**Figure 1.**
(**a**) Bathymetry and the sea surface temperature climatology in the COMS/MI full-disk area in (**b**) February and (**c**) August.

**Figure 1.**
(**a**) Bathymetry and the sea surface temperature climatology in the COMS/MI full-disk area in (**b**) February and (**c**) August.

**Figure 2.**
An example of time series of in-situ drifter temperatures (°C) (**a**) before and (**b**) after QC procedure to eliminate abnormal temperatures in 2011. (**c**,**d**) present another example of the QC procedure.

**Figure 2.**
An example of time series of in-situ drifter temperatures (°C) (**a**) before and (**b**) after QC procedure to eliminate abnormal temperatures in 2011. (**c**,**d**) present another example of the QC procedure.

**Figure 3.**
(**a**) Spatial distribution of the number of matchup points between COMS and in situ measurements in bins of 2° × 2° and the histograms of the matchups with respect to (**b**) month, (**c**) local time (hour), (**d**) latitude (°N), and (**e**) in situ temperature (°C) for the period from April 2011 to March 2015.

**Figure 3.**
(**a**) Spatial distribution of the number of matchup points between COMS and in situ measurements in bins of 2° × 2° and the histograms of the matchups with respect to (**b**) month, (**c**) local time (hour), (**d**) latitude (°N), and (**e**) in situ temperature (°C) for the period from April 2011 to March 2015.

**Figure 4.**
Distributions of NLSST errors of OSI SAF during the (**a**) daytime and (**b**) nighttime, MCSST errors during the (**c**) daytime and (**d**) nighttime, and NLSST errors in the (**e**) daytime and (**f**) nighttime as a function of the in situ SST (°C). The color represents the percentage of the data to the total number of matchup points in each bin; error bars represent the standard deviation of SST errors of each interval.

**Figure 4.**
Distributions of NLSST errors of OSI SAF during the (**a**) daytime and (**b**) nighttime, MCSST errors during the (**c**) daytime and (**d**) nighttime, and NLSST errors in the (**e**) daytime and (**f**) nighttime as a function of the in situ SST (°C). The color represents the percentage of the data to the total number of matchup points in each bin; error bars represent the standard deviation of SST errors of each interval.

**Figure 5.**
Comparison between MCSSTs and in situ temperatures during the (**a**) daytime and (**b**) nighttime and between NLSSTs and in situ temperatures at (**c**) daytime and (**d**) nighttime, where the color represents the percentage of the data to the total number of matchup points in a bin of 0.5 × 0.5 °C. Bias, root mean square error (RMSE), scatter index (SI), and correlation coefficient (R) are given in each plot.

**Figure 5.**
Comparison between MCSSTs and in situ temperatures during the (**a**) daytime and (**b**) nighttime and between NLSSTs and in situ temperatures at (**c**) daytime and (**d**) nighttime, where the color represents the percentage of the data to the total number of matchup points in a bin of 0.5 × 0.5 °C. Bias, root mean square error (RMSE), scatter index (SI), and correlation coefficient (R) are given in each plot.

**Figure 6.**
Spatial distribution of mean differences between MCSSTs and in situ temperatures for the (**a**) daytime and (**b**) nighttime and between NLSSTs and in situ temperatures in the (**c**) daytime and (**d**) nighttime in bins of 5° × 5° for the period from April 2011 to March 2014.

**Figure 6.**
Spatial distribution of mean differences between MCSSTs and in situ temperatures for the (**a**) daytime and (**b**) nighttime and between NLSSTs and in situ temperatures in the (**c**) daytime and (**d**) nighttime in bins of 5° × 5° for the period from April 2011 to March 2014.

**Figure 7.**
Distributions of SST residuals (satellite SST—in situ SST) of MCSST during the (**a**) daytime and (**b**) nighttime and NLSST during the (**c**) daytime and (**d**) nighttime as a function of wind speed (m s^{−1}). The color represents the percentage of the data to the total number of matchup points in each bin; error bars represent the standard deviation of SST errors of each interval.

**Figure 7.**
Distributions of SST residuals (satellite SST—in situ SST) of MCSST during the (**a**) daytime and (**b**) nighttime and NLSST during the (**c**) daytime and (**d**) nighttime as a function of wind speed (m s^{−1}). The color represents the percentage of the data to the total number of matchup points in each bin; error bars represent the standard deviation of SST errors of each interval.

**Figure 8.**
Distributions of MCSST errors during the (**a**) daytime and (**b**) nighttime and NLSST errors during the (**c**) daytime and (**d**) nighttime as a function of satellite zenith angle (SZA) in range of 0°–60° and mean bias and standard deviation of SST errors (blue and green) of each interval during the (**e**) daytime and (**f**) nighttime with respect to SZA in range of 20°–80°, where the red bars represent SST errors in case of SZA correction term $(\mathit{sec}\left(\theta \right)-1$) included in the SST equations.

**Figure 8.**
Distributions of MCSST errors during the (**a**) daytime and (**b**) nighttime and NLSST errors during the (**c**) daytime and (**d**) nighttime as a function of satellite zenith angle (SZA) in range of 0°–60° and mean bias and standard deviation of SST errors (blue and green) of each interval during the (**e**) daytime and (**f**) nighttime with respect to SZA in range of 20°–80°, where the red bars represent SST errors in case of SZA correction term $(\mathit{sec}\left(\theta \right)-1$) included in the SST equations.

**Figure 9.**
Distributions of MCSST errors during the (**a**) daytime and (**b**) nighttime and NLSST errors during the (**c**) daytime and (**d**) nighttime as a function of the latitude (°N).

**Figure 9.**
Distributions of MCSST errors during the (**a**) daytime and (**b**) nighttime and NLSST errors during the (**c**) daytime and (**d**) nighttime as a function of the latitude (°N).

**Figure 10.**
A plot of averaged MCSST errors during the (**a**) daytime and (**b**) nighttime and NLSST errors during the (**c**) daytime and (**d**) nighttime as a function of distance from the coast (km). The error bars indicate the standard deviation of SST errors in each interval.

**Figure 10.**
A plot of averaged MCSST errors during the (**a**) daytime and (**b**) nighttime and NLSST errors during the (**c**) daytime and (**d**) nighttime as a function of distance from the coast (km). The error bars indicate the standard deviation of SST errors in each interval.

**Figure 11.**
Comparison between MCSSTs and in situ temperatures during the (**a**) daytime and (**b**) nighttime and between NLSSTs and in situ temperatures during the (**c**) daytime and (**d**) nighttime for the period from April 2014 to March 2015, where the color represents the percentage of the data to the total number of matchup points in a bin of 0.5 °C × 0.5 °C. Bias, root mean square error (RMSE), scatter index (SI), and correlation coefficient (R) are given in each plot.

**Figure 11.**
Comparison between MCSSTs and in situ temperatures during the (**a**) daytime and (**b**) nighttime and between NLSSTs and in situ temperatures during the (**c**) daytime and (**d**) nighttime for the period from April 2014 to March 2015, where the color represents the percentage of the data to the total number of matchup points in a bin of 0.5 °C × 0.5 °C. Bias, root mean square error (RMSE), scatter index (SI), and correlation coefficient (R) are given in each plot.

**Figure 12.**
Spatial distribution of (**a**) root mean square errors (RMSEs) and (**b**) mean biases of MCSSTs and (**c**) RMSEs and (**d**) mean biases of NLSST from in situ temperatures in bins of 5° × 5° for the period from April 2014 to March 2015.

**Figure 12.**
Spatial distribution of (**a**) root mean square errors (RMSEs) and (**b**) mean biases of MCSSTs and (**c**) RMSEs and (**d**) mean biases of NLSST from in situ temperatures in bins of 5° × 5° for the period from April 2014 to March 2015.

**Figure 13.**
Example of daily SST fields (°C) from (**a**) COMS/MI-derived SST and (**b**) OSTIA on March 26, 2015. (**c**,**d**) are SSTs in the seas around the Korean Peninsula, enlarged images from (**a**,**b**), respectively.

**Figure 13.**
Example of daily SST fields (°C) from (**a**) COMS/MI-derived SST and (**b**) OSTIA on March 26, 2015. (**c**,**d**) are SSTs in the seas around the Korean Peninsula, enlarged images from (**a**,**b**), respectively.

**Table 1.**
Spectral channel characteristics of the Meteorological Imager (MI) on Communication, Ocean and Meteorological Satellite (COMS) (VIS: visible, SWIR: shortwave infrared, WV: water vapor, IR1: infrared 1, IR2: infrared 2).

**Table 1.**
Spectral channel characteristics of the Meteorological Imager (MI) on Communication, Ocean and Meteorological Satellite (COMS) (VIS: visible, SWIR: shortwave infrared, WV: water vapor, IR1: infrared 1, IR2: infrared 2).

Channel | Description | Wavelength (μm) | Bandwidth (μm) | NEDT at 300K (K) | Spatial Resolution (km) |
---|

1 | VIS | 0.675 | 0.55–0.80 | | 1 |

2 | SWIR | 3.75 | 3.5–4.0 | <0.10 | 4 |

3 | WV | 6.75 | 6.5–7.0 | <0.12 | 4 |

4 | IR1 | 10.8 | 10.3–11.3 | <0.12 | 4 |

5 | IR2 | 12.0 | 11.5–12.5 | <0.20 | 4 |

**Table 2.**
List of criteria and thresholds for the removal of cloudy and cloud-contaminated pixels and pixels with inappropriate satellite zenith angle and sun reflection angle (VIS: Albedo of visible channel, BT_{IR1}: Brightness temperature of IR1 channel, BT_{IR2}: Brightness temperature of IR2 channel, BT_{SWIR}: Brightness temperature of SWIR channe, STD: standard deviation in a 3 × 3 window, Max: maximum value in a 3 × 3 windo, Min: minimum value in a 3 × 3 window, Max. BT_{IR1} for 10 day: maximum value of 10 consecutive days, SZA: satellite zenith angle, SRA: solar reflection angle, SST: retrieved SST, FGSST: first-guess SST).

**Table 2.**
List of criteria and thresholds for the removal of cloudy and cloud-contaminated pixels and pixels with inappropriate satellite zenith angle and sun reflection angle (VIS: Albedo of visible channel, BT_{IR1}: Brightness temperature of IR1 channel, BT_{IR2}: Brightness temperature of IR2 channel, BT_{SWIR}: Brightness temperature of SWIR channe, STD: standard deviation in a 3 × 3 window, Max: maximum value in a 3 × 3 windo, Min: minimum value in a 3 × 3 window, Max. BT_{IR1} for 10 day: maximum value of 10 consecutive days, SZA: satellite zenith angle, SRA: solar reflection angle, SST: retrieved SST, FGSST: first-guess SST).

Criteria | Thresholds |
---|

Day | Night |
---|

VIS | >5% | - |

BT_{IR1} | <−3.5 °C | <−3.5 °C |

STD VIS | >0.5% | - |

STD BT_{IR1} | >0.7 °C | >0.5 °C |

STD BT_{IR2} | >0.7 °C | >0.5 °C |

Max. VIS–Min. VIS | >3% | - |

Max. BT_{IR1}–Min. BT_{IR1} | >0.7 °C | >0.5 °C |

Max. BT_{IR2}–Min. BT_{IR2} | >0.7 °C | >0.5 °C |

BT_{IR1}–BT_{IR2} | >0.0032 * BT_{IR1}^{2} + 0.0996 * BT_{IR1} + 1.607 (°C) (if BT_{IR1} ≤ 20 °C) >6 °C (if BT_{IR1} > 20 °C) |

BT_{SWIR}–BT_{IR2} | - | <exp(−9.375 + 0.0342 * BT_{IR1}) (°C) |

BT_{IR1}–Max. BT_{IR1} for 10 day | <−3 °C | <−3 °C |

SZA | >60° | >60° |

SRA | <15° | - |

|SST–FGSST| | >3 °C | >3 °C |

**Table 3.**
Formula based on MCSST and NLSST algorithms ($B{T}_{3.7}$, $B{T}_{11}$, and $B{T}_{12}$: Brightness temperature at 3.7 $\mathsf{\mu}\mathrm{m}$ (SWIR), 11 $\mathsf{\mu}\mathrm{m}$ (IR1), and 12 $\mathsf{\mu}\mathrm{m}$ (IR2), SS${T}_{FG}$: first-guess SST (in Celsius), ${S}_{\theta}=\mathit{sec}\left(\theta \right)-1$, $\theta $: satellite zenith angle, ${a}_{0}$, ${a}_{1}$, ${a}_{2}$, and ${a}_{3}$: regression coefficients).

**Table 3.**
Formula based on MCSST and NLSST algorithms ($B{T}_{3.7}$, $B{T}_{11}$, and $B{T}_{12}$: Brightness temperature at 3.7 $\mathsf{\mu}\mathrm{m}$ (SWIR), 11 $\mathsf{\mu}\mathrm{m}$ (IR1), and 12 $\mathsf{\mu}\mathrm{m}$ (IR2), SS${T}_{FG}$: first-guess SST (in Celsius), ${S}_{\theta}=\mathit{sec}\left(\theta \right)-1$, $\theta $: satellite zenith angle, ${a}_{0}$, ${a}_{1}$, ${a}_{2}$, and ${a}_{3}$: regression coefficients).

Alg. | Time | Equation | Ref. | RMSE (°C) | Bias (°C) |
---|

MCSST | Night | $SST={a}_{0}+{a}_{1}B{T}_{11}+{a}_{2}B{T}_{3.7}+{a}_{3}B{T}_{12}+{a}_{4}{S}_{\theta}\left(B{T}_{3.7}-B{T}_{12}\right)+{a}_{5}{S}_{\theta}$ | [50] | 0.55 | −0.01 |

NLSST | Day | $SST={a}_{0}+{a}_{1}B{T}_{11}+{a}_{2}SS{T}_{FG}\left(B{T}_{11}-B{T}_{12}\right)+{a}_{3}{S}_{\theta}\left(B{T}_{11}-B{T}_{12}\right)$ | [50] | 0.58 | −0.01 |

Day | $SST={a}_{0}+{a}_{1}B{T}_{11}+{a}_{2}SS{T}_{FG}\left(B{T}_{11}-B{T}_{12}\right)+{a}_{3}\left(B{T}_{11}-B{T}_{12}\right)+{a}_{4}{S}_{\theta}\left(B{T}_{11}-B{T}_{12}\right)$ | [51] | 0.55 | 0.01 |

Night | $SST={a}_{0}+{a}_{1}B{T}_{11}+{a}_{2}SS{T}_{FG}\left(B{T}_{3.7}-B{T}_{12}\right)+{a}_{3}\left(B{T}_{3.7}-B{T}_{12}\right)\phantom{\rule{0ex}{0ex}}+{a}_{4}{S}_{\theta}\mathit{VOr}\_\mathit{ernayyuated}\text{}\mathit{using}\text{}\mathit{both}\text{}\mathit{ol}\text{}\left(\mathit{QC}\right)\text{}\mathit{proceduresy}\text{}\mathit{applyin}\text{}\mathit{an}\text{}\mathit{algorithm}\text{}\mathit{optimized}\text{}\mathit{for}\text{}\mathit{each}\text{}\mathit{sensor}.$ | 0.52 | −0.04 |

All | $SST={a}_{0}+{a}_{1}B{T}_{11}+{a}_{2}SS{T}_{FG}\left(B{T}_{11}-B{T}_{12}\right)+{a}_{3}{S}_{\theta}\left(B{T}_{11}-B{T}_{12}\right)$ | [18] | 0.58 | −0.01 |

0.71 | −0.08 |

All | $SST={a}_{0}+{a}_{1}B{T}_{11}+{a}_{2}\left(B{T}_{11}-B{T}_{12}\right)+{a}_{3}{S}_{\theta}\left(B{T}_{11}-B{T}_{12}\right)+{a}_{4}SS{T}_{FG}$ | [52] | 0.34 | −0.01 |

0.29 | 0.00 |

All | $SST={a}_{0}+({a}_{1}+{a}_{2}{S}_{\theta})B{T}_{11}+\left({a}_{3}+{a}_{4}SS{T}_{FG}+{a}_{5}{S}_{\theta}\right)\left(B{T}_{11}-B{T}_{12}\right)+{a}_{6}{S}_{\theta}$ | [20] | 0.55 | 0.01 |

0.63 | −0.02 |

**Table 4.**
Coefficients of MCSST and NLSST algorithms in the Equations (1)–(4) and root mean squared errors (RMSEs) and bias errors.

**Table 4.**
Coefficients of MCSST and NLSST algorithms in the Equations (1)–(4) and root mean squared errors (RMSEs) and bias errors.

Algorithm | Type | Time | Coefficients | RMSE (°C) | Bias (°C) |
---|

a0 | a1 | a2 | a3 |
---|

MCSST | Split | Day | −0.4907 | 1.0039 | 1.9956 | 0.7340 | 0.67 | −0.02 |

Split | Night | 0.6351 | 1.0196 | 1.5888 | 0.7250 | 0.79 | −0.04 |

Triple | Night | 2.0183 | 0.9849 | 0.7737 | 0.4149 | 0.56 | −0.02 |

NLSST | Split | Day | 2.1785 | 0.9071 | 0.0650 | 0.7499 | 0.58 | −0.01 |

Split | Night | 2.7423 | 0.9272 | 0.0563 | 0.6946 | 0.71 | −0.08 |

Triple | Night | 3.2185 | 0.9381 | 0.0259 | 0.4450 | 0.52 | −0.04 |