cloud fraction of offline CICE

[hspark1@gmail_com]

I would like to know the exact definition of 'cloud fraction (cldf)' in CICE 5.1.2. Does it imply total cloud fraction?  I've prepared CICE input data from ERA-interim. Cloud fraction (cldf) is an input data for calculating the surface radiative fluxes (both shorwave and longwave). I used the ERA-interim's total cloud fraction, which is mostly higher than 0.9 over the Arctic. Then, I've integrated CICE-slab ocean model for 16 years (from year 1979 to 1995) and found that sea ice becomes thicker than 10 meters in 16 years.  This is probably because ERA-interim's total cloud fraction is very large in the Arctic (mostly larger than 0.9), which weakens incoming shortwave radiation to the surface. In July, shortwave to the surface is only around 150-200 W/m2, which is far small than the usual (250-300 W/m2). The summer shortwave radiation at the surface and sea ice thickness were simulated reasonably well in the case when I used ERA-interim's high cloud fraction, which is around 0.6-0.7 (smaller than the total cloud fraction). Am I supposed to use high cloud cover for the CICE cloud fraction, cldf?

 

[dbailey]

The variable cldf is total cloud fraction. This is used to modify the incoming LW and SW fluxes. See the subroutines longwave_parkinson_washington, longwave_rosati_miyakoda, and compute_shortwave in the ice_forcing.F90 module. I would suggest choosing different formulae for these if this is not what you want. Note I've moved your question to the appropriate group here.

 

[dbailey]

The variable cldf is total cloud fraction. This is used to modify the incoming LW and SW fluxes. See the subroutines longwave_parkinson_washington, longwave_rosati_miyakoda, and compute_shortwave in the ice_forcing.F90 module. I would suggest choosing different formulae for these if this is not what you want. Note I've moved your question to the appropriate group here.

 

[dbailey]

The variable cldf is total cloud fraction. This is used to modify the incoming LW and SW fluxes. See the subroutines longwave_parkinson_washington, longwave_rosati_miyakoda, and compute_shortwave in the ice_forcing.F90 module. I would suggest choosing different formulae for these if this is not what you want. Note I've moved your question to the appropriate group here.

 

[dbailey]

The variable cldf is total cloud fraction. This is used to modify the incoming LW and SW fluxes. See the subroutines longwave_parkinson_washington, longwave_rosati_miyakoda, and compute_shortwave in the ice_forcing.F90 module. I would suggest choosing different formulae for these if this is not what you want. Note I've moved your question to the appropriate group here.

 

[dbailey]

The variable cldf is total cloud fraction. This is used to modify the incoming LW and SW fluxes. See the subroutines longwave_parkinson_washington, longwave_rosati_miyakoda, and compute_shortwave in the ice_forcing.F90 module. I would suggest choosing different formulae for these if this is not what you want. Note I've moved your question to the appropriate group here.

 

[eclare@lanl_gov]

I'll add a little more to this:I’m not surprised that the simulation does weird things, since it was tuned for the other forcing.  There are a couple of things that I think you should try.  First, try to find out if there are biases in the forcing data compared with observations, such as the total cloud fraction.  Is it too big or small for parts of the year?  I’m not familiar with it, so this would be helpful information for me.  Then, try switching the longwave calculation in ice_forcing.F90 from the Rosati and Miyakoda formulation to Parkinson and Washington.  The one we currently have as the default has some problems (this is on our list to fix when we create a new forcing data set for the model).  I think the forcing that we tuned the model for is biased, so you might just need to retune.

 

[eclare@lanl_gov]

I'll add a little more to this:I’m not surprised that the simulation does weird things, since it was tuned for the other forcing.  There are a couple of things that I think you should try.  First, try to find out if there are biases in the forcing data compared with observations, such as the total cloud fraction.  Is it too big or small for parts of the year?  I’m not familiar with it, so this would be helpful information for me.  Then, try switching the longwave calculation in ice_forcing.F90 from the Rosati and Miyakoda formulation to Parkinson and Washington.  The one we currently have as the default has some problems (this is on our list to fix when we create a new forcing data set for the model).  I think the forcing that we tuned the model for is biased, so you might just need to retune.

 

[eclare@lanl_gov]

I'll add a little more to this:I’m not surprised that the simulation does weird things, since it was tuned for the other forcing.  There are a couple of things that I think you should try.  First, try to find out if there are biases in the forcing data compared with observations, such as the total cloud fraction.  Is it too big or small for parts of the year?  I’m not familiar with it, so this would be helpful information for me.  Then, try switching the longwave calculation in ice_forcing.F90 from the Rosati and Miyakoda formulation to Parkinson and Washington.  The one we currently have as the default has some problems (this is on our list to fix when we create a new forcing data set for the model).  I think the forcing that we tuned the model for is biased, so you might just need to retune.

 

[eclare@lanl_gov]

I'll add a little more to this:I’m not surprised that the simulation does weird things, since it was tuned for the other forcing.  There are a couple of things that I think you should try.  First, try to find out if there are biases in the forcing data compared with observations, such as the total cloud fraction.  Is it too big or small for parts of the year?  I’m not familiar with it, so this would be helpful information for me.  Then, try switching the longwave calculation in ice_forcing.F90 from the Rosati and Miyakoda formulation to Parkinson and Washington.  The one we currently have as the default has some problems (this is on our list to fix when we create a new forcing data set for the model).  I think the forcing that we tuned the model for is biased, so you might just need to retune.

 

[eclare@lanl_gov]

I'll add a little more to this:I’m not surprised that the simulation does weird things, since it was tuned for the other forcing.  There are a couple of things that I think you should try.  First, try to find out if there are biases in the forcing data compared with observations, such as the total cloud fraction.  Is it too big or small for parts of the year?  I’m not familiar with it, so this would be helpful information for me.  Then, try switching the longwave calculation in ice_forcing.F90 from the Rosati and Miyakoda formulation to Parkinson and Washington.  The one we currently have as the default has some problems (this is on our list to fix when we create a new forcing data set for the model).  I think the forcing that we tuned the model for is biased, so you might just need to retune.

 

[hspark1@gmail_com]

Thank you for your reply. Yes, ERA-interim’s cloud fraction is generally large throughout the season.Is there a way to prescribe downward longwave and shortwave radiation to the surface directly into CICE? Personally I think it is more convenient to prescribe surface longwave & shortwave radiation data provided by reanalysis (e.g. ERA-interim, MERRA, NCEP) into CICE.    In this version of CICE model, cloud fraction is climatologically fixed, whereas time-varying temperature & water vapor are used to calculate longwave & shortwave radiation to the surface. 

 

[hspark1@gmail_com]

Thank you for your reply. Yes, ERA-interim’s cloud fraction is generally large throughout the season.Is there a way to prescribe downward longwave and shortwave radiation to the surface directly into CICE? Personally I think it is more convenient to prescribe surface longwave & shortwave radiation data provided by reanalysis (e.g. ERA-interim, MERRA, NCEP) into CICE.    In this version of CICE model, cloud fraction is climatologically fixed, whereas time-varying temperature & water vapor are used to calculate longwave & shortwave radiation to the surface. 

 

[hspark1@gmail_com]

Thank you for your reply. Yes, ERA-interim’s cloud fraction is generally large throughout the season.Is there a way to prescribe downward longwave and shortwave radiation to the surface directly into CICE? Personally I think it is more convenient to prescribe surface longwave & shortwave radiation data provided by reanalysis (e.g. ERA-interim, MERRA, NCEP) into CICE.    In this version of CICE model, cloud fraction is climatologically fixed, whereas time-varying temperature & water vapor are used to calculate longwave & shortwave radiation to the surface. 

 

[hspark1@gmail_com]

Thank you for your reply. Yes, ERA-interim’s cloud fraction is generally large throughout the season.Is there a way to prescribe downward longwave and shortwave radiation to the surface directly into CICE? Personally I think it is more convenient to prescribe surface longwave & shortwave radiation data provided by reanalysis (e.g. ERA-interim, MERRA, NCEP) into CICE.    In this version of CICE model, cloud fraction is climatologically fixed, whereas time-varying temperature & water vapor are used to calculate longwave & shortwave radiation to the surface. 

 

[hspark1@gmail_com]

Thank you for your reply. Yes, ERA-interim’s cloud fraction is generally large throughout the season.Is there a way to prescribe downward longwave and shortwave radiation to the surface directly into CICE? Personally I think it is more convenient to prescribe surface longwave & shortwave radiation data provided by reanalysis (e.g. ERA-interim, MERRA, NCEP) into CICE.    In this version of CICE model, cloud fraction is climatologically fixed, whereas time-varying temperature & water vapor are used to calculate longwave & shortwave radiation to the surface. 

 

[eclare@lanl_gov]

Yes, you can certainly read in shortwave and longwave radiation and use them directly instead of calculating them.  Use the subroutines in ice_forcing.F90 for guidance and write a subroutine that does what you need.

 

[eclare@lanl_gov]

Yes, you can certainly read in shortwave and longwave radiation and use them directly instead of calculating them.  Use the subroutines in ice_forcing.F90 for guidance and write a subroutine that does what you need.

 

[eclare@lanl_gov]

Yes, you can certainly read in shortwave and longwave radiation and use them directly instead of calculating them.  Use the subroutines in ice_forcing.F90 for guidance and write a subroutine that does what you need.

 

[eclare@lanl_gov]

Yes, you can certainly read in shortwave and longwave radiation and use them directly instead of calculating them.  Use the subroutines in ice_forcing.F90 for guidance and write a subroutine that does what you need.