Overview

This dataset contains the model ouputs, scripts, and figures used to check the analyses and figures prepared for publication (Yamagami et al. 2026).

Notes

The experiment names correspond as follows: pi.gulf2 to NAGA, pi.kuro2 to NPGA, pictl.gm3 to CTL, pilow.gulf2 to NAGA-LR, and pilow.kuro2 to NPGA-LR.
For the coupled model experiments, the ten ensemble members from run06 through run15 are used in this analysis. The names run01 through run05 were assigned to preliminary experiments with different configurations and are not used in this study.
The "agcm" data is from atmosphere-only model experiments. The member numbers run01 through run10 used there are separate from those of the coupled model experiments described above.
Because uploading spatial fields for all individual ensemble members does not complete successfully, only ensemble-mean spatial fields are provided through Zenodo.

Contents

# scripts.zip
Python scripts for producing the main figures. Multiple original code files were refactored into a version suitable for public release.

# figs.zip
Files of the publication figures.

# u60N_agcm_{clim,kuro,gulf}_y2100_2120.nc
Outputs from atmosphere-only model experiments. The data are divided into agcm_clim, forced with climatological SST; agcm_gulf, forced with SST anomalies in the Gulf Stream region; and agcm_kuro, forced with SST anomalies in the Kuroshio region. For each experiment, the file contains daily zonal-mean zonal wind at 60°N (u60N) from y2100 through y2120. y2100 is a spinup and is not used for analysis.

# csv.tar.gz
CSV files containing monthly SST indices calculated for each experiment. They contain regional-mean time series for the Kuroshio and Gulf Stream regions, the tropics, the Northern and Southern Hemisphere midlatitudes, Niño 3.4, and other regions, organized by experiment and ensemble member.

# compo_annual.tar.gz
Results of composite analyses based on the Kuroshio SST index. For NAGA and NAGA-LR, this directory contains composite fields of SST (sst), sea surface height (sho), and surface velocity (velo), together with statistical significance data (*_sig.nc).

# compo_daily_zonal_u60N.tar.gz
Daily zonal-wind composites at 60°N based on the DJF Gulf Stream or Kuroshio SST index. This directory contains NetCDF data for NAGA and NPGA, and statistical significance data (*_sig.nc).

# reg_atm_field.tar.gz
Monthly lag-regression results for atmospheric fields against the DJF Gulf Stream and Kuroshio SST indices. The directory contains regression coefficients and statistical significance data (*_sig.nc) for sea level pressure (slp), 500 hPa geopotential height (z500), and 30 hPa geopotential height (z30).

# reg_xy_ocn.tar.gz
Annual-mean lag-regression results for the ocean mixed-layer heat budget against the DJF SST index. The directory contains temperature tendency (dtdt), surface heat flux (hf), advection (adv), and entrainment (ent), together with statistical significance data for each term (*_sig.nc).

# for.linear.rossby.zip
Input data prepared for the linear Rossby wave model. For NAGA, NAGA-LR, and CTL, this directory contains zonal and meridional surface wind stress (otaux and otauy) and sea surface height (sho). All data were converted to the 1°x1° resolution.

# sst_ssp245.tar.gz
SST and regional-mean time series for 1980–2024. The historical directory corresponds to 1980–2014, and ssp245 contains the standard SSP2-4.5 experiment for 2015–2024. The sensitivity experiments for 2015–2024 are pogacobe.y2015.ssp245 for the tropical Pacific (TP), pacecobe.y2015.ssp245 for the North Atlantic (NA), and tpgscobe.y2015.ssp245 for TP+NA. Each directory contains ensemble-mean SST in gtool format, while the CSV files at the top level contain regional-mean SST indices calculated for each run and for the observational datasets COBE-SST2 and OISSTv2.

Main File-Name Conventions

clim: climatology
anom: anomaly
enmean: ensemble mean
lagmN: field N years or months before the index
lag0: field at the same time as the index
sig: statistical significance data
y2100-2199, etc.: analysis period