We appreciate the reviewer’s time and their insightful assessment of this study. Their professional suggestions have been invaluable in improving the quality of our manuscript. Our responses to the specific points are provided below.

1. We sincerely thank the reviewer for raising this critical point regarding taxonomic assignment and validation.

1.1 Taxonomic Classification Methodology:

In the original manuscript, we omitted the detailed parameters for taxonomic annotation. We have now clarified this in the revised Methods section. Specifically, after quality filtering, OTUs with fewer than 10 reads or present in fewer than 3 samples were removed to minimize artifacts. The representative sequences of the remaining OTUs were taxonomically annotated using BLASTn (version 2.7.1) against the Protist Ribosomal Reference database (PR2) (Guillou et al., 2012). OTUs identified as Foraminifera but unable to be assigned to specific major orders were categorized as "Others".

1.2. Phylogenetic Analysis and Micro-diversity:

We fully agree that a phylogenetic analysis provides essential proof of the taxonomic affiliations.

Since our dataset contains nearly 2,000 OTUs, constructing a single tree with all sequences would result in visual clutter and make interpretation difficult. Therefore, to verify the classification and visualize micro-diversity as requested, we selected the top 50 most abundant OTUs to represent the community structure (comprising 20 Rotaliida, 20 Monothalamids, 5 Textulariida, and 5 Milioliida). We constructed a Maximum Likelihood tree (Supplementary Fig. S1) using these representative OTUs alongside corresponding reference sequences extracted from the PR2 database, rooted with Gromia oviformis as the outgroup.

The resulting tree confirms that our OTUs cluster consistently with known foraminiferal lineages, validating our taxonomic assignments. Furthermore, it reveals high micro-diversity within dominant groups, particularly in the Ammonia (Rotaliida) and Ovammia (Monothalamids) clades.

1. Thank you for your suggestions. We will add detailed morphological and eDNA data to the supplementary materials of the manuscript and include the numerical values for total abundance at the top of the bar chart in Figure 3A.

The responses to the specific comments are as follows:

line 58 (and line 75-76): Yes, it should be phrased as "13 different salinity levels." Thank you for pointing this out.

line 84: We used the Guillard f/2 medium for algal culture, and this will be added to the manuscript.

lines 108-110: In this study, the amplified target fragment was 380 bp in length. Paired-end sequencing with 250 bp reads was employed, resulting in a potential overlapping region of approximately 120 bp between forward and reverse reads. Thank you for this thoughtful suggestion. These technical details will be incorporated into the main text.

lines 111-114: OTUs with fewer than 10 reads or appearing in fewer than 3 samples were removed to eliminate potential contaminants. Representative sequences were extracted from the OTU table and taxonomically annotated against the PR2 database (Protist Ribosomal Reference Database, Guillou et al., 2012) using BLAST (version 2.7.1). OTUs not classified under Foraminifera (rank 3 in the PR2 database) were subsequently discarded. We will also include the detailed workflow in the corresponding section of the manuscript.

lines 115-116: We appreciate your feedback. The original description was indeed too brief, and we have revised the relevant section in the manuscript to provide a more detailed explanation. “The OTU table was used to quantify the species number and relative abundance of each foraminifera taxon under different salinity conditions, as well as the alpha diversity indices of the community. Spearman correlation analysis between each taxon (as well as community diversity) and salinity was performed using SPSS software. For data showing significant correlations, figures were generated in Origin with linear fitting equations, plotted using a 95% confidence interval.”

Figure 2: Yes, the samples were fixed and stained with Rose Bengal dye to differentiate living organisms. As mentioned in Section 2.2 (Materials and Methods), we have incorporated this point into the figure caption based on your suggestion to enhance reader comprehension.

Figure 5: Thank you for the opportunity to clarify Figure 5. We have revised the figure legend to address your questions directly.

The chord diagram visually compares the foraminiferal community composition revealed by eDNA (Molecular) versus morphological methods. Each outer semicircle represents a method’s dataset, independently normalized to 100% to show its internal taxonomic proportions. The inner colored arcs represent taxa; their width equals the sum of that taxon’s relative abundance from both methods, indicating its overall detection strength across the study. The connecting ribbons show how each taxon’s abundance is partitioned between the two methods—ribbon width at the connection point equals its relative abundance within that method.

You correctly interpreted the key patterns: eDNA was dominated by Monothalamiids, Rotaliida, and “Others,” with minimal signal from Textulariida and Milioliida, which were primarily detected morphologically. A wide ribbon connecting to a large outer segment (e.g., Rotaliida to Morphology) confirms that taxon as a major component of that method’s community.

line 171: In this study, OTUs annotated as Foraminifera yet unassigned at the Order level or lower were collectively categorized as "Others." As for the detailed explanation regarding the phylogenetic tree, it has been addressed in the response to the first point; please refer to that section.

line 204: Thank you for pointing out the missing reference. We have now added the appropriate citation to the manuscript.

In environments with low salt concentrations, there is a demonstrable decline in the feeding and metabolic efficiency of foraminifera (Nigam et al., 2006; Sujata et al., 2011). Conversely, the collective of foraminifera exhibits a higher degree of tolerance to physiological stress induced by elevated levels of salinity (Debenay, 1990; Gull et al., 2025).

Debenay, J.-P.: Recent foraminiferal assemblages and their distribution relative to environmental stress in the paralic environments of West Africa (Cape Timiris to Ebrie Lagoon), J. Foraminifer. Res., 20(3):267–282, https://doi.org/10.2113/gsjfr.20.3.267, 1990.

Gull, H. M., Tawabini, B. S., Amao, A. O., Prayudi, S. D., Ayranci, K., and Kaminski, M. A.: Benthic Foraminiferal Response to a Salinity Gradient in the Uqair Hypersaline Lagoonal System, Gulf Coast of Saudi Arabia, Micropaleontology, 71(3):261–279, https://doi.org/10.47894/mpal.71.3.04, 2025.

Nigam, R., Saraswat, R., and Kurtarkar Raikar, S.: Laboratory experiment to study the effect of salinity variations on benthic foraminiferal species – Pararotalia nipponica (Asano), J. Geol. Soc. India, 67:41–46, https://doi.org/10.1007/BF02709367, 2006.

Sujata, K. R., Nigam, R., Saraswat, L., and Rajeev, N.: Regeneration and abnormality in benthic foraminifera Rosalina leei: Implications in reconstructing past salinity changes, Riv. Ital. Paleontol. Stratigr., 117(2):309–318, https://doi.org/10.13130/2039-4942/5970, 2011.

line 228: Thank you for your suggestion. We have implemented it by annotating the abundance values directly above the corresponding bars in Figure 3A, which enhances clarity for readers. As shown, the data confirm that foraminiferal abundance is lower under low-salinity conditions compared to moderate- and high-salinity conditions.

line 230: We thank you for their careful observation regarding Figure 3B. You are absolutely correct that Figure 3B shows that the Milioliida group consists of only one species (Quinqueloculina seminula) and this species richness remains constant across salinities. However, the discussion in Line 230 refers specifically to the "relative abundance" (proportion of the total community) rather than species richness. As illustrated in Figure 3A, the relative abundance of Milioliida clearly increases with salinity (from ~0% at 0 psu to ~45% at 55 psu), which aligns with our discussion on biomineralization mechanisms favoring high-Mg calcite in hypersaline waters.

Changes in Manuscript:

To prevent future confusion, we have modified the sentence to explicitly cite Figure 3A and clarified the distinction between abundance and richness:

"At higher salinities, the formation of shells was found to be more favorable for Milioliida, thereby explaining the higher relative abundance observed at elevated salinities in classical morphological analyses (Fig. 3A)."

line 249: Thank you for this important clarification regarding terminology. We agree that eDNA sequencing primarily identifies Operational Taxonomic Units (OTUs) rather than biological species directly.

Changes in Manuscript: "Firstly, while morphological analysis in this experiment identified only 13 species, the eDNA method demonstrated considerable sensitivity, identifying 1,977 OTUs which were annotated to 103 species, thereby better reflecting the diversity within the foraminiferan group."

line 262: We agree with you that the diversity of Milioliida in the morphological dataset is very low (only one species). In the original text, we used the term "relatively monotypic" to describe this. To be more precise and address your comment, we have revised the sentence to explicitly acknowledge that this group consists of a single species, while emphasizing that its high relative abundance (which eDNA failed to capture) is the key insight provided by morphology.

Revised text:" In the eDNA dataset, the porcelaneous Milioliida group constituted a negligible proportion of the total reads; consequently, the community response to salinity gradients within this group could not be detected. In contrast, a statistical analysis of classical morphology reveals that (Fig. 3A/B), although the Milioliida assemblage is monospecific (represented only by Quinqueloculina seminula), its abundance is substantial, constituting a high proportion of the total biological count."

lines 263-265: Thank you for your suggestion. This statistical description is based on the results presented in Figures 3A/B. Detailed quantitative data on foraminiferal abundance will be provided in the Supplementary Materials. Please refer to the specific language revisions in the previous response (Line 262).

lines 277-280: Thank you for providing the references; they are greatly helpful for our manuscript and future research. We have incorporated the suggested content and made the following revisions:

To address this challenge, environmental RNA (eRNA) has emerged as a promising tool for identifying metabolically active communities, due to its rapid degradation outside living cells which better reflects contemporary biological activity (Chen et al., 2025; Qiao et al., 2025). Although eRNA methodology in foraminiferal research is still evolving and not yet standardized, recent studies—such as those revealing anaerobic metabolic adaptations in subseafloor foraminifera and the role of primary production in shaping benthic eukaryotic interactions—highlight its potential to uncover physiological and ecological responses inaccessible to traditional approaches. The continued development and integration of eRNA with controlled laboratory culture systems will help directly resolve how salinity fluctuations influence active foraminiferal community structure and function, thereby extending the findings of this study toward more mechanistic, ecophysiologically meaningful insights.

General comments: Thank you for your suggestion. We will carefully review the manuscript. The term "in situ" has been revised as advised.

We would like to sincerely thank Dr. Christopher Lowery for his positive evaluation of our study and his constructive comments. Our detailed responses to his specific suggestions are presented below.

1. We fully agree with you regarding the critical importance of data transparency and apologize for the omission of these datasets in the initial submission. To address this, we have now compiled the raw morphological count data and the eDNA OTU tables, which will be provided as Supplementary Tables in the revised manuscript. Furthermore, regarding the visualization of total abundance, we have modified Figure 3A to display the total abundance values directly above each bar. We would also like to clarify that since our statistical analyses focused primarily on community-level parameters or the relative abundance of major taxonomic groups, detailed species-level abundance was not visualized in the main figures; however, these specific data are now fully accessible in the Supplementary Tables for reference.
2. We appreciate your request for further details on our sampling and experimental setup, which we have now fully clarified in the revised Methods section. To obtain a representative initial community, surface sediments (top 0–2 cm) were collected from approximately 20 random locations within the intertidal zone of Qingdao Bay and pooled into a single composite sample. Upon returning to the laboratory, this composite sample was gently homogenized. Crucially, prior to culturing, the sediment was wet-sieved through a 300-mesh (~48 μm) silk screen using gentle seawater flow. This pre-treatment step was essential to remove fine clay particles and excess organic detritus, thereby preventing pore-water hypoxia and bacterial overgrowth (which could lead to sediment fouling) during the long-term static culture.

From this processed sediment, an initial subsample (~50 g) was preserved as the baseline (T0), and the remaining material was divided into 13 equal aliquots for the salinity gradient experiment. Each aliquot (approximately 50 g wet weight) was placed in a sterilized glass crystallizing dish (10 cm diameter) and covered with an equivalent volume of salinity-adjusted seawater. At the conclusion of the 10-week culture period, each sample was split, with approximately 20 g allocated for morphological analysis and 20 g reserved for eDNA extraction.

1. We appreciate your suggestion to deepen the discussion on the strengths of this combined approach. In the revised manuscript, we have expanded Section 4.3 to explicitly address the trade-offs and necessity of coupling these methods. We highlight that morphological analysis alone introduces a significant bias by failing to detect soft-shelled taxa; our eDNA results demonstrated that Monothalamiids constitute a substantial proportion of the community abundance, representing a critical ecological component that would be entirely overlooked in traditional surveys. Conversely, the eDNA method in our study yielded low sequence counts for the Milioliida group, obscuring their response patterns. Morphology was therefore essential to visualize their dynamics, revealing that Milioliida are sensitive to low salinity (maintaining < 20% relative abundance at 5–20 PSU) but exhibit remarkable tolerance to hypersaline conditions, with relative abundances remaining high (~40%) at 40–60 PSU and peaking at 55 PSU. Furthermore, we clarified the observation at 0 PSU, where the relative abundance of Milioliida appeared slightly higher than at 5 PSU; this is an artifact resulting from the sharp decline in the total community abundance under extreme freshwater stress, rather than a preference for freshwater.

Finally, our discussion now emphasizes that while morphology remains the standard for reconstructing fossilizable assemblages, eDNA provides indispensable insights into the "hidden" soft-bodied diversity. Only by integrating both datasets can we achieve a holistic understanding of benthic foraminiferal responses to environmental stress.

Below, we provide point-by-point responses to the reviewer's specific suggestions, organized by line number.

Line 10: We appreciate this correction. We concur that ‘and’ provides a better logical flow here and have modified the text accordingly.

Line 11: We thank the reviewer for this insightful suggestion. We fully agree that the proposed revision not only resolves the ambiguity regarding the term "their" but also better contextualizes the study by highlighting the importance of foraminifera in estuarine settings. We have adopted the phrasing suggested by the reviewer in the revised Abstract.

Changes in Manuscript: The sentence has been rephrased as follows:

"Foraminifera are commonly used to assess salinity changes in estuary settings, but quantitative experimental studies on their responses to salinity gradients are scarce."

Line 13 (and throughout): We thank you for pointing this out. We apologize for the oversight and have capitalized the acronym as "PSU" throughout the revised manuscript.

Line 15: Thank you for this suggestion to improve quantitative clarity. Since our conclusions are based on two independent datasets (morphological and molecular) that yield distinct numerical ranges, providing a single value in the Abstract could be misleading. To address this, we have:

Revised the Abstract to emphasize that the reported trends are consistent, significant, and quantitatively robust across both methods;

Added full quantitative results (e.g., percentage changes, regression coefficients) in the Results section (Section 3.1&3.2, Fig. 3, and Supplementary Table).

This approach maintains conciseness in the Abstract while providing complete numerical transparency in the main text.

Line 22: Thank you for the correction. This sentence in the manuscript will be revised to: "Benthic foraminifera were found to survive across a range of salinity levels."

Line 30-32: Thank you for your suggestion. We will add the relevant reference (Boltovskoy et al., 1991) to the manuscript.

Line 33: We thank you for pointing out this nomenclatural inconsistency. According to the World Foraminifera Database (WoRMS) and following Loeblich & Tappan (1987), Streblus Fischer de Waldheim, 1817 is an objective junior synonym of Ammonia Brünnich, 1771. We have therefore standardized the genus name to Ammonia throughout the manuscript. 

Line 49: We thank you for this insightful suggestion. As foreshadowing of our eDNA results—which reveal high relative abundances of Monothalamiids undetected by morphological analysis—we have revised the Introduction to explicitly mention that soft-bodied Monothalamous foraminifera, despite their modern-day ecological prevalence, are virtually absent from the fossil record due to their non‑mineralized tests. This modification strengthens the logical link between the methodological rationale and our key findings.

Line69: Thank you for your careful observation. The refractometer has a precision of 1 PSU, and the value of 0.1 PSU was obtained by visual estimation (interpolation) between scale markings. We have now added a clarification in the Methods section: (measured by handheld refractometer, precision ±1‰, interpolated to 0.1 PSU).

Figure 1: Thank you for this helpful suggestion. We have revised Figure 1 accordingly: (1) a panoramic photograph of the sampling site and a close-up image of the surface sediment have been added to the blank space within the map box; (2) the scale bar has been manually redrawn, enlarged, and bolded for better readability. We believe these modifications significantly improve the clarity and visual appeal of the figure. You can see the revised figure in supplementary materials of this Author Comment.

Line 76: Thank you for pointing this out. We agree that the term “gradients” was imprecise in this context. As noted by another reviewer, what we refer to are in fact discrete salinity levels (5, 10, 15, …, 60 PSU) established in individual culture dishes. We have therefore replaced all instances of “salinity gradients” with “salinity levels” throughout the manuscript to accurately reflect the experimental design.

Line 86: Thank you for your question. In this study, salinity was the primary variable under investigation, and pH was not actively adjusted. To maintain stable culture conditions, at least half of the overlying seawater—previously adjusted to the corresponding salinity level—was replaced every three days. During this period, pH was monitored but not manipulated.

Line 92: Thank you for your detailed and constructive questions regarding our foraminiferal processing methodology. We have now expanded the description in the Methods section to provide a complete and transparent workflow. Specifically:

Samples were fixed and stained for 48 hours using a mixture of 95% ethanol and 1 g/L Rose Bengal to distinguish living specimens at the time of collection. After staining, samples were dried at 50 °C for 12 hours, weighed, and wrapped in 300‑mesh sieve netting, then soaked and rinsed in seawater. Foraminifera were concentrated by flotation using carbon tetrachloride (CCl₄), followed by a second drying step. The residues were then dry‑sieved through 0.150 mm and 0.063 mm mesh screens and split into two size fractions: >0.150 mm and 0.063–0.150 mm (Fontanier et al., 2002; Jian et al., 1999). The 0.063-0.150 mm fraction, consisting mainly of juveniles and very small specimens, was not picked. All stained (i.e., living) individuals from the >0.150 mm fraction were hand‑picked under a stereomicroscope. Although all specimens were counted, only Rose Bengal‑stained individuals were included in the final community analyses.

To provide a baseline for comparison, the original intertidal sediment samples were processed identically. The pre‑experimental community composition is shown as “T0” in Figure 3. A/B. Complete description of this protocol has now been added to the revised Methods section.

We thank the reviewer again for helping us improve the clarity and reproducibility of our methodology.

Line 125-127: Thank you for this insightful question. The total abundance of foraminifera showed a clear increasing trend with rising salinity. To make this pattern more accessible to readers, we have now added the total abundance values directly above the corresponding bars in Figure 3A. This addition clearly illustrates the positive relationship between salinity and foraminiferal abundance, complementing the relative abundance data already presented.

Line 129: In our manuscript, “the total” refers to the total number of individuals for morphological data, and total number of reads for eDNA data. Accordingly, relative abundance is defined as the proportion of individuals (or reads) belonging to a specific taxonomic group relative to the total foraminiferal individuals (or total reads) in a given sample. We have now added these definitions explicitly in the Methods section to avoid any ambiguity. We appreciate your careful reading and helpful suggestion.

Line 161: Thank you for your interest. The eDNA analysis yielded a total of 103 annotated species; however, listing all species names in the main text would be excessively lengthy. Since our statistical analyses are primarily based on community-level parameters and relative abundances of major taxonomic groups, we have provided the complete OTU table, including all species annotations and read counts, in the Supplementary Materials (Table S2). Readers interested in the detailed species composition can refer to this file.

Figure 5: Thank you for pointing this out. The colored curves in Figure 5A represent individual samples from each salinity treatment, including replicates. We have now added a clear explanation in the figure legend to avoid any confusion.

Line 170-176: Thank you for raising this important methodological concern. You are absolutely correct that eDNA extracted from bulk sediments may include relic DNA from specimens that were alive prior to sample collection or even from dead individuals, which is a well-recognized limitation of sediment-based eDNA approaches. To minimize the impact of this legacy DNA and ensure comparability across treatments, we thoroughly homogenized the original sediment and split it equally into each culture dish prior to the start of the experiment. Therefore, all salinity treatments began with the same initial sediment pool and thus an identical baseline eDNA signal. Any differences observed at the end of the culture period can therefore be primarily attributed to differential responses of the foraminiferal community to the salinity treatments during the incubation.

We fully acknowledge that relic DNA may still introduce background noise and potentially dampen the detection of rapid community shifts. This may partly explain the discrepancy between morphological and eDNA data for certain groups, such as the Miliolida. In the morphological dataset (Fig. 3A), Miliolida showed a clear increase in relative abundance with increasing salinity, consistent with their known preference for hypersaline conditions. In contrast, their eDNA signal was consistently low across most salinity treatments, making trends difficult to discern (Fig. 3C). Nevertheless, a closer examination of the eDNA data reveals that the relative abundance of Miliolida did increase from 0.29% at 0 PSU to 1.23% at 60 PSU, indicating that the same directional trend—though much weaker—is still detectable. The attenuated signal may reflect lower DNA extraction efficiency, faster DNA degradation, or lower ribosomal copy numbers in this group, compounded by the masking effect of relic DNA from other taxa.

We have now expanded the Discussion section to explicitly address this limitation, its potential impact on our interpretations, and the methodological steps we took to ensure treatment comparability. These additions are integrated with the detailed sediment processing protocol already included in the Methods in response to your earlier comments.

We thank the reviewer again for pushing us to critically evaluate the interpretation of eDNA signals in experimental sediment systems.

Line 197-201: We sincerely thank you for this insightful comment and for directing us to the important work of Goetz et al. (2025). We agree that the extensive cryptic diversity within the genus Ammonia is a critical context often missing in discussions of its ecological tolerance. As suggested, we have now incorporated this perspective into the Discussion. The revised text acknowledges that the broad salinity ranges reported for Ammonia tepida and A. beccarii in the literature may represent composite signals from multiple cryptic species, rather than the plasticity of a single biological entity.

Regarding the reviewer's suggestion to compare our DNA results with existing genetic data: while we fully agree this is a fascinating and important direction, we believe a rigorous, species-level phylogenetic comparison is beyond the scope of the current manuscript. Our eDNA metabarcoding approach was designed to assess community-level diversity patterns across salinity gradients, and the current dataset—while robust for this purpose—does not provide the single-species resolution required for detailed taxonomic revision. We have therefore framed this as a key future research priority in the revised Discussion. We are grateful for this suggestion and believe it significantly strengthens our manuscript by clearly defining its scope and acknowledging its limitations.

Line 205-207: We thank you for this insightful comment and for referencing Poag’s classic work on the Gulf of Mexico. We fully agree with the reviewer and Poag (1981) that in hypersaline environments (e.g., Laguna Madre), diversity often declines due to physiological stress, similar to what is observed in low-salinity estuaries. The relationship between salinity and diversity is generally non-linear (hump-shaped) across a full global spectrum.

In our study, the "positive correlation" (Margalef index vs. Salinity) is primarily driven by the steep increase in diversity from hyposaline (0–20 PSU) to marine/hypersaline (35–60 PSU) conditions. At 0–20 PSU: Strong osmotic stress severely limits survival, resulting in extremely low diversity. At 35–60 PSU: While this is hypersaline, for the specific taxa in Qingdao Bay (e.g., opportunistic Quinqueloculina), this range appears to be within their tolerance limit or even favorable for biomineralization, thus sustaining a relatively higher diversity compared to the lethal freshwater end. However, we acknowledge that if salinity increased further (e.g., >70 PSU), diversity would likely drop, consistent with Poag's model.

Regarding the wild populations in Qingdao Bay, Lei et al. (2017) conducted a thorough field survey in this specific intertidal zone. They reported that salinity (ranging from 31 to 38 PSU in their study) was positively correlated with species diversity, as freshwater input from rainfall/runoff acts as the primary disturbance. Our experimental results validate this local field trend but extend the understanding to extreme limits.

Line 214: We appreciate your attention to the terminology. We agree that the term "primitive" carries normative connotations that are inappropriate in an evolutionary context. Our intention was to refer to the fact that Miliolida represents a lineage that diverged early in foraminiferal evolution and retains specific ancestral calcification traits (high-Mg calcite), as discussed in Bentov et al. (2006).

We have replaced "evolutionarily primitive" with "representing an early-diverging lineage" to be more scientifically precise and neutral.

Line 227: We agree with the reviewer that the phrase "it can be deduced" was imprecise, given that we have direct count data supporting this observation.

As mentioned in our response to your previous comment regarding data availability, we have now included the raw count data in Supplementary Table S1, and explicitly labeled the total abundance values above the bars in Figure 3A.

We have rephrased the sentence to be direct and fact-based, removing the speculative tone. The sentence now reads: "Our quantitative data directly show that, in conditions of low salinity, there was a sharp decline in the abundance of foraminifera."

Line 228: We thank the reviewer for this critical observation. We acknowledge that we did not perform specific taphonomy analyses to quantify dissolution features (etching or breakage) on the tests in this study. The statement was intended as a mechanistic inference to explain the observed sharp decline in Milioliida and the relative dominance of Rotaliida under low salinity. This inference is based on well-established geochemical principles: the high-Mg calcite skeletons of Milioliida are chemically more soluble in hyposaline (often lower saturation state) waters compared to the low-Mg calcite of Rotaliida (e.g., Ammonia), as documented by previous studies (e.g., various citations in the previous paragraph). We have revised the manuscript: “Among these, the high-magnesium calcite shells of porcelaneous Miliolida were prone to dissolution, whereas the low-magnesium calcite shells of hyaline Rotaliida resisted dissolution more readily,”

Line 230-231: We appreciate this insightful question, which highlights the fundamental differences between morphological and molecular signals in our long-term experiment. It is important to note that although the starting stocks were identical, the 10-week culture period allowed for significant biological turnover and community divergence driven by physiological selection. In low-salinity conditions, Milioliida likely experienced high mortality and population collapse due to osmotic stress—in addition to test dissolution—which would result in the degradation of their DNA over time, rather than its preservation.

Furthermore, the consistently low eDNA reads for Milioliida, even in high-salinity groups where morphological analysis confirmed their thriving populations, indicate a systematic methodological bias (e.g., lower PCR amplification efficiency or extraction challenges for porcelaneous taxa). Consequently, the discrepancy between datasets is not because dissolution obscured the DNA signal, but rather because eDNA methods failed to capture Milioliida effectively due to primer bias, while morphological counts accurately reflected their high abundance in favorable hypersaline conditions. This distinction underscores the critical necessity of coupling morphology with eDNA to correct for such molecular "blind spots."

Line 238-239: Thank you for this helpful suggestion. We agree that acknowledging previous successful applications of benthic foraminifera as paleosalinity indicators provides essential context and support for our statement.

We have added key references to the revised manuscript, including Murray (2006) for general ecological distribution patterns and Strachan et al. (2015) for quantitative transfer functions. The revised sentence now reads: "Therefore, after controlling for temperature effects, foraminifera also serve as amplifiers of salinity variations, enabling salinity determination based on specific species assemblages (Murray, 2006; Strachan et al., 2015)."

Line 240-241: We thank you for pointing out these foundational references. We fully acknowledge that the impact of salinity on Mg/Ca-derived temperatures is a well-established phenomenon. We have revised the sentence to explicitly cite Lea et al. (1999) and the review by Katz et al. (2010). We have also adjusted the tone to state that our findings reinforce the importance of this established practice, rather than presenting it as a novel recommendation.

"Given that salinity influences Mg/Ca-based thermometry (e.g., Lea et al., 1999; Katz et al., 2010), our results further underscore the necessity of conducting foraminiferal composition analyses to constrain salinity backgrounds, thereby enhancing the accuracy of paleo-oceanic temperature reconstructions."

Line 247: Thank you for pointing this out. We have cited Pawlowski et al. (2014) in the revised manuscript to substantiate the statement regarding the power of molecular methods in describing diversity.

Line 250: We appreciate your positive assessment of this finding. We fully agree that the dominance of Monothalamiids is a critical result that reveals significant gaps in our understanding of benthic ecosystems when relying solely on traditional methods. In the revised Section 4.3, we have expanded the discussion to explicitly address these implications. Specifically, we argue that overlooking this group leads to a severe underestimation of total benthic standing stock and biodiversity, particularly in stressed environments where they appear to replace calcified taxa as the dominant guild.

Furthermore, we discuss how this omission can lead to a misinterpretation of ecosystem resilience. While morphological data showing a decline in calcified taxa might be interpreted as a "system collapse" under salinity stress, our molecular data reveal a functional shift towards a soft-bodied community rather than a total loss of biological activity. We also note that ignoring this "hidden majority" limits our understanding of energy flow in the benthic food web, as soft-bodied taxa likely exhibit distinct turnover rates and trophic roles compared to their shelled counterparts.

Line 286: Thank you for pointing out this ambiguity. By "functional relationships," we intended to refer to the quantitative linear regression models we established between salinity and specific biotic parameters (Rotaliida abundance and Margalef index).

We agree that the original phrasing was vague and potentially confusing (as "functional" can also imply ecological functioning). We have rephrased the sentence to be precise: "...Specifically, robust linear regression models were established linking salinity to both the relative abundance of Rotaliida and community diversity."

Line 289-291: We appreciate you for raising this important point regarding the applicability of our regression models. We fully agree that estuarine environments are complex and that local controls (e.g., temperature, food availability, distinct genotypes) can influence foraminiferal responses. We do not intend to claim that the exact linear coefficients derived from Qingdao Bay are universally applicable to all estuaries globally without local validation.

Instead, we propose that: The quantitative regression models serve as a robust, locally-calibrated proxy specifically for the Yellow Sea and Bohai Sea regions, where such experimental data are scarce. The biological trends (e.g., the positive correlation of diversity with salinity; the distinct tolerance limits of Rotaliida vs. Milioliida) provide a general reference framework for understanding physiological constraints in similar temperate marginal marine systems. We have revised the Conclusion to specify that these relationships are most applicable to "marginal marine environments, particularly in the Yellow Sea region." We have also added a sentence in the Discussion (Section 4.1) acknowledging the need for regional calibration when applying these proxies broadly.