High-Resolution Sea Surface Temperatures Derived from Landsat 8: A Study of Submesoscale Frontal Structures on the Pacific Shelf off the Hokkaido Coast, Japan

Hiroshi Kuroda,Yuko Toya

doi:10.3390/rs12203326

Abstract

Coastal and offshore waters are generally separated by a barrier or “ocean front” on the continental shelf. A basic question arises as to what the representative spatial scale across the front may be. To answer this question, we simply corrected skin sea surface temperatures (SSTs) estimated from Landsat 8 imagery with a resolution of 100 m using skin SSTs estimated from geostationary meteorological satellite Himawari 8 with a resolution of 2 km. We analyzed snapshot images of skin SSTs on 13 October 2016, when we performed a simultaneous ship survey. We focused in particular on submesoscale thermal fronts on the Pacific shelf off the southeastern coast of Hokkaido, Japan. The overall spatial distribution of skin SSTs was consistent between Landsat 8 and Himawari 8; however, the spatial distribution of horizontal gradients of skin SSTs differed greatly between the two datasets. Some parts of strong fronts on the order of 1 °C km−1 were underestimated with Himawari 8, mainly because of low resolution, whereas weak fronts on the order of 0.1 °C km−1 were obscured in the Landsat 8 imagery because the signal-to-noise ratios were low. The widths of the strong fronts were estimated to be 114–461 m via Landsat 8 imagery and 539–1050 m via in situ ship survey. The difference was probably attributable to the difference in measurement depth of the SST, i.e., about 10-μm skin layer by satellite and a few dozen centimeters below the sea surface by the in situ survey. Our results indicated that an ocean model with a grid size of no more than ≤100–200 m is essential for realistic simulation of the frontal structure on the shelf.

Highlights

Sea surface temperature (SST) is one of the most essential variables in the study of oceanic, atmospheric, and marine ecosystem processes
This low-temperature bias relative to Himawari 8 was much more serious for brightness temperatures from Band 11 (Figure 2b). These results indicated that the brightness temperatures from the Landsat 8 needed to be corrected to obtain precise skin SSTs
We proposed a very simple method to correct skin SSTs from Landsat 8 imagery by comparisons with skin SSTs from Himawari 8, and we described the horizontal structure of SST fronts on the Pacific shelf off the Hokkaido coast of Japan, together with in situ SSTs measured along the ship track

Summary

Introduction

Sea surface temperature (SST) is one of the most essential variables in the study of oceanic, atmospheric, and marine ecosystem processes. In 1981, global measurements of high-quality SSTs began using an innovative multi-spectral sensor referred to as the Advanced Very High Resolution Radiometer (AVHRR) on board the National Oceanic and Atmospheric Administration (NOAA) polar-orbiting operational environmental satellites [1,2]. The Advanced Himawari Imager has 16 observation bands. The greatest advance offered by Himawari 8 has been the high frequency of the Full Disk observations, i.e., every 10 min. High-frequency measurements have several important advantages compared with low-frequency samplings by non-geostationary satellites: high-frequency measurements make it possible to distinguish biases of skin SSTs between nighttime and daytime observations; they reduce cloud noise from skin SST; and they facilitate comparisons with data from other satellites with a temporal error of less than 10 min. The skin SSTs from Himawari 8 are informative enough to enhance understanding of variations in the ocean [12,13,14] and enable appropriate initialization of an ocean forecast system [15]

Objectives

Methods

Results

Conclusion