Corrections requested for egusphere-2025-6092 (Albright, Stevens, Wirth),
typeset manuscript version 2
==========================================================================

Summary: While verifying the analysis against the proofs, we found that the
fixed air-sea temperature difference had been derived by applying the
cool-skin adjustment with the wrong sign (yielding 1.3 K instead of 1.0 K);
correcting this changes Delta_a T to 1.0 K and requires replacing Figure 9
(updated file attached) and updating the derived statistics in the
abstract, in Sects. 3 and 5, and in the captions of Figs. 2, 8, and 9, as
itemized below (correlation coefficients are unaffected).

--------------------------------------------------------------------------
1. ABSTRACT (p. 1, ~lines 17-21)
--------------------------------------------------------------------------
(a) Replace: "mean bias 0.33 g kg-1 compared to observed"
    with:    "mean bias 0.43 g kg-1 compared to observed"

(b) Replace: "with low mean biases (-0.06 and -0.03 g kg-1 compared to
              observed)"
    with:    "with low mean biases (0.04 and 0.06 g kg-1 compared to
              observed)"

--------------------------------------------------------------------------
2. SECTION 3.1, error-propagation estimate (p. 4, ~lines 75-78)
--------------------------------------------------------------------------
(a) Replace: "For typical values W_a = 0.90, Delta_a T = 1.3 K, and
              chi ~ 1/16 K-1,"
    with:    "For typical values W_a = 0.90, Delta_a T = 1.0 K, and
              chi ~ 1/16 K-1,"

(b) In the displayed (unnumbered) equation that follows:
    Replace: "epsilon_q ~ 1.32 epsilon_h + 0.42 epsilon_T"
    with:    "epsilon_q ~ 1.26 epsilon_h + 0.36 epsilon_T"

--------------------------------------------------------------------------
3. FIGURE 2 CAPTION (p. 4)
--------------------------------------------------------------------------
Replace: "with a fixed Delta_a T = 1.3 K."
with:    "with a fixed Delta_a T = 1.0 K."

--------------------------------------------------------------------------
4. SECTION "Estimates of dW/dz and Delta_a T" (p. 7, ~lines 99-105)
--------------------------------------------------------------------------
Replace the two sentences:
   "Because the relevant temperature for surface fluxes is the
    sea-surface skin temperature, we take it to be 0.3 K cooler than the
    measured bulk temperature, yielding a representative offset of
    Delta_a T = 1.3 K. This fixed offset is applied to the sea-surface
    temperatures corresponding to each radiosonde launch."
with:
   "Because the relevant temperature for surface fluxes is the
    sea-surface skin temperature, which is approximately 0.2 K cooler
    than the measured bulk temperature (Fairall et al., 1996; Yan et
    al., 2024), we adopt a representative offset of Delta_a T = 1.0 K.
    In practice, the near-surface air temperature at each radiosonde
    launch is obtained by subtracting 1.2 K (the mean bulk-air
    difference referenced to z_a = 40 m) from the measured bulk
    sea-surface temperature. The retrieved q_a is not sensitive to the
    precise choice, changing by about 0.1 g kg-1 per 0.1 K in
    Delta_a T (Eq. 8)."

--------------------------------------------------------------------------
5. FIGURE 8 CAPTION (p. 9)
--------------------------------------------------------------------------
Replace: "(b) Distribution of temperature differences,
          Delta_a T = T_s - T_a, with vertical lines for the mean and
          median values."
with:    "(b) Distribution of differences between the measured (bulk)
          sea-surface and near-surface air temperatures, with vertical
          lines for the mean and median values."
(Reason: T_s is defined in Sect. 2 as the skin temperature, whereas the
histogram shows the bulk measurement.)

--------------------------------------------------------------------------
6. SECTION "Retrieval skill in q_a" (p. 9, ~lines 10-25)
--------------------------------------------------------------------------
Replace: "overestimates q_a by 0.33 g kg-1 (5th-95th percentile range:
          -0.74 g kg-1 to 1.8 g kg-1), with a median absolute error of
          0.47 g kg-1"
with:    "overestimates q_a by 0.43 g kg-1 (5th-95th percentile range:
          -0.65 g kg-1 to 1.9 g kg-1), with a median absolute error of
          0.52 g kg-1"

(The following sentence, "r = 0.54 overall, rising to r = 0.76 for bases
below 600 m", is unchanged.)

--------------------------------------------------------------------------
7. FIGURE 9 (p. 10): REPLACE FIGURE FILE AND UPDATE CAPTION
--------------------------------------------------------------------------
Please replace f09.pdf with the attached updated version (recomputed
with the corrected offset; font sizes harmonized; correlation values
removed from the panel legends -- they remain quoted in the caption).

Caption changes:
(a) Replace: "a median absolute error of 0.47 g kg-1 and a mean bias of
              0.33 g kg-1"
    with:    "a median absolute error of 0.52 g kg-1 and a mean bias of
              0.43 g kg-1"
(b) Replace: "(dashed blue, r = 0.64)"        [panel c]
    with:    "(dashed blue, r = 0.61)"
    (this also corrects a stale value; the correct value, 0.61, is the
    one quoted in the main text)
(c) Replace: "a median absolute error of 0.26 g kg-1 and a mean bias of
              -0.06 g kg-1"                   [panel d]
    with:    "a median absolute error of 0.31 g kg-1 and a mean bias of
              0.04 g kg-1"

--------------------------------------------------------------------------
8. TEXT FOLLOWING FIGURE 9 (p. 10, ~lines 1-5)
--------------------------------------------------------------------------
Replace: "with a small mean bias of -0.06 g kg-1 and a median absolute
          error of 0.26 g kg-1 on January 28, 2020. On February 2, 2020,
          the correlation is slightly lower (r = 0.57), with a mean bias
          of -0.03 g kg-1 and a median absolute error of 0.27 g kg-1."
with:    "with a small mean bias of 0.04 g kg-1 and a median absolute
          error of 0.31 g kg-1 on January 28, 2020. On February 2, 2020,
          the correlation is slightly lower (r = 0.57), with a mean bias
          of 0.06 g kg-1 and a median absolute error of 0.31 g kg-1."

--------------------------------------------------------------------------
ATTACHMENT
--------------------------------------------------------------------------
f09.pdf  -- updated Figure 9
