VALSO-VALTRANS

Table of Contents

1. Introduction
2. Installation and running
3. Output
4. Observations
5. Authors
6. Licence
7. Acknowledgements

Introduction

A software package for ocean scientists to calculate and plot the following evaluation metrics to compare Southern Ocean biases and Global biases in transports/exchanges through straits between CMIP models with a NEMO ocean:

• VALSO metrics (Southern Ocean assessment):

  • Drake Passage net eastward transport (ACC)
  • Weddell gyre strength
  • Ross gyre strength
  • Salinity of High Salinity Shelf Water (HSSW) in west Weddell and west Ross Seas
  • Intrusion of Circumpolar Deep Water (CDW) in Amundsen sea
  • Intrusion of CDW on East Ross shelf

• VALTRANS metrics (Straits transports and exchanges):

  • North Atlantic deep overflows: Denmark Strait and Faroe Bank Channel.
  • Marginal Seas exchanges: Gibraltar, Bab el Mandeb, Strait of Hormuz.
  • Indonesian Throughflow: Lombok Strait, Ombai Strait, Timor Passage.

Note that there is also a set of metrics called VALGLO but this needs debugging.

Currently works for output from eORCA1, eORCA025 and eORCA12 models.

Installation and running

Clone the MARINE_VAL repository:

git clone https://github.com/JMMP-Group/MARINE_VAL

Build the CDFTOOLS executables. Note the make macro and modules shown below work for the current Met Office linux servers

module load gcc/8.1.0 mpi/mpich/3.2.1/gnu/8.1.0 \
            hdf5/1.8.20/gnu/8.1.0               \
            netcdf/4.6.1/gnu/8.1.0
cd MARINE_VAL/CDFTOOLS-4.0/src
ln -s ../Macrolib/macro.gfortran_metoffice make.macro
make

Edit environment variables in param.bash to fit your setup/need.

• mesh mask location with mesh mask name
• location of the CDFTOOLS toolbox
• where to store the data output (or link to existing data location)

Edit param.bash to define which metrics you want to calculate, normally a package like VALSO or VALTRANS, but you can pick and choose individual metrics.

Edit style.db to define labels, colours and line styles for the integrations you want to plot (some examples provided).

Process the data to generate the timeseries data:

./run_all.bash -C [chunksize] -B [BATHY] [MESHMASK] [YEARB] [YEARE] [FREQ] [RUNID list]

for example :

./run_all.bash -C 20 -B bathymetry_eORCA025-GO6.nc mesh_mask_eORCA025-GO6.nc 1981 1990 1y u-cl681

• [chunksize] is the number of dates that should be restored from MASS at a time, recommended value 10-20
• [BATHY] is the name of the bathymetry file found in $MSKPATH
• [MESHMASK] is the name of the mesh_mask file found in $MSKPATH.
• [FREQ] options currently 1y for annual means or 1m for monthly means.

Note that the bathymetry file is only required for metrics involving transports through straits. It can be created from the mesh_mask file using SCRIPT/bathy_from_dommesh.py.

Output from the processing scripts appears under the SLURM directory.

Build the plot for the Southern Ocean:

./run_plot_VALSO.bash [KEY] [FREQ] [RUNID list]

for example :

./run_plot_VALSO.bash cpl_and_forced 1y u-am916 u-az867 u-ba470 u-ar685 u-bj000 u-bn477

• [KEY] is an arbitrary label that will be used to name the output PNG file.

Output

VALSO output:

VALTRANS output:

Other output :

• bsf, september mld netcdf file for each year in your DATPATH directory.
• all individual time series are saved in FIGURES along with the txt file describing the exact command line done to build it

Observations

This section gives a bit more detail on where the observations used for the various metrics come from.

VALSO

Transports

The ACC transport value is taken from Donohue, K. A., Tracey, K. L., Watts, D. R., Chidichimo, M. P., and Chereskin, T. K.: "Mean Antarctic Circumpolar Current transport measured in Drake Passage", Geophysical Research Letters, 43, 11,760–11,767, https://doi.org/https://doi.org/10.1002/2016GL070319, (2016).

The Weddell Gyre strength is taken from Klatt, O., Fahrbach, E., Hoppema, M., and Rohardt, G.: "The transport of the Weddell Gyre across the Prime Meridian", Deep Sea Research Part II: Topical Studies in Oceanography, 52, 513–528, https://doi.org/https://doi.org/10.1016/j.dsr2.2004.12.015, "Direct observations of oceanic flow: A tribute to Walter Zenk" (2005). The model metric is the peak positive value of the streamfunction integrating from the southern boundary. It therefore includes the transport of the Antarctic Shelf Current and the recirculating gyre. The Klatt et al measurement is comparable to this.

The Ross Gyre strength is taken from Dotto, T. S., Naveira Garabato, A., Bacon, S., Tsamados, M., Holland, P. R., Hooley, J., Frajka-Williams, E., Ridout, A., and Meredith, M. P.: "Variability of the Ross Gyre, Southern Ocean: Drivers and Responses Revealed by Satellite Altimetry", Geophysical Research Letters, 45, 6195–6204, https://doi.org/https://doi.org/10.1029/2018GL078607 (2018). In this case they measure the transport of the recirculating gyre and the throughflow (Shelf current) separately so I have added these values in order to compare to the model. The variability (standard deviation) of the observations is only quoted for the recirculating gyre so I have used this value.

Shelf temperatures and salinities

The original version of VALSO took average bottom salinities and temperatures over the four boxes defined on the shelf. There was no reference for the corresponding observational values (just P.Mathiot), so as of June 2023 I have updated these to use means over deep boxes (below 400m) instead. For the observational comparison I have used the EN4.2.2.g10 profile dataset which is probably more reliable than the EN4 analysis in this region. I took means and standard deviations for all the profiles with data below 400m in the defined boxes (see the script mk_deepTS.bash for the box lat/lon limits). Because the time and depth sampling of the profiles is uneven I weighted the statistics by month and depth: I calculated means and variances for each month and for 100m depth bins and then took the means of these to give the overall mean and variance.

Note that the final values used for the means of the observations don't differ much from the original observational values for the bottom fields (meaning that Pierre's expert judgment was reasonable) and the change in the calculation of the model metrics from using averages of bottom fields to averages of deep boxes doesn't generally seem to make much difference to these values either (meaning that the deep water isn't very stratified). So the net result is that updating your VALSO calculations to use the new metrics will (probably) not make a material difference to the conclusions you draw, but will make it a bit easier to include the plots in papers.

The pictures here are from a presentation I gave and give some indication of the time distribution of the profiles and the distribution in temperature and salinity space. The first set of plots shows the time distribution of all the available EN4 profiles, the lines colour coded by month. For the EROSS and WWED areas the profiles consist mainly of CTD measurements from cruises and are weighted heavily towards the summer. When the EROSS and WWED metrics are calculated for the model fields we use DJF means rather than annual means to make the comparison with obs more valid. For the AMU and WROSS regions there is good coverage through out the year for some recent years. For these regions the data is mainly seal data. Weddell seals dive surprisingly deep - up to 700m.

The second set of plots shows the distribution of all observations within the relevant boxes (although note that for these plots I used a top-of-box depth of 300m whereas in the end I settled on 400m). The number of obs reduces with increasing depth as you might expect but it is encouraging to note that the deepest values appear to converge in all cases. The third set of plots show model bathymetries (from eORCA12) for the four areas and you can see that the depth coverage of the observations is good in all cases. The shelves around Antarctica are generally deep - of order 500m - and in the AMU and WROSS regions there are deep trenches to over 1000m.

References:

• Good, S. A., Martin, M. J., and Rayner, N. A.: "EN4: quality controlled ocean temperature and salinity profiles and monthly objective analyses with uncertainty estimates", J. Geophys. Res.-Oceans, 118, 6704–6716, https://doi.org/10.1002/2013JC009067, 2013
• Storkey et al: to be submitted to GMD

Weddell mixed layer depth

TBD

VALTRANS

TBD

Authors

• pmathiot
• DaveStorkey

Licence

Acknowledgements

• This is the Met Office development of the original VALSO code by Pierre Mathiot. Pierre's repository is here