A Nowcast/Forecast System for Japan’s Coasts Using Daily Assimilation of Remote Sensing and In Situ Data

Yasumasa Miyazawa,Misako Kachi,Hiroshi Murakami,Toru Miyama,Sergey M Varlamov,Yukio Kurihara

doi:10.3390/rs13132431

Abstract

We have developed an ocean state nowcast/forecast system (JCOPE-T DA) that targets the coastal waters around Japan and assimilates daily remote sensing and in situ data. The ocean model component is developed based on the Princeton Ocean Model with a generalized sigma coordinate and calculates oceanic conditions with a 1/36-degree (2–3 km) resolution and an hourly result output interval. To effectively represent oceanic phenomena with a spatial scale smaller than 100 km, we adopted a data assimilation scheme that explicitly separates larger and smaller horizontal scales from satellite sea surface temperature data. Our model is updated daily through data assimilation using the latest available remote-sensing data. Here we validate the data assimilation products of JCOPE-T DA using various kinds of in situ observational data. This validation proves that the JCOPE-T DA model output outperforms those of a previous version of JCOPE-T, which is based on nudging the values of temperature and salinity toward those provided by a different coarse grid data-assimilated model JCOPE2M. Parameter sensitivity experiments show that the selection of horizontal scale separation parameters considerably affects the representation of sea surface temperature. Additional experiments demonstrate that the assimilation of daily-updated satellite sea surface temperature data actually improves the model’s efficiency in representing typhoon-induced disturbances of sea surface temperature on a time scale of a few days. Assimilation of additional in situ data, such as temperature/salinity/ocean current information, further improves the model’s ability to represent the ocean currents near the coast accurately.

Highlights

Ocean state forecast (OSF), targeting ocean mesoscale phenomena, has been mainly using satellite sea surface height anomaly (SSHA) data for capturing the target phenomena through data assimilation (DA) since pioneering works in the 1990s (e.g., [1]).Other typical satellite remote-sensing data of ocean, sea surface temperature (SST), have been used by OSF for correcting the water mass property mainly above the surface mixed layer [2]
We examined the DA impacts of the in situ temperature/salinity profiles (TS) data using a result of a sensitivity experiment (INSTS) which assimilated the in situ TS data into the SUMM3 case, which assimilated the only remote-sensing data
The model validation using the skill metrics based on the independent observational data clearly demonstrated that Japan Coastal Ocean Predictability Experiment (JCOPE)-T DA based on the direct DA approach outperformed the previous version of JCOPE-T based on the indirect DA approach

Summary

Introduction

Ocean state forecast (OSF), targeting ocean mesoscale phenomena, has been mainly using satellite sea surface height anomaly (SSHA) data for capturing the target phenomena through data assimilation (DA) since pioneering works in the 1990s (e.g., [1]). Other typical satellite remote-sensing data of ocean, sea surface temperature (SST), have been used by OSF for correcting the water mass property mainly above the surface mixed layer [2]. The targeted spatiotemporal scales of the basin-scale operational OSF systems are O (100 km) and O (10 day) [3,4]. The satellite SSH measurements with a minimum sampling scale of 100 km and a major sampling period of 10 days fulfill the requirements for DA in a basin-scale operational OSF system

Methods

Results

Discussion

Conclusion