SST Diagnostic
**************

Here we compute some statistics on both Sea surface temperature from Croco and from SEVRI product

The statistics have been computed over a period of 1 year between July 2012 and July 2013.

SEVIRI is described here :

https://podaac.jpl.nasa.gov/dataset/SEVIRI_SST-OSISAF-L3C-v1.0


#. List of diagnostics 

   * Monthly average comparison
   * Extrema  (percentiles 5 and 95) comparison over the year
   * Root Mean Square Error computed monthly
   * Fitting of the SST on a function and compare the terms between Croco and Seviri 
   
   .. math:: 
      :label: euler
   
       SST(t)=a*cos(2*pi*t/Ta+phi)+b*t+c
   
   
   The parameters of the function allow to compare differents components of the SST and show the skills of the model.
   
   * *a* is the annual amplitude
   * *Ta* is annual period
   * *phi* is phase lagging from the begining of simulation 
   * *b* is trending term reprensting the inter annual variability
   * *c* is the annual mean around which the SST is varying
   
   .. image:: images/SST/sst_fitting.png
   .. image:: images/SST/sst_monthly_diff_croco_seviri.png
   .. image:: images/SST/sst_monthly_rmse_croco_seviri.png
   .. image:: images/SST/sst_stats_croco_seviri.png
   .. image:: images/SST/sst_daily_mean_basins.png

#. What scripts ?
   
    * **calc_save_sst_stats.ipynb** is the notebook which compute the statistics and save the results as a netcdf file
    * **plot_sst.ipynb** is the notebook which read the intermediate files and do the plots