[ecco-support] Access to ECCOV4

Wed Apr 5 21:55:52 EDT 2017

Ian did quite a good job explaining where to start.  I just wanted to
mention that I have done budgets for ECCO v4 for heat content (not
salinity).   I also did the separation on the linear advective convergences
into Ekman and geostrophic parts.  I took the approach that Ian described
(first close budgets, then look at Ekman/geostrophic decomposition). See
http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-13-00316.1
http://journals.ametsoc.org/doi/10.1175/JCLI-D-14-00579.1

I can provide some code as examples for your salinity budget.
Martha

On Wed, Apr 5, 2017 at 5:21 PM, Fenty, Ian G (329C) <Ian.Fenty at jpl.nasa.gov>
wrote:

> Guancheng,
>
> Just a clarification if you are using the ECCO v4-r2 solution, we call
> that the “ECCO Central Production version 4, release 2” state estimate.
> Our ECCO2 solutions use a higher resolution, 18 km eddy-permitting grid.
>
> As to your problem, our state estimates will close heat and salt budgets
> on the native model grid.  It looks like you are trying to close the MLS
> budget using only Ekman and geostrophic transports estimated using wind
> stress and SSH gradients.
>
> Using the method you described it is very unlikely that you would exactly
> close the MLS budget for several reasons.  First, actual salinity fluxes in
> the MLS include eddy advection and MLS depth changes.  Second, you are
> using monthly-mean wind stress and SSH gradients while salinity fluxes in
> the model are driven by time-varying wind stress and SSH gradients.  Third,
> you are looking at the ‘interp_monthly’ fields which are not on the native
> model grid.  'interp' fields are on an interpolated lat-lon grid.
>
> My suggestion is that you work backwards.  At first, use the native model
> grid and fluxes on the native model grid and use the output flux terms and
> surface forcing and prove to yourself that you can close the MLS budget.
>
> To proceed, examine the individual advective and diffusive flux terms
> across the faces of the model grid cells.  The units of the flux terms are
> in g/kg m3/s because they are volume fluxes.  g/kg (m/s) m^2 or S*U*A, with
> S being g/kg, U being (m/s) equivalent velocity, and A (m^2) being the area
> of the face of the grid cell through which the flux is considered.
>
> For each surface cell there are 5 advective terms, one for each face:
>
> 2 terms for the 'x' dimension lateral faces: ADVx_SLT,   "U Comp.
> Advective Flux of Salinity"
> 2 terms for the 'y' dimension lateral faces: ADVy_SLT,   "V Comp.
> Advective Flux of Salinity"
> 1 terms for the 'z' dimension bottom faces: ADVr_SLT,   "Vertical
> Advective Flux of Salinity"
>
> and 6 diffusive terms, the bottom face now has an additional term:
>
> 2 terms for the 'x' dimension lateral faces: DFx_SLT,   "U Comp. Diffusive
> Flux of Salinity"
> 2 terms for the 'y dimension lateral faces: DFy_SLT,   "V Comp. Diffusive
> Flux of Salinity"
> 1 terms for the 'z' dimension bottom faces: DFrE_SLT,   "Vertical
> Advective Flux of Salinity (explicit part)"
> 1 terms for the 'z' dimension bottom faces: DFrI_SLT,   "Vertical
> Advective Flux of Salinity (implicit part)"
>
> If a grid cell below the surface is considered, then you need to consider
> the additional advective and diffusive fluxes in the 'z' dimension across
> the top face as well.
>
> To fully close the budget for grid cells at the surface you must also
> consider the air-sea freshwater flux term, oceFWflux.
>
> While doing that it may be helpful to examine the 3-day T and S average
> tendency terms which are included in our output.
>
> ftp://mit.ecco-group.org/ecco_for_las/version_4/release2/
> nctiles_tendencies/README
>
> Stend    : "Salinity Tendency"
> Sdi      : "Salinity Tendency Due To Diffusion"
> Sad      : "Salinity Tendency Due To Advection"
> Sforc    : "Salinity Tendency Due To Forcing"
> SALT     : "Salinity Snapshots"
>
> The total salinity tendency variable “Stend” is the sum of the individual
> tendency terms.  We also provide 3-daily snapshots of salinity so you can
> verify that salinity evolves as expected.
>
> After you have convinced yourself that you can exactly close the salt
> budget, then you can attempt to reconcile the advective fluxes into the
> Ekman and geostrophic transports.  These diffusive fluxes can be thought of
> as parameterizing eddy fluxes which are not accounted for in the
> approximation that dS/dt is due to only Ekman and geostrophic fluxes.
>
> -Ian
>
> On 4/5/17, 12:45 AM, "ecco-support-bounces at mit.edu on behalf of Ligc" <
> ecco-support-bounces at mit.edu on behalf of liguancheng15 at mails.ucas.ac.cn>
> wrote:
>
>
>     Dear Sir/Medam,
>           Firstly, thank you for your reply so much! Now,please allow me
> to introduce my method and questions.
>      I used the Budget equation of the mixed layer salinity to calculate
> the MLS variability around North Pacific using by ECCO2 original data (from
> ECCO v4-r2/ interp_monthly).The left term is the
>      rate of change of the MLS, and the right terms represent the
> freshwater flux (E-P), advection of the MLS by Ekman and geostrophic
> transports, the entrainment of the MLS. Among these variables，Ekman
> Velocity was estimated from wind stress data and geostrophic
>      Velocity from free surface height Anomalies using by the related
> formula.
>      I tried to calculate all of these terms at each lon-lat grid by using
> the original variables, but i found that the total range at the right hand
> is a bit more widely than the left, which may
>      be resulted by the wrong or larger result of the advection of MLS
> because other term is fitted with the conclusion of other articles
> well，including the freshwater flux is the leading dominant term in this
> area.
>       So i was wondering if there's any available variable or method to
> test my results. And, i want to know about how to calculate advective flux
> of Salinity in a better way.
>       Moreover, Could you tell me how to get the ADVe_SLT and ADvn_SLT in
> this dataset, because i can not figure out the source of 'm^3' in these
> unit. I am looking forward for your helpless if it
>      is convenient.
>     Thank you for your patient reading! I really appreciate you help!
>
>
>     Best regards,
>     Guancheng
>
>
>
>
>
>
>
>
>
> _______________________________________________
> ecco-support mailing list
> ecco-support at mit.edu
> http://mailman.mit.edu/mailman/listinfo/ecco-support
>

-- 
Martha W. Buckley
marthabuckley at gmail.com
http://sites.google.com/site/marthabuckley/home
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/ecco-support/attachments/20170405/2c4e2da8/attachment.html