SeaIceRT is a python interface for CESM3 Delta Eddington radiative transfer for sea ice. The radiative transfer code is written in Fortran 77. The python wrapper allows the sea ice parameters to be set, the code run and output returned.
The seaicert directory contains the class SeaIceRT, which
initializes parameters and call the code.
The fortran code is in the 1D_SIR_DE directory. The main fortran
routine has been modified to allow the python wrapper to set and get
parameters. The main fortran calling routine crm has been changed
from a Fortan 77 PROGRAM to a SUBROUTINE. This allows the fortran
code to be built as a library, which is called from python using the
ctypes package.
A full description of the original model can be found here
The easiest way to install the wrapper and model is using git.
git clone git@github.com:andypbarrett/seaice_radiative_transfer.git
cd seaice_radiative_transfer
I strongly suggest creating a new environment for running and creating the model. This will ensure that the dependencies are installed.
conda env create -f environment.yml
or
mamba env create -f environment.yml
This will create a new environment called seaice_radiative_transfer
Start the environment using
conda activate seaice_radiative_transfer
or
conda activate seaice_radiative_transfer
The radiative transfer model is written in fortran. The source code must be compile to create a dynamic library containing the model. There is a makefile in 1D_dE_CCSM.
You will need a fortran compiler. I've used gfortran from https://gcc.gnu.org/wiki/GFortran.
cd 1D_dE_CCSM
make
This will create libcrm.so or libcrm.dylib, if you are on a Mac.
Note
The fortran code has been written and compiled on a Ubuntu Linux machine using gfortran. The compiler flags in the makefile work for this architecture. If you have MacOS and get compile >errors, you may need to add -fallow-argument-mismatch to the compiler switches. You might need to play around with other compiler switches.
Warning
The fortran compile step has not be tried on a Windows machine.
Running the model can be done from a python IDE, either python or ipython.
(seaice_radiative_transfer) nsidc-abarrett-442:seaice_radiative_transfer$ cd seaicert/
(seaice_radiative_transfer) nsidc-abarrett-442:seaicert$ ipython
Python 3.7.6 | packaged by conda-forge | (default, Mar 5 2020, 15:27:18)
Type 'copyright', 'credits' or 'license' for more information
IPython 7.17.0 -- An enhanced Interactive Python. Type '?' for help.
In [1]: from ccsm3_sir_de import SeaIceRT
In [2]: model = SeaIceRT()
In [3]: model.run()
In [4]: model.print_results()
----------------------------------------------------------------------
CCSM3 Sea Ice Delta Eddington calculation
----------------------------------------------------------------------
----------------------------------------------------------------------
Visible and near-ir direct and diffuse albedos
Visible: 0.2 to 0.7 micrometers
Near-IR: 0.7 to 5.0 micrometers
----------------------------------------------------------------------
Albedo shortwave direct: 0.17
Albedo shortwave diffuse: 0.19
Albedo longwave direct: 0.06
Albedo longwave diffuse: 0.06
----------------------------------------------------------------------
Surface ansorption and Albedos
----------------------------------------------------------------------
Visible solar absorbed by ocean: 27.5656681060791
Near-IR absorbed by ocean: 0.0
----------------------------------------------------------------------
Surface absorption ad albedos
----------------------------------------------------------------------
Solar vs direct surface irradiance: 0.12 Wm-2
----------------------------------------------------------------------
Snow/Sea ice transmitted flux (Tr fraction) and absorption (Q Wm-2)
----------------------------------------------------------------------
Level depth Tr_vs Q_vs Tr_ni Q_ni Q_total
----------------------------------------------------------------------
0 surface 26.88 68.01 94.89
0.000 1.0000 1.0000
1 pond 12.76 67.06 79.82
0.250 0.9494 0.0130
2 pond 12.08 0.93 13.01
0.500 0.8625 0.0002
3 ice 2.05 0.02 2.07
0.050 0.7888 0.0000
4 ice 10.84 0.00 10.84
0.375 0.6228 0.0000
5 ice 9.41 0.00 9.41
0.750 0.4730 0.0000
6 ice 6.78 0.00 6.78
1.125 0.3551 0.0000
7 ice 4.61 0.00 4.61
1.500 0.2612 0.0000
8 ocean 27.57 0.00 27.57
Model parameters can be changed simply by modify model parameters, for example:
In [11]: model.pond_depth
Out[11]: 0.5
In [12]: model.pond_depth = 1.
A list of model parameters is given below.
- :day_of_year: day of year, 1..365, where day 1 = January 1
- :latitude: latitude (-90 to 90) (test=80.)
- :surface_pressure: Surface pressure in mb (test=1008 mb)
- :co2_volume_mixing_ratio: CO2 volume mixing ratio (test 3.7e-04)
- :surface_air_temperature: Surface air temperature (K) (test=273.16 K)
- :ground_temperature: Surface skin temperature (K) (test=273.17 K)
- :snow_depth: Physical snow depth in meters (test=0 m)
- :snow_density: Snow density (kg/m3) (test=330 kg/m3)
- :snow_grain_radius: Snow grain radius in microns (um) (test=50. um) -:pond_depth: Physical pond depth in meters (test=0.5 m)
- :pond_tuning_parameter: Pond tuning parameter in standard deviations (test=-1.)
- :sea_ice_thickness: Physical ice thickness in meters (test=1.5 m)
- :sea_ice_tuning_parameter: Sea ice tuning parameter in standard deviations (test=0.)
- :level: number id of level 1..18
- :pressure: Pressure in mb
- :air_temperature: air temperature in Kelvin
- :water_vapor_mixing_ratio: Water vapour mixing ration (g/g)
- :ozone_mixing_ratio: Ozone mixing ration (g/g)
- :cloud_cover: Cloud cover - non-dimension 0.-1.
- :cloud_liquid_water_path: Cloud liquid water path (g/m2)
A sensitivity analsysis of selected parameters can be found here
If you need to add new packages:
- hand-edit the
environment.ymlfile. conda env updateto update the current environment with the new entry.conda env export > environment-lock.ymlto save the working version of the environment.- finally commit these changes.