Jim Rosinski (rosinski@cgd.ucar.edu)
Mon, 29 Mar 1999 09:38:36 -0700 (MST)
On Mon, 29 Mar 1999, Gibbas, Mark J. wrote:
> The definesurf code and the CCM code is functioning correctly. The seaice
> disappeared because the input SST data did not set seaice temps low enough
> to be identified by the ccm threshold for seaice. Specifically our input
> SST data set seaice at -1.75C and the CCM uses -1.7999C as the threshold.
> To solve the problem we set the variable
> tsice = -1.74 in the file src/dom/parsst.h
I recently encountered a similar phenomenon. Sea ice was disappearing due to
a separate offline code of mine which was writing SST values in 32-bit
precision. When "interpsst" (contained in ccm/tools/interpsst) read this
32-bit data, something in the underlying algorithm changed the sea ice flag
values from -1.8 to -1.75 degrees C. Your solution of changing the value of
tsice in the model code may work just fine. I got around the problem by
making sure the input to interpsst is always maintained in 64-bit precision.
Don't know if you were encountering a similar situation.
Jim Rosinski
CCM Core Group
This archive was generated by hypermail 2.0b3 on Tue Aug 24 1999 - 11:27:15 MDT