This folder contains the model simulation output of the Veros ocean model that underlies the publication "Risk of tipping the overturning circulation due to increasing rates of ice melt" by Johannes Lohmann and Peter D. Ditlevsen (2021)

The spin-up simulations at present-day forcing are found in the folder "spinup".

"s30" corresponds to simulations with a 30-day salinity relaxation time scale.
"s360" corresponds to simulations with a 360-day salinity relaxation time scale.
"windstress" corresponds to simulations with 360-day salinity relaxation time scale and ERA-40 windstress forcing.

The transient simulations where the North Atlantic freshwater anomaly is slowly ramped up and down again are found in the folder "hysteresis". The data corresponds to the results shown in Fig.1B, Fig. S1 and Fig. S2 of the publication.

The simulations showing the rate-dependence of tipping are located in the folder "rate_dependence", and correspond to the data underlying Fig. 1C-D and 3A. The subfolders correspond to the individual realizations with different ramping durations, where e.g. "1" corresponds to 1,000 years of ramping, and "001" to 10 years.

The folder "init_dependence" contains simulations underlying Fig. 3B-C. The subfolder "instant_ramping" contains the simulations uncerlying Fig. 4.
The different subfolders "2y", "2mo" and "1w" separate the results where initial conditions were sampled from the attractor with a temporal spacing of 2 years, 2 months, and 1 week, respectively. Finally, the numbered subfolders contain the individual realizations with the initial conditions sampled along the attractor.

For all simulations, a snapshot of all model variables has been saved every 5 years in *.h5 files. Furthermore, 5-year averages of key variables have been saved in the *averages.nc files. The *overturning.nc files are not to be used. Instead, the overturning stream functions may be computed using the calc_overturning.py scripts. Individual files contain 20 consecutive samples, and thus 100 years of model output. Consecutive 100-year segments of a simulation are numbered *.0000.*, *.0001.*, *.0002.*, etc. Exceptions are the spinup simulations in "init_dependence", where output has been written more frequently.

For further questions contact johannes.lohmann@nbi.ku.dk