Validation of Satellite Sea Surface Temperatures and Long-Term Trends in Korean Coastal Regions over Past Decades (1982–2018)

Abstract

Validation of daily Optimum Interpolation Sea Surface Temperature (OISST) data from 1982 to 2018 was performed by comparison with quality-controlled in situ water temperature data from Korea Meteorological Administration moored buoys and Korea Oceanographic Data Center observations in the coastal regions around the Korean Peninsula. In contrast to the relatively high accuracy of the SSTs in the open ocean, the SSTs of the coastal regions exhibited large root-mean-square errors (RMSE) ranging from 0.75 K to 1.99 K and a bias ranging from −0.51 K to 1.27 K, which tended to be amplified towards the coastal lines. The coastal SSTs in the Yellow Sea presented much higher RMSE and bias due to the appearance of cold water on the surface induced by vigorous tidal mixing over shallow bathymetry. The long-term trends of OISSTs were also compared with those of in situ water temperatures over decades. Although the trends of OISSTs deviated from those of in situ temperatures in coastal regions, the spatial patterns of the OISST trends revealed a similar structure to those of in situ temperature trends. The trends of SSTs using satellite data explained about 99% of the trends in in situ temperatures in offshore regions (>25 km from the shoreline). This study discusses the limitations and potential of global SSTs as well as long-term SST trends, especially in Korean coastal regions, considering diverse applications of satellite SSTs and increasing vulnerability to climate change.

Primary production by phytoplankton in the territorial seas of the Republic of Korea

“…The annual mean SSTs in the eastern South Sea and East Sea significantly increased during the 17-year study period (Figs 4 & 5). Although the increase in the annual mean temperatures in other regions was not significant in the present study, there was a tendency to increase the SST in the Korean coastal waters, which is in agreement with previous studies Park 2020).…”

A Song of Wind and Ice: Increased Frequency of Marine Cold‐Spells in Southwestern Patagonia and Their Possible Effects on Giant Kelp Forests

“…In the complete area of study, three validation sites in the immediate vicinity of OISST pixels exist, plus two more validation sites at a considerable distance to the satellite records. Despite this, the levels of error and bias are in line with previous studies (Lee & Park, 2020; Stobart et al., 2015), which is something to be expected for global SST data adjacent to coastlines. The larger RMSE of Caleta Meteoro (1.80) and Puerto Williams (1.25) may be explained by the complex coastline and bathymetry of the Pacific side of the Strait of Magellan and the turbulent waters around the Cape Horn, which contributed to an increased variability in the data.…”

“…The accuracy of satellite‐detected SST data in the inner waters of Patagonia could be potentially lower than that in the open seas of the Southern Hemisphere (Reynolds et al., 2007). Furthermore, previous studies have warned about the importance of calculating the level of certainty and bias when using satellite‐detected SST in coastal areas (Lee & Park, 2020; Smale & Wernberg, 2009; Stobart et al., 2015). To explore this, we compared OISST pixels with in situ automatic SST records from five sites covering five marine ecosystems of our area of study (Table 1).…”

Application of Non-stationary Extreme Value Analysis to Satellite-Observed Sea Surface Temperature Data for Past Decades

“…Concentrated on the isotherm of approximately 26 • C, the maximum SSTs from the model simulation were spatially dominant in the coastal region of the Korean Peninsula up to 42 • N compared to the offshore region in the central part of the EJS. The limitation of this characteristic distribution is that it is difficult to compare accurately because both satellite and model data have relatively low accuracies in coastal regions (Lee and Park, 2020). In particular, for different spatial resolutions, 1 • resolution model data and 0.25 • resolution satellite data produced relatively large errors in the coastal region.…”