TOA and surface cloud radiative kernels calculated with RRTM

These cloud radiative kernels (CRK) are calculated with RRTM, with meteorological variables from ERA-interim as inputs. The file format is netcdf4, and was created by python. To read these files, any software supporting netcdf4 can be used. 

    Longwave cloud radiative effect at surface is primarily decided by cloud base properties, while top-of-atmosphere (TOA) cloud radiative effect is primarily decided by cloud top properties, so the standard version of surface CRK is a function of latitude, longitude, month, cloud optical thickness (τ) and cloud base pressure (CBP), and the TOA CRK is a function of latitude, longitude, month,τ and cloud top pressure (CTP).  Considering that the cloud property histograms provided by climate models are functions of CTP instead of CBP at present, we created a set of surface CRK on CTP-τ cloud fraction histograms using the statistical relationship between CTP, CBP and τ from collocated CloudSat-MODIS observations.

    There are five individual files. "SFC_CRK_CBP2.nc" is for surface CRK on CBP-τ histograms, "SFC_CRK_CTP2.nc" is for surface CRK on CTP-τ histograms, and "TOA_CRK_CTP2.nc" is for TOA CRK on CTP-τ histograms. The atmospheric CRK can be calculated as the difference between "TOA_CRK_CTP2.nc" and "SFC_CRK_CTP2.nc". The other two files denote separate CRK for ice and liquid clouds. 

    Notes: (1) this version is used in the submitted draft for publication, and it might be renewed in the future.

              (2) To avoid NAN values, the value of CBP/CTP is set to be the lowest level near surface for cloud bins with CBP/CTP greater than surface air temperature (cloud base or top is below surface).