Emil M.
Constantinescu
Computational Mathematician
emconsta[at]mcs.anl.gov
Office #: 2151
• LANS
• Mathematics and Computer Science
• Argonne National Laboratory

• Scientist at large at CASE
• University of Chicago

Cloud processes in global climate models

Funding: Internal LDRD, Argonne National Laboratory.

Synopsis: Cloud processes in global climate models (GCMs) are highly parameterized and often tuned to obtain the Earth’s radiative balance and global mean temperature. As a result, clouds constitute the single-largest uncertainty in terms of radiative forcings and climate feedbacks. Improving the cloud processes and especially interactions with aerosols in high-resolution climate models is definitely a priority for the coming decade.

NEWS:

http://energy.gov/eere/articles/energy-department-announces-25-million-improve-wind-forecasting

Online calibration results

Test 2D-Guassian

Cloud convection ()

MAP output

Median output

Mean output

Configure & build PETSc with debug options (pgi on the Fusion machine):

git clone git@bitbucket.org:petsc/petsc.git petsc-dev

export PETSC_DIR=%path to petsc%
export PETSC_ARCH=arch-mac

./configure --PETSC_ARCH=arch-linux2-c-debug --with-mpi-dir=/soft/mvapich2/1.4.1-pgi-9.0.4 --with-shared-libraries --download-sundials=yes --download-netcdf=yes --download-hdf5=yes --download-metis=yes --download-netcdf-shared=1

make PETSC_DIR=$PWD PETSC_ARCH=arch-linux2-c-debug all test

<move to netcdf-fortran-xx download it, and read instructions>

export FC=$PETSC_DIR/$PETSC_ARCH/bin/mpif90

export CC=$PETSC_DIR/$PETSC_ARCH/bin/mpicc

export LD_LIBRARY_PATH=$PETSC_DIR/$PETSC_ARCH/lib:$LD_LIBRARY_PATH

CPPFLAGS="-DgFortran -I$PETSC_DIR/$PETSC_ARCH/include" LDFLAGS=-L$PETSC_DIR/$PETSC_ARCH/lib ./configure --prefix=$PETSC_DIR/$PETSC_ARCH

make
make check
make install