<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN">
<html><body>
<p>Hello to all,</p>
<p>I am trying to configure the namelist of GGL90 package in order to have the most realistic representation of vertical mixing as possible. As I am a new user of MITgcm I wonder if someone could share her or his knowledge of this package. I am using a regional configuration of the Mediterranean Sea at 1/16° with increased resolution at the Strait of Gibraltar (up to 1/200°) and 100 vertical levels.</p>
<p>For the moment here is the parameters that I would change from the reference values, and some justifications:</p>
<ul>
<li><strong>GGL90TKEmin = 1.0 _d -10</strong></li>
</ul>
<p>(in order to have a minimal mixing length that is not too big, see next point)</p>
<ul>
<li><strong>GGL90mixingLengthMin = 10^-6 / (GGL90ck* sqrt(GGL90TKEmin)) = 1</strong></li>
</ul>
<p>I chose this formulation to obtain a molecular viscosity (10^-6 m^2/s) when the TKE reaches its minimal value.</p>
<ul>
<li><strong>mxlMaxFlag = 2 </strong></li>
</ul>
<p>As in [Gaspar et al 1990]</p>
<ul>
<li><strong>GGL90_dirichlet = .FALSE.</strong></li>
</ul>
<p>To allow more regular connections to the bottom boundary.</p>
<ul>
<li><strong>calcMeanVertShear = .TRUE.</strong></li>
</ul>
<p>It seems better to me to take the mean vertical shear than to calculate the vertical shear of the mean speeds.</p>
<p> </p>
<p>Thank you in advance for your time and expertise.</p>
<p>Best regards,</p>
<p>Nicolas</p>
<p> </p>
<div> </div>
</body></html>