this out, it is a typo. In the revised manuscript it will be corrected.
Citation: https://doi.org/10.5194/egusphere-2025-2033-AC2
-
AC2: 'Reply on RC2', Robert Frouin, 04 Aug 2025
-
RC3: 'Comment on egusphere-2025-2033', Anonymous Referee #3, 29 Sep 2025
General comments
This paper presents a statistical analysis of phytoplankton bloom phenology along the Argentine Continental Shelf based on a 22-year satellite record of surface chlorophyll a concentration, and its relationships with light radiation and mixed layer depth (MLD) at 7 selected locations. Indeed, I agree with Reviewer 2 that there are no significant flaws regarding the technical part, as similar methodologies have been applied in many previous studies. Having said that, the weaknesses and limitations inherent to satellite-based studies on this topic are also well recognized in ocean color communities. Since Reviewer 1 has already elaborated on these, I will not repeat them here.For strengthening the paper’s contribution, I wonder if there are any in situ chlorophyll or ancillary observations available in the study area. Are there monitoring stations near the 7 selected sites that could provide complementary validation of satellite-derived chlorophyll? Even a simple comparison would enhance confidence in the results.
In addition, I found many imprecise uses of terminology, particularly for physical processes, which make the manuscript difficult to follow. Terms such as “mixing,” “stratification,” and “mixed layer depth” are often used interchangeably, though they represent distinct concepts (e.g., stratification is a water property, not a process). The manuscript also frequently mentions local dynamics such as tidal mixing and upwelling. However, these discussions are mostly descriptive in nature, without quantitative analysis, definition, or explanation. In the current version, they are mixed into the results and distract from the focus on MLD and light analysis. The authors could consider reorganizing these sentences and paragraphs for better clarity. The introduction is also quite loose. Overall, the manuscript would benefit from a substantial writing revision. I have tried to provide some indications in the specific comments, but I did not do it thoroughly.
To me, another concern is the removal of seasonality from chlorophyll data. Bloom characteristics are not inherently seasonal. Figure 5 is an example showing large variations in bloom timing and magnitude. As a result, it becomes unclear what component is removed, or whether this step simply reduces variance and artificially increases the variance explained by the multiple regression model. I recommend that the authors also present an analysis without seasonal removal for comparison.
I do not have an issue with the use of a threshold-based method for bloom detection. However, different phenological indices and methods have been studied and proposed in recent years. I think the authors need to conduct a more thorough literature review that includes recent studies on bloom phenology (e.g., Marchese et al., 2017; Zhao et al., 2022). At present, the introduction cites many older references while overlooking newer methods such as Gaussian fitting. The examples I list here are just two I am aware of.
Marchese, C., Albouy, C., Tremblay, J. É., Dumont, D., D’Ortenzio, F., Vissault, S., & Bélanger, S. (2017). Changes in phytoplankton bloom phenology over the North Water (NOW) polynya: a response to changing environmental conditions. Polar Biology, 40(9), 1721-1737.
Zhao, H., Matsuoka, A., Manizza, M., & Winter, A. (2022). Recent changes of phytoplankton bloom phenology in the northern high‐latitude oceans (2003–2020). Journal of Geophysical Research: Oceans, 127(10), e2021JC018346.Specific comments
Line 21: The meaning of “geographic trends” is unclear. Do the authors mean spatial distribution patterns?
Line 25: The attribution “due to deeper MLDs and colder Subantarctic waters” is not clearly supported. Please clarify.
Line 34: The term “mixing” is used frequently. Consider using “vertical mixing” for clarity. Also, please be cautious: mixed layer depth and mixing represent different concepts, and stratification is a property of water, not a process.
Line 37–39: This paragraph may need revision, particularly what n represents.
Line 87: The manuscript often mentions local processes. What is the purpose of highlighting them, given that they occur at different scales and are not the focus of analysis? Could they potentially conflict with the large-scale variables analyzed (e.g., MLD, PAR)? This requires clarification.
Line 285: Does “mean timing and intensity” refer to climatological averages over 22 years, or another metric? Please specify.
Line 296: The statement “deeper mixed layers require more time for warming and stratification” is not convincing as written. Either rephrase or provide more explanation.
Line 313–318: This paragraph reads more like a discussion. It would benefit from additional explanation, e.g., “frontal systems promote vertical mixing” or “a deeper mixed layer moves phytoplankton below the euphotic zone.
Line 325–326: This sentence is not clear to me. What does ‘the maximum frequency’ refer to?
Figure 4: Four variables are plotted with two y-axes. Were Chl a and PAR standardized with the other variables? Please explain the method clearly.Citation: https://doi.org/10.5194/egusphere-2025-2033-RC3 - AC3: 'Reply on RC3', Robert Frouin, 20 Oct 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-2033', Emmanuel Boss, 07 Jul 2025
Review of ‘On the Role of Light and Mixing in Shaping Southwestern Atlantic Shelf Blooms’ by Dogliotti et al.
Reviewer: Emmanuel Boss, UMaine
Full disclosure: I have met Ana Dogliotti multiple times (in Argentina as well) and am an active collaborator of R. Frouin. The topic of this paper is also close to my heart as I have worked on it from about 15yrs. Despite these potential conflicts of interest, I feel I can give this paper a fair read and provide useful comments for its improvement.
This paper sets out to study the phenology of phytoplankton bloom off Argetina and the influence of light (PAR, MLD) and mixing (via wind parameters) on it. The latter is done via statistical analysis and the derivation of multiple regression relating the latter parameters to the bloom initiation, bloom termination and bloom peak value.
I have some major comments that I feel need to be addressed for this paper to be a significant contribution. This paper ignores a whole body of literature questioning the approach they have espoused.
To start with the bloom of phytoplankton is analyzed from the perspective of surface chlorophyll concentration. There are major flaws with this approach:
- Chlorophyll is not phytoplankton biomass but a proxy (e.g. Cullen, 1985). There are other proxies accessible from remote-sensing that could allow one to evaluate the degree to which the observed chlorophyll is affected by photo-acclimation. The latter is a phenomena, well known from house plants, in which cells exposed to low light produce more chlorophyll per biomass compared to cells exposed to high light. Given the topic of the analysis done in this paper, looking at the influence of light on chlorophyll, accounting to this phenomenon is necessary (for example by looking also to C_phyto derived from particulate backscattering a product of remote sensing). In phytoplankton culture Carbon/chl ratios vary from 30-300, an order of magnitude.
- Phytoplantkon biomass in the ocean is not well described by its surface concentration, as phytoplankton occupy, throughout the year and in most environment the ML and depths below having sufficient light to grow in. When one looks only at the surface (and does not even integrate over the MLD), one will confuse a deep mixing event (which could either increase surface chl concentration if the chlorophyll max is entrained or decrease surface chl if the waters below are depleted of chlorophyll) that do not affect the integrated phytoplankton concentration as events that are ecologically important in terms of biomass.
- Which bring us to the problem of the euphotic depth. Phytoplankton care about the absolute photon flux and recent evidence from the field suggest they can photosynthesize near their theoretical limit, which is more than three orders of magnitude of the value suggested by Sverdrup, 1953 (see Hoppe et al, 2024, Nature Communication, Rendelhoff et al., 2020 Science advances, and discussion in Behrenfeld and Boss, 2017, which include lab studies. The 1% light level cannot act as a substitute to Zeu for this reason. While Zeu will be affected by factors affecting Kd (e.g. absorption by water constituents) it will not be affected by seasons and latitude, both of which clearly affect the amount of light in the upper ocean.
- Phytoplankton in the ocean surface divide on a scale of a day (e.g. review by Ed Laws in ARMS) even in the most nutrient deplete regions of the ocean. However, most of this production is consumed on the same time scale. The small accumulation in time we designate as ‘the bloom’, occur on time scales of months and is due to a slight deviation from steady state as growth conditions of upper ocean phytoplankton keep improving (termed quasi steady state by Evans and Parslow). The ‘end’ of a bloom coincides with maximal growth rate rather than a collapse of the bloom. It is simply that loss rate catching up with the phytoplankton growth rate resulting in the stopping of accumulation. To constrain loss process from space, one can look at NPP (another satellite product). NPP/biomass=growth rate. Contrasting it with the accumulation rate (change of biomass in time) one can get an estimate of the loss rate (=growth rate-accumulation rate). Doing this you can get a sense of how ecological processes (controlling loss) may affect your observations.
- The use of the 5% threshold above the median for bloom initiation is not well explained in the paper cited or here. How sensitive are your result to this value and how do your uncertainties in Chlorophyll affect the actual date of the 55th percentile?
Dear authors. I am often wrong. If you feel that my review is “off the mark” feel free to contact me and if convinced, I would be more than happy to change my review.
-
AC1: 'Reply on RC1', Robert Frouin, 15 Jul 2025
We thank the reviewer for the thoughtful and constructive review, which we believe has helped clarify the scope, strengths, and limitations of our study. We respond below to each major point, incorporating supporting evidence and contextual rationale for our methodological choices.
Reviewer Comment:
"Chlorophyll is not phytoplankton biomass but a proxy (e.g. Cullen, 1985). There are other proxies accessible from remote-sensing that could allow one to evaluate the degree to which the observed chlorophyll is affected by photo-acclimation. The latter is a phenomena, well known from house plants, in which cells exposed to low light produce more chlorophyll per biomass compared to cells exposed to high light. Given the topic of the analysis done in this paper, looking at the influence of light on chlorophyll, accounting to this phenomenon is necessary (for example by looking also to C_phyto derived from particulate backscattering a product of remote sensing). In phytoplankton culture Carbon/chl ratios vary from 30-300, an order of magnitude."
Response:
We agree with the reviewer that chlorophyll-a concentration is not a direct measure of phytoplankton carbon biomass and is subject to variability due to photoacclimation, species composition, and environmental conditions, as originally outlined by Cullen (1985) and in subsequent studies (e.g., Geider, 1987; Sathyendranath et al., 2009). Indeed, photoacclimation leads to significant changes in intracellular chlorophyll content, particularly in response to varying light regimes, which is highly relevant when studying the influence of light on phytoplankton dynamics.
While we fully recognize these limitations, our choice to use surface chlorophyll-a (Chl-a) as the primary indicator in this study was motivated primarily by its robustness and consistency as a satellite-based proxy available for long-term, large-scale studies of phytoplankton dynamics. This is particularly true for studies aiming to investigate phenological patterns over two decades across an extensive and heterogeneous shelf system such as the Southwestern Atlantic.
Alternative remote sensing products, such as particulate backscattering-derived phytoplankton carbon (C_phyto), have limited validation in our study region and insufficient temporal coverage for the 1998–2019 period. Furthermore, previous tests using the carbon-based productivity model (CbPM; Behrenfeld et al., 2005) in this region (Dogliotti et al., 2014) showed poor performance, further constraining the utility of satellite-derived C_phyto in our analysis.
Observed regional variability in the C:Chl ratio supports the reviewer’s point, and some of us have indeed investigated this in other studies. For example, C:Chl ratios ranged from 4.8 to 187.7 at the EPEA station on the northern Argentine coast (Silva et al., 2009) and were as low as 3–12 in the Magellan Strait (Lutz et al., 2016). These findings reflect the influence of both environmental gradients and species composition across this broad shelf ecosystem.
Given these constraints, our study does not attempt to quantify absolute biomass or productivity but rather focuses on analyzing the spatiotemporal variability of bloom intensity and timing based on Chl-a patterns. The phenological metrics we compute (e.g., bloom initiation, peak timing, bloom anomalies) are relative in nature and rely on internal thresholds rather than absolute biomass values, thereby minimizing the impact of photoacclimation-driven variability on our conclusions.
Nevertheless, we acknowledge that interpreting Chl-a variability solely as a reflection of biomass changes can be misleading without accounting for photoacclimation. We will explicitly clarify this limitation in the revised manuscript and acknowledge that Chl-a-based interpretations may, in some cases, reflect changes in intracellular pigment concentration rather than actual phytoplankton accumulation.
In future work, we agree that incorporating regionally validated C_phyto products or in situ C:Chl ratio observations would significantly enhance our understanding of the physiological and ecological controls on bloom development. For this study, however, Chl-a remains the most appropriate and practical parameter for long-term spatially explicit phenological analysis.
Reviewer Comment:
"Phytoplankton biomass in the ocean is not well described by its surface concentration, as phytoplankton occupy, throughout the year and in most environment the ML and depths below having sufficient light to grow in. When one looks only at the surface (and does not even integrate over the MLD), one will confuse a deep mixing event (which could either increase surface chl concentration if the chlorophyll max is entrained or decrease surface chl if the waters below are depleted of chlorophyll) that do not affect the integrated phytoplankton concentration as events that are ecologically important in terms of biomass."
Response:
We fully agree with the reviewer that surface chlorophyll concentration may not accurately represent the vertically integrated phytoplankton biomass, as well reviewed by Cullen (2015), and that this limitation can lead to potential misinterpretations. Some of us have reported differences in the vertical distribution of phytoplankton in the water column in the Argentine Sea (Lutz et al., 2010; Segura et al., 2021). However, we believe that our approach remains justified for several reasons, which we clarify and expand upon below.
Surface Chl-a remains the only consistently available satellite-based metric with sufficient spatiotemporal coverage to conduct multi-decadal phenological analysis at the regional scale addressed in our study. Integrated biomass from ocean color observations is not currently operationally available, and attempts to derive such products require assumptions (e.g., Chl vertical profiles, bio-optical models) that are not yet robust or validated across the Argentine Shelf.
While surface Chl-a does not directly represent total biomass, we do not assume it does in our interpretation. Rather, we use it as a practical index of surface-layer bloom dynamics and phenology, consistent with other remote sensing-based studies (e.g., Siegel et al., 2002; Racault et al., 2011; Ferreira et al., 2021). Our analysis is framed in terms of surface bloom anomalies and timing, not total water column productivity or standing stock.
That said, we acknowledge that surface Chl-a can sometimes be decoupled from vertical biomass structure, particularly during deep mixing events. To partially address this issue, our study includes not only instantaneous surface conditions (e.g., MLD, PAR, Zeu) at the time of the bloom peak, but also lagged environmental variables, specifically, 2- and 4-week averages of MLD and wind forcing. These lagged variables were included precisely to capture the influence of prior mixing history, which may affect whether subsurface biomass is entrained or diluted in the surface layer.
Regionally, field data from the Argentine Shelf show that surface Chl often tracks biomass variability reasonably well, especially during bloom periods. For instance, in the same region, it has been shown that primary production estimates based on uniform biomass profiles perform reliably despite the lack of vertical integration (Dogliotti et al., 2014), largely because the depth of significant production is limited by light and frequently shallower than the physical mixed layer.
Moreover, our phenology metrics (e.g., timing of bloom peak, anomalies) are calculated relative to long-term local baselines and are used to evaluate variability within site rather than across depths. This minimizes the impact of potential systematic offsets between surface and integrated biomass, especially since we do not attempt to interpret absolute production or biomass changes.
In conclusion, while we agree with the reviewer that vertically integrated biomass would be a more ecologically meaningful measure, the lack of reliable, long-term, vertically resolved datasets over the study domain precludes such an approach for the current analysis. However, by incorporating both contemporaneous and lagged physical predictors, we attempt to capture some of the vertical dynamics indirectly. We will strengthen the manuscript by clearly stating this limitation and by emphasizing that our conclusions apply to surface bloom behavior, with recognition that deeper biomass dynamics may diverge in certain regimes.
Reviewer Comment:
“Which bring us to the problem of the euphotic depth. Phytoplankton care about the absolute photon flux and recent evidence from the field suggest they can photosynthesize near their theoretical limit, which is more than three orders of magnitude of the value suggested by Sverdrup, 1953 (see Hoppe et al, 2024, Nature Communications, Rendelhoff et al., 2020 Science Advances, and discussion in Behrenfeld and Boss, 2017, which include lab studies). The 1% light level cannot act as a substitute to Zeu for this reason. While Zeu will be affected by factors affecting Kd (e.g. absorption by water constituents), it will not be affected by seasons and latitude, both of which clearly affect the amount of light in the upper ocean.”
Response:
We appreciate this insightful comment concerning the use of Zeu, traditionally defined as the depth where photosynthetically available radiation (PAR) is reduced to 1% of its surface value. We fully agree with the reviewer that phytoplankton can photosynthesize at irradiance levels well below this conventional threshold, as demonstrated in both laboratory and in situ studies. These results highlight that the 1% light level does not represent a true physiological limit for productivity, particularly for light-adapted phytoplankton populations.
In previous modeling work for this region (Dogliotti et al., 2014), we accounted for photosynthesis occurring below the 1% light level by integrating production over the full water column, acknowledging that productivity decreases gradually rather than ceasing abruptly at Zeu(1%). Although the present study does not address primary production, that experience informs our current use of Zeu(1%) as a conservative and pragmatic proxy for the depth of significant light penetration. Our goal here is to examine bloom phenology and its relationship with light–mixing dynamics, for which Zeu(1%) provides a consistent and interpretable satellite-derived index. While we recognize this threshold does not represent a physiological cutoff, it remains a widely used metric, particularly when in situ irradiance profiles or full light budgets are unavailable.
Importantly, we use Zeu in the context of the Zeu:MLD ratio, a diagnostic meant to represent the balance between light availability and vertical mixing. This ratio has been shown to be a useful indicator of bloom-conducive conditions in multiple studies (e.g., Siegel et al., 2002; Krug et al., 2018; Ferreira et al., 2021), despite the fact that photosynthesis continues below the 1% light level. From this perspective, our use of Zeu(1%) is not meant to imply a hard physiological cutoff, but rather to provide a practical, satellite-derived measure of how far light penetrates into the water column relative to the mixed layer, which in our study is derived from ocean reanalysis products.
Moreover, the 1% PAR level offers some accounting for vertical light limitation in a region such as the Southwestern Atlantic, where phytoplankton biomass is relatively high and the water column exhibits strong seasonal variability in both optical and physical properties. It is therefore reasonable to assume that the zone above Zeu captures the bulk of the biomass contributing to bloom events and that the Zeu:MLD ratio remains informative for interpreting seasonal changes in light–mixing dynamics.
In summary, we agree that Zeu(1%) does not reflect the minimum irradiance at which photosynthesis is possible, and we will clarify this point more explicitly in the revised manuscript. Nonetheless, it remains a consistent, conservative, and interpretable satellite-based index that serves our goal of characterizing relative bloom drivers over a 22-year time series.
Reviewer Comment:
"Phytoplankton in the ocean surface divide on a scale of a day (e.g. review by Ed Laws in ARMS) even in the most nutrient deplete regions of the ocean. However, most of this production is consumed on the same time scale. The small accumulation in time we designate as ‘the bloom’, occur on time scales of months and is due to a slight deviation from steady state as growth conditions of upper ocean phytoplankton keep improving (termed quasi steady state by Evans and Parslow). The ‘end’ of a bloom coincides with maximal growth rate rather than a collapse of the bloom. It is simply that loss rate catching up with the phytoplankton growth rate resulting in the stopping of accumulation. To constrain loss process from space, one can look at NPP (another satellite product). NPP/biomass=growth rate. Contrasting it with the accumulation rate (change of biomass in time) one can get an estimate of the loss rate (=growth rate-accumulation rate). Doing this you can get a sense of how ecological processes (controlling loss) may affect your observations."
Response:
We appreciate the reviewer’s insightful discussion of the dynamic balance between phytoplankton growth and loss processes in shaping the temporal evolution of surface blooms. We agree that the accumulation of biomass observed in surface chlorophyll time series reflects not just enhanced growth, but also a temporary mismatch between growth and loss rates, as described in quasi–steady state frameworks (e.g., Evans and Parslow, 1985). Furthermore, we acknowledge the conceptual value of contrasting net primary production (NPP) and biomass to infer ecological losses, as NPP/biomass provides an estimate of division rate, and the difference between division and accumulation can, in principle, yield a loss rate.
That said, implementing such an approach with the current satellite data products presents substantial challenges in our study region. Available global NPP products, such as those based on the Vertically Generalized Production Model (VGPM; Behrenfeld and Falkowski, 1997) and the Carbon-based Production Model (CbPM; Behrenfeld et al., 2005), rely on assumptions about photophysiology, nutrient limitation, and temperature dependencies that may not be appropriate for the highly dynamic and optically complex waters of the Southwestern Atlantic shelf. In particular, factors such as frontal systems, episodic nutrient inputs, and non-temperature-driven regulation of photosynthesis complicate the application of these global models in our region. Previous work (Dogliotti et al., 2014) showed that these satellite-based models do not perform well here when compared to in situ productivity measurements.
Moreover, deriving robust biomass estimates from space, especially in terms of phytoplankton carbon (C_phyto), also remains problematic due to the limited validation of backscattering-based algorithms (e.g., Graff et al., 2015), the scarcity of local in situ carbon data, and the difficulty of disaggregating contributions from detritus or other particles. This makes the denominator in the proposed NPP/biomass diagnostic potentially unreliable, compounding the uncertainty in any inferred growth or loss rate estimates.
Even if suitable NPP and biomass estimates were available, implementing a temporal differencing method to estimate accumulation (d[biomass]/dt) and infer loss rates would require high-quality, temporally smooth biomass time series, a level of fidelity not always achievable with currently available satellite products in shelf environments. While we recognize that this approach could yield deeper insight into the underlying ecological processes governing bloom evolution, we feel it falls beyond the scope of the present study.
Our aim was to explore the associations between bloom phenology (i.e., timing and magnitude anomalies) and physical drivers related to light and mixing. We do not attempt to close the biomass budget or disentangle physiological growth from ecological loss terms. However, we agree this is a promising avenue for future work, particularly as satellite algorithms and regional biogeochemical models continue to improve in accuracy and specificity for coastal and frontal systems. We will clarify in the revised manuscript that our analysis focuses on bloom variability observable at the surface and does not resolve underlying production or loss dynamics explicitly.
Reviewer Comment:
"The use of the 5% threshold above the median for bloom initiation is not well explained in the paper cited or here.How sensitive are your results to this value and how do your uncertainties in Chlorophyll affect the actual date of there 55thpercentile?"
Response::
We thank the reviewer for raising this important point. The use of a relative threshold, such as the 5% increase above the median chlorophyll-a (Chl-a) concentration, is indeed a methodological choice that warrants justification and transparency. Our approach follows a number of prior studies that have used similar percentile-based or threshold-based definitions to characterize bloom initiation and termination (e.g., Siegel et al., 2002; Racault et al., 2012; Ferreira et al., 2021; Delgado et al., 2023). The rationale is to detect significant increases in Chl-a relative to each pixel’s own climatological baseline, allowing the detection of phenological events that are locally meaningful and robust to regional gradients.
The choice of a 5% increase above the median was made to capture moderate but ecologically meaningful bloom events while limiting sensitivity to noise. This threshold has precedent in the literature, including Racault et al. (2012), and aligns with methods that define bloom onset based on anomalies relative to a climatological baseline. To test the robustness of this choice, we varied the threshold across a plausible range (3%, 5%, 7%) and found that the resulting spatial and interannual patterns in bloom initiation, peak, and termination remained consistent. The 3% and 7% values were selected to bracket the 5% threshold, representing slightly weaker and stronger bloom signals while avoiding spurious detection due to noise or transient variability. These findings support the plausibility of the selected threshold from a data-driven perspective.
Regarding uncertainty in the Chl-a product itself, we note that we use weekly composites from the CCI-v5.0 dataset, which blends observations from multiple satellite sensors using a consistent and validated processing chain. The use of weekly means reduces short-term retrieval noise and minimizes the impact of transient atmospheric artifacts. While instantaneous or daily satellite-derived Chl-a retrievals typically exhibit uncertainties in the range of ±30–40%, weekly-averaged products such as CCI-v5.0 generally show lower uncertainties, on the order of 15–25%, due to temporal averaging and multi-sensor blending (e.g., Sathyendranath et al., 2019).
In terms of their effect on the timing of bloom initiation, these residual uncertainties could, in principle, influence the exact timing at which the Chl-a time series first exceeds our threshold (median + 5%). However, because we work with weekly (8-day) composite data, the temporal resolution of our phenology metrics is inherently limited to ±4 days. In practice, bloom development in the Southwestern Atlantic shelf tends to occur gradually over several weeks, with Chl-a increasing steadily rather than fluctuating erratically near the threshold. This smooth progression means that threshold crossing typically occurs in a clear upward trend, making the estimated initiation date relatively insensitive to the expected level of retrieval uncertainty. In addition, because our bloom metrics are based on relative anomalies with respect to the local climatological median, rather than on absolute concentrations, systematic biases and residual inter-sensor differences in the blended product are minimized, enhancing the robustness of bloom onset detection. This approach helps ensure that bloom detection is driven by meaningful ecological departures rather than noise or retrieval error.
We will clarify in the revised manuscript both the rationale for the 5% threshold and the results of our sensitivity tests, and we will explicitly note that the threshold defines a relative increase in Chl-a above a site-specific baseline, not a universal bloom criterion. We agree with the reviewer that this kind of methodological clarity is important for reproducibility and interpretation of results.
Our responses above aim to explain the rationale for our methodological choices and acknowledge the trade-offs inherent in using satellite-derived surface chlorophyll to study large-scale phytoplankton dynamics. We recognize the ecological complexity involved, from vertical structure and photoacclimation to the balance of growth and loss, and have addressed each of the reviewer’s thoughtful concerns within the limits of the available data and tools.
While our approach simplifies some biological processes, it is grounded in precedent and shaped by the constraints of regional-to-decadal satellite and reanalysis data. Our goal is not to resolve every mechanistic detail but to identify robust patterns in bloom phenology and their links to key physical drivers, in line with established satellite-based studies.
Based on the reviewer’s constructive feedback, we will revise the manuscript to better articulate the limitations of our approach, strengthen the justification for our methods, and situate our findings more clearly within the broader context of global phytoplankton monitoring. We hope these clarifications convey the relevance of our work to understanding bloom dynamics in the complex and understudied Southwestern Atlantic shelf.
Citation: https://doi.org/10.5194/egusphere-2025-2033-AC1
-
RC2: 'Comment on egusphere-2025-2033', Anonymous Referee #2, 29 Jul 2025
Review of ‘On the Role of Light and Mixing in Shaping Southwestern Atlantic Shelf Blooms’ by Dogliotti et al.
This paper presents a study of phytoplankton phenology on the shelf of South America in the Southwestern Atlantic Ocean. Its primary focus is on the relationship between light penetration depth, mixed layer depth, and chlorophyll-a concentration, which serves as a proxy for phytoplankton biomass, building on Sverdrup's Critical Depth Hypothesis.
The study is well-structured, easy to follow, and a pleasure to read. I appreciate the extensive literature background and the solid introduction. While I do not see any significant flaws, I believe the paper could benefit from clarifying and explaining a few points.
1. As someone who lives and works in a very different part of the world, I would appreciate an additional figure illustrating the bathymetry, the main currents, and the extents of the Central and Patagonian Shelf. This would enhance the reader's understanding of the regional specifics, which was emphasized in the text.
2. I understand the challenges related to phenology metric estimation, as I have studied phenology in a different region. Defining the year is crucial, and this is challenging in your study area, as blooms begin at nearly opposite times of the year in the southern and northern ends. Did you observe any clear boundary in the locations where you transitioned from the conventional definition of the year to the June-May definition? Did this require any statistical “blending or merging” of the data?
3. In the modelling section, you mentioned that you analyzed values of the abiotic factors two and four weeks before bloom initiation to account for pre-bloom conditions. Did you explore time lag analysis, or were these time ranges adopted from previous studies? The rationale for the two- and four-week ranges was unclear.
4. I found it striking that there was only a minor reference to river outflow in the Río de la Plata area. Although shallow (<20 m) areas of extremely high turbidity were masked out, river runoff affects the surrounding waters in many ways. For example, the additional nutrients fuel primary productivity, also far from the river mouth; introduced detritus and chromophoric dissolved organic matter (CDOM) can limit the euphotic zone, and water stratification may be enhanced due to significant salinity differences. Interannual and seasonal changes in river discharge were not included in the modelling section, which may have been limited by data availability. However, I would expect more direct references in the discussion section, not just minor references to frontal zones.
5. One last comment is, why in line 44 did you refer to PAR as the photosynthetically available region, while even later in your manuscript (line 201) you defined PAR as the Photosynthetically Available Radiation. The latter, seems to me to be a more commonly used definition.
-
AC2: 'Reply on RC2', Robert Frouin, 04 Aug 2025
1. As someone who lives and works in a very different part of the world, I would appreciate an additional figure illustrating the bathymetry, the main currents, and the extents of the Central and Patagonian Shelf. This would enhance the reader's understanding of the regional specifics, which was emphasized in the text.
We appreciate the reviewer’s suggestion that would help improve the manuscript. We will include a map with the bathymetry, main currents and shelf regions in figure 1.
2. I understand the challenges related to phenology metric estimation, as I have studied phenology in a different region. Defining the year is crucial, and this is challenging in your study area, as blooms begin at nearly opposite times of the year in the southern and northern ends. Did you observe any clear boundary in the locations where you transitioned from the conventional definition of the year to the June-May definition? Did this require any statistical “blending or merging” of the data?
We agree with the reviewer that defining the year is important and challenging given the large latitudinal range the study area encompasses. We used a different definition of the year only for the analysis at the specific sites, but not for the climatological analysis over the whole study area given that it is not easy to set an objective way of defining which year to use on a pixel by pixel basis. We found that the conventional Jan-Dec year was more suitable to analyze the phytoplankton growth cycle in the northern part of the study area (north of 39°), like EP and C4 sites, while for the other five sites the Jun-May year suited better. Given that most of the study area is located in the southern region, the Jun-May definition was used for the climatological analysis. We will clarify this point more explicitly in the revised manuscript.
3. In the modelling section, you mentioned that you analyzed values of the abiotic factors two and four weeks before bloom initiation to account for pre-bloom conditions. Did you explore time lag analysis, or were these time ranges adopted from previous studies? The rationale for the two- and four-week ranges was unclear.
We thank the reviewer for raising this point. There have been prior studies that analyzed different physical forcings preconditioning the water column prior to surface bloom initiation using different time ranges in different regions of the world with different results but with no clear consensus on a given time frame. For example, Henson et al. (2006) suggested that wind conditions in the previous winter influenced the starting date of the spring bloom in the North Atlantic (Henson et al. 2006), whereas Krug et al. (2018) found regionally variable responses to seasonal wind and MLD anomalies in the SW Iberia region. In our study we selected 2 and 4-week averages of MLD and wind-speed previous to the bloom peak as pragmatic time windows for two main reasons. First, phytoplankton growth responses to physical drivers such as MLD shoaling or wind-driven nutrient input typically unfold over timescales of one to several weeks, given division rates and the time required for nutrient replenishment or turbulence-induced cell damage or export out of the euphotic zone. The chosen windows capture these short-term preconditioning effects without being too close to the peak or too long to dilute the influence of immediate pre-bloom forcings. Second, given the 8-day temporal resolution of both the satellite and reanalysis datasets, the shortest meaningful lag that can be calculated corresponds to two consecutive time steps (about 2 weeks). This is the finest temporal scale available to represent pre-bloom conditions while remaining sufficiently separated from contemporaneous measurements at the bloom peak, thereby avoiding redundancy among predictors in the regression model. The 4-week lag (four time steps) extends this window to capture slightly longer-term preconditioning effects. These fixed intervals ensure that lagged averages are defined consistently across all sites and years and provide a balance between capturing relevant pre-bloom forcing and retaining enough independent observations relative to the number of predictors in the analysis.
4. I found it striking that there was only a minor reference to river outflow in the Río de la Plata area. Although shallow (<20 m) areas of extremely high turbidity were masked out, river runoff affects the surrounding waters in many ways. For example, the additional nutrients fuel primary productivity, also far from the river mouth; introduced detritus and chromophoric dissolved organic matter (CDOM) can limit the euphotic zone, and water stratification may be enhanced due to significant salinity differences. Interannual and seasonal changes in river discharge were not included in the modelling section, which may have been limited by data availability. However, I would expect more direct references in the discussion section, not just minor references to frontal zones.
We agree that the Río de la Plata outflow has an important influence in its surrounding waters. However, this was not considered in the modeling section given that six out of seven of the selected sites are far from and are not clearly influenced by the Río de la Plata or any other big river discharge. Only at the EP (EPEA) site the influence of the Río de la Plata waters has been previously reported. On particular occasions, not every year, lower salinity values have been measured in summer when the river outflow reaches its maximum southern extension (Carreto et al., 1995). The present study focuses on the main bloom peak, and at EPEA it usually occurs in autumn-winter, therefore a strong influence of the river waters is not expected at this site. In the context of the climatological analysis, the influence of the Río de la Plata has not been addressed given that, even though the more turbid waters have been masked, it has been shown that satellite-derived Chl-a is overestimated in the area influenced by the Río de la Plata (Armstrong et al., 2004; Garcia et al., 2005, 2006) probably due to high detritus (non-algal particles) and yellow substances (CDOM). We will mention this in the revised manuscript.
5. One last comment is, why in line 44 did you refer to PAR as the photosynthetically available region, while even later in your manuscript (line 201) you defined PAR as the Photosynthetically Available Radiation. The latter, seems to me to be a more commonly used definition.
We thank the reviewer for pointing this out, it is a typo. In the revised manuscript it will be corrected.
Citation: https://doi.org/10.5194/egusphere-2025-2033-AC2
-
AC2: 'Reply on RC2', Robert Frouin, 04 Aug 2025
-
RC3: 'Comment on egusphere-2025-2033', Anonymous Referee #3, 29 Sep 2025
General comments
This paper presents a statistical analysis of phytoplankton bloom phenology along the Argentine Continental Shelf based on a 22-year satellite record of surface chlorophyll a concentration, and its relationships with light radiation and mixed layer depth (MLD) at 7 selected locations. Indeed, I agree with Reviewer 2 that there are no significant flaws regarding the technical part, as similar methodologies have been applied in many previous studies. Having said that, the weaknesses and limitations inherent to satellite-based studies on this topic are also well recognized in ocean color communities. Since Reviewer 1 has already elaborated on these, I will not repeat them here.For strengthening the paper’s contribution, I wonder if there are any in situ chlorophyll or ancillary observations available in the study area. Are there monitoring stations near the 7 selected sites that could provide complementary validation of satellite-derived chlorophyll? Even a simple comparison would enhance confidence in the results.
In addition, I found many imprecise uses of terminology, particularly for physical processes, which make the manuscript difficult to follow. Terms such as “mixing,” “stratification,” and “mixed layer depth” are often used interchangeably, though they represent distinct concepts (e.g., stratification is a water property, not a process). The manuscript also frequently mentions local dynamics such as tidal mixing and upwelling. However, these discussions are mostly descriptive in nature, without quantitative analysis, definition, or explanation. In the current version, they are mixed into the results and distract from the focus on MLD and light analysis. The authors could consider reorganizing these sentences and paragraphs for better clarity. The introduction is also quite loose. Overall, the manuscript would benefit from a substantial writing revision. I have tried to provide some indications in the specific comments, but I did not do it thoroughly.
To me, another concern is the removal of seasonality from chlorophyll data. Bloom characteristics are not inherently seasonal. Figure 5 is an example showing large variations in bloom timing and magnitude. As a result, it becomes unclear what component is removed, or whether this step simply reduces variance and artificially increases the variance explained by the multiple regression model. I recommend that the authors also present an analysis without seasonal removal for comparison.
I do not have an issue with the use of a threshold-based method for bloom detection. However, different phenological indices and methods have been studied and proposed in recent years. I think the authors need to conduct a more thorough literature review that includes recent studies on bloom phenology (e.g., Marchese et al., 2017; Zhao et al., 2022). At present, the introduction cites many older references while overlooking newer methods such as Gaussian fitting. The examples I list here are just two I am aware of.
Marchese, C., Albouy, C., Tremblay, J. É., Dumont, D., D’Ortenzio, F., Vissault, S., & Bélanger, S. (2017). Changes in phytoplankton bloom phenology over the North Water (NOW) polynya: a response to changing environmental conditions. Polar Biology, 40(9), 1721-1737.
Zhao, H., Matsuoka, A., Manizza, M., & Winter, A. (2022). Recent changes of phytoplankton bloom phenology in the northern high‐latitude oceans (2003–2020). Journal of Geophysical Research: Oceans, 127(10), e2021JC018346.Specific comments
Line 21: The meaning of “geographic trends” is unclear. Do the authors mean spatial distribution patterns?
Line 25: The attribution “due to deeper MLDs and colder Subantarctic waters” is not clearly supported. Please clarify.
Line 34: The term “mixing” is used frequently. Consider using “vertical mixing” for clarity. Also, please be cautious: mixed layer depth and mixing represent different concepts, and stratification is a property of water, not a process.
Line 37–39: This paragraph may need revision, particularly what n represents.
Line 87: The manuscript often mentions local processes. What is the purpose of highlighting them, given that they occur at different scales and are not the focus of analysis? Could they potentially conflict with the large-scale variables analyzed (e.g., MLD, PAR)? This requires clarification.
Line 285: Does “mean timing and intensity” refer to climatological averages over 22 years, or another metric? Please specify.
Line 296: The statement “deeper mixed layers require more time for warming and stratification” is not convincing as written. Either rephrase or provide more explanation.
Line 313–318: This paragraph reads more like a discussion. It would benefit from additional explanation, e.g., “frontal systems promote vertical mixing” or “a deeper mixed layer moves phytoplankton below the euphotic zone.
Line 325–326: This sentence is not clear to me. What does ‘the maximum frequency’ refer to?
Figure 4: Four variables are plotted with two y-axes. Were Chl a and PAR standardized with the other variables? Please explain the method clearly.Citation: https://doi.org/10.5194/egusphere-2025-2033-RC3 - AC3: 'Reply on RC3', Robert Frouin, 20 Oct 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 814 | 175 | 48 | 1,037 | 27 | 41 |
- HTML: 814
- PDF: 175
- XML: 48
- Total: 1,037
- BibTeX: 27
- EndNote: 41
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Ana I. Dogliotti
Reinaldo A. Maenza
Moira Luz Clara
Vivian A. Lutz
Robert Frouin
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2201 KB) - Metadata XML
Review of ‘On the Role of Light and Mixing in Shaping Southwestern Atlantic Shelf Blooms’ by Dogliotti et al.
Reviewer: Emmanuel Boss, UMaine
Full disclosure: I have met Ana Dogliotti multiple times (in Argentina as well) and am an active collaborator of R. Frouin. The topic of this paper is also close to my heart as I have worked on it from about 15yrs. Despite these potential conflicts of interest, I feel I can give this paper a fair read and provide useful comments for its improvement.
This paper sets out to study the phenology of phytoplankton bloom off Argetina and the influence of light (PAR, MLD) and mixing (via wind parameters) on it. The latter is done via statistical analysis and the derivation of multiple regression relating the latter parameters to the bloom initiation, bloom termination and bloom peak value.
I have some major comments that I feel need to be addressed for this paper to be a significant contribution. This paper ignores a whole body of literature questioning the approach they have espoused.
To start with the bloom of phytoplankton is analyzed from the perspective of surface chlorophyll concentration. There are major flaws with this approach:
Dear authors. I am often wrong. If you feel that my review is “off the mark” feel free to contact me and if convinced, I would be more than happy to change my review.