| 1 |
Seaice-only estimation in the Labrador Sea |
| 2 |
------------------------------------------ |
| 3 |
09-Sep-2005 |
| 4 |
heimbach@mit.edu, ifenty@mit.edu, menemenlis@jpl.nasa.gov |
| 5 |
|
| 6 |
This is a 2x2 degree configuration of a seaice-only model |
| 7 |
in the Labrador Sea (Nx*Ny*Nz = 20*16*23). |
| 8 |
The seaice model is forced by monthly-mean atmospheric state. |
| 9 |
The lower boundary condition is SST, also provided as monthly-mean |
| 10 |
fields (rather than being computed by an ocean component). |
| 11 |
In its current form the seaice model is thermodynamic-only, i.e. |
| 12 |
#undef SEAICE_ALLOW_DYNAMICS |
| 13 |
The estimation problem consists of a least-squares fit of |
| 14 |
daily-mean modeled vs. observed seaice fractional area. |
| 15 |
Observations are from satellite-borne sounding microwave radiometers. |
| 16 |
|
| 17 |
The setup consists of 2 parts: |
| 18 |
1. the forward/adjoint model component which computes the gradient of |
| 19 |
the model vs. data mistif costfunction w.r.t. to the time-varying |
| 20 |
atmospheric state. |
| 21 |
2. the optimization/linesearch routine that uses the gradient information |
| 22 |
to provide an improved guess of the control vector. |
| 23 |
|
| 24 |
Instructions to configure: |
| 25 |
|
| 26 |
0. Preparation |
| 27 |
-------------- |
| 28 |
# from model base dir., e.g. MITgcm/ |
| 29 |
set basedir = `pwd` |
| 30 |
cd ${basedir}/verification/lab_sea/ |
| 31 |
mkdir exe |
| 32 |
mkdir exe/optim |
| 33 |
|
| 34 |
1. forward/adjoint model component |
| 35 |
---------------------------------- |
| 36 |
cd ${basedir}/verification/lab_sea/build/ |
| 37 |
../../../tools/genmake2 -mods ../code_ad_seaice_only |
| 38 |
make depend |
| 39 |
make adall |
| 40 |
|
| 41 |
2. optimization/linesearch routines |
| 42 |
----------------------------------- |
| 43 |
# This part needs more attention: |
| 44 |
# Compiler options in the Makefile are hard-coded, |
| 45 |
# make sure you choose the right ones for your platform |
| 46 |
# |
| 47 |
# Assume you are in model base dir., e.g. MITgcm/ |
| 48 |
# |
| 49 |
# A: lsopt |
| 50 |
# -------- |
| 51 |
cd ${basedir}/lsopt/ |
| 52 |
# edit Makefile to choose right compiler |
| 53 |
make |
| 54 |
# |
| 55 |
# B: optim |
| 56 |
# -------- |
| 57 |
cd ${basedir}/optim/ |
| 58 |
# Makefile in ${basedir}/optim/ needs 4 adjustments: |
| 59 |
# 1: -DMAX_INDEPEND=1000000 |
| 60 |
# make sure this number is >= number of control elements |
| 61 |
# 2: Make sure you've chosen the right compiler for your platform |
| 62 |
# 3: Make sure you've chosen the right INCLUDEDIRS for the header files, |
| 63 |
# in this case it is |
| 64 |
# -I../verification/lab_sea/build/ |
| 65 |
# 4: Make sure you've got the right set of BLAS libraries |
| 66 |
# some are provided in lsopt/ |
| 67 |
# See the README file in lsopt/ on how to get platform-optimized |
| 68 |
# BLAS libraries, and how to best install them. |
| 69 |
# If you're done with all these, just type |
| 70 |
make |
| 71 |
# |
| 72 |
# C: move executable and parameter files to executable directory |
| 73 |
# ------- |
| 74 |
cp optim.x ../verification/lab_sea/exe/optim/. |
| 75 |
cp ../lsopt/data.ecco ../lsopt/data.ctrl ../verification/lab_sea/exe/optim/. |
| 76 |
|
| 77 |
3. Run state estimation |
| 78 |
----------------------- |
| 79 |
cd ${basedir}/verification/lab_sea/exe/optim |
| 80 |
ln -s ../data.optim . |
| 81 |
cd ${basedir}/verification/lab_sea/exe |
| 82 |
ln -s ../build/mitgcmuv_ad . |
| 83 |
ln -s ../input_ad_seaice_only/* . |
| 84 |
ln -s ../input/*.labsea1979 . |
| 85 |
# |
| 86 |
./run_estim.csh & |
Parent Directory
| 
Revision Log
| 
Revision Graph