Comments on “Glass plate sampling efficiency for trace gases in the sea surface microlayer”

EGUsphere-2025-5361

 General Comments

This manuscript addresses an important methodological gap in sea surface microlayer (SML) research by quantifying the sampling efficiency of the glass plate technique for short-lived trace gases (DMS, isoprene, CS₂). The topic is highly relevant for SML biogeochemistry and air–sea exchange studies. The manuscript is generally well structured, and the authors make a commendable effort to analyse their data comprehensively.

However, there are several limitations that prevent the study from supporting some of its conclusions in the current form. The experimental configuration is constrained by practical conditions (homogeneous tank, no natural organic enrichment, oversaturated conditions) and does not fully represent the complexity of natural SMLs. The key assumption that SML and ULW have identical concentrations (C_SML = C_ULW) requires further clarification, especially under oversaturation and in surfactant treatments. Moreover, differences between experiments are confounded by changes in tank material, location, and artificial sea salt formulation. In addition to the major points above, several specific issues in the manuscript should also be addressed.

Specific Comments

1. The manuscript proposes to use sampling efficiencies of 0.13±0.01 (DMS/isoprene) and 0.12±0.01 (CS₂) as thresholds to identify cases of net production in the SML. However, this recommendation requires significant caution. The experimental setup represents a specific scenario: a stagnant, homogeneous system with trace gas oversaturation. The steep concentration gradient driven by this oversaturation likely promotes rapid outgassing during the sampling process, potentially resulting in a lower sampling efficiency than might occur under near-equilibrium field conditions. Conversely, the lack of wind-driven turbulence in the lab might overestimate efficiency compared with rough seas. Therefore, these “thresholds” should be framed as system-specific empirical values derived from a high-outgassing regime, rather than general operational criteria.
2. The central definition of sampling efficiency assumes C_SML = C_ULW so that any difference between SML and ULW samples reflects pure sampling artefacts. This assumption is particularly challenging under the experimental conditions of strong oversaturation and during surfactant treatment, where ongoing outgassing and the sampling sequence may introduce gradients not fully accounted for by the "well-mixed" test. It is therefore recommended that the Discussion more explicitly acknowledge this limitation, clarifying that the reported sampling efficiencies likely reflect a combination of true sampling artefacts and contemporaneous water-air exchange.
3. The decision to omit explicit calibration and rely only on PA ratios is reasonable in principle, provided the GC-MS response is linear with negligible intercept over the concentration range used. However, the manuscript does not show any supporting calibration data, yet draws conclusions that depend on distinguishing efficiencies around 0.12–0.13. Even small deviations from linearity or a non-zero intercept could affect the ratio PA__SML/PA__ULW, especially if SML and ULW concentrations differ by a factor of ~8–10. A brief demonstration of linearity (in Supplementary Material) or a more explicit discussion of this assumption would strengthen the quantitative claims.

Minor Comments

Page 1, line 18 – “in the field. . Instead,” → remove extra period.

Page 4, Line 121: Two types of artificial seawater (Tropic Marin® Pro-Reef Sea Salt vs. ordinary aquarium salt) are used. It would help to explain the rationale for these choices and how their compositions differ (e.g. trace metals, alkalinity, additives). Since sampling efficiency differs between experiments B and C, the role of salt composition should be discussed more concretely.

Page 6, Line 140: The text notes that a single glass plate dip yields 1.8–6.3 mL of medium, leading to substantial variability in sample volume per dip. Could you clarify what factors drive this large range in collected volume per dip? Understanding these drivers is important for assessing whether volume variability introduces bias to sampling efficiency.

Page 7, Line 190: The decision to use peak area ratios without GC-MS calibration is stated, but the assumption of a linear relationship between peak area and concentration is not validated. Including a brief note on how this assumption was tested or acknowledging its limitations would enhance transparency.

Page 13 Line 337: Correct figure reference formatting. Several places refer to “Fig. Figure X” (e.g. Figure 1, Figure 2, Figure 3, Figure 5, Figure 6, Figure 7), which is clearly a typesetting artefact.

Page 21, Lines 490–495: The total water temperature range is modest (ΔT ≈ 5.9 °C). You might add that this limited range likely contributes to the relatively weak apparent temperature effect in the raw data, despite statistical significance in the MLR.