id	summary	reporter	owner	description	type	status	priority	milestone	component	version	resolution	keywords	cc
609	Spatially varying Jerlov water type	arango	arango	"Added the option to have spatially varying Jerlov water types for light absorption in the water column.  This is used when '''SOLAR_SOURCE''' or '''LMD_SKPP''' CPP options are activated.

A new CPP option '''WTYPE_GRID''' is introduced to activate spatially varying Jerlov water types.  The ROMS internal variable '''Jwtype''' is now a floating-point variable to facilitate I/O manipulation within the metadata model.  The spatially varying Jerlov water type index can be specified in two ways:

 * A variable, '''wtype_grid''', in the ROMS '''GRID NetCDF''' file:
{{{
        double wtype_grid(eta_rho, xi_rho) ;
                wtype_grid:long_name = ""Jerlov water type index"" ;
                wtype_grid:coordinates = ""lon_rho lat_rho"" ;
}}}

 * or an analytical function in new routine '''ana_wtype.h''', which can be activated with new CPP option '''ANA_WTYPE'''.  For example, we can have an expression to set the Jerlov water type index in terms of the bathymetry to give a distribution of values between 5 (shallow; turbid water) and 1 (deep; clear water): 
{{{
      fac=1.0/1000.0_r8                   ! Inverse bathymetry threshold
      DO j=JstrT,JendT
        DO i=IstrT,IendT
          Jwtype(i,j)=ANINT(5.0_r8-4.5_r8*(TANH(h(i,j)*fac)))      ! 1:5
        END DO
      END DO
}}}

The variable '''Jwtype''' is used in '''lmd_swfrac''' to compute the fraction of shortwave flux penetrating the water column (light absorption), modeled as a double exponential decay function in Jerlov water type.  The double exponential parameters ('''lmd_mu1''', '''lmd_mu2''', '''lmd_r1''') were also enhanced to add additional Jerlov coastal water types: '''1''', '''3''', '''5''', '''7'''.  Currently, the following Jerlow water types are supported:

{{{
   Array     Jerlov
   Index   Water Type   Examples
   -----   ----------   --------

     1         I        Open Pacific
     2         IA       Eastern Mediterranean, Indian Ocean
     3         IB       Western Mediterranean, Open Atlantic
     4         II       Coastal waters, Azores
     5         III      Coastal waters, North Sea
     6         1        Skagerrak Strait
     7         3        Baltic
     8         5        Black Sea
     9         7        Coastal waters, dark
}}}

The range of indices 1:9 are ordered by increasing absorption: from clear water (type '''I''') to dark turbidity water (type '''7''').  The indices correspond to the paramenters ('''lmd_mu1''', '''lmd_mu2''', '''lmd_r1''') used to model the light absorption into the water column using a double exponential fitting function of (Paulson and Simpson, 1997).

The following figure shows the Jerlov's water type classification based on transmittance along 1 m of seawater:

[[Image(https://www.myroms.org/trac/Jerlov_WaterTypes.png, 500)]]

The water type '''9''' was not included in the above array classification indices because its double exponential fitting parameters are not available.

----

I also introduced routine '''ana_dqdsst.h''' to specify the surface net heat flux sentivity to SST, '''d(Q)/d(SST)''', with analytical expressions (internal ROMS variable '''dqdt''').  This is activated with new CPP option '''ANA_DQDSST'''.  It is now possible to specify the SST in a NetCDF forcing file and the '''d(Q)/d(SST)''' field with an analytical function. No longer we need both variables in the NetCDF file.

I haven't coded the function that computes '''d(Q)/d(SST)''' using the formula reported in Barnier ''et al.'' (1995).  Several additional atmospheric field are required (see their equation '''6b''').  The values for '''d(Q)/d(SST)''' can be negative or positive and has units of Watts meters-2 Celsius-1 (or internal kinematic units of m/s/Celsius).

If you are planning to use '''QCORRECTION''' in your application, I suggest you to read Barnier ''et al.'' (1995) paper first.  There are some messages in the forum suggesting the summer warming trends cannot be fixed with '''QCORRECTION'''.  I just tried in the Gulf of Mexico and it worked well.  It is just a matter how you specify the appropriate '''d(Q)/d(SST)''' values. I usually use monthly composite SST from the NOAA !CoastWatch OpenDAP catalog (http://oceanwatch.pfeg.noaa.gov/thredds/catalog.html), see product '''SST, Blended, Global, EXPERIMENTAL''' (BA/ssta/mday).

'''Reference:'''

Barnier, B., L. Siefridt, and P. Marchesiello, 1995: Thermal forcing for a global ocean circulation model using a three-year climatology of ECMWF analyses, ''J. Marine Systems'', '''6''', 363-380."	upgrade	closed	major	Release ROMS/TOMS 3.7	Nonlinear	3.7	Done