Welcome to pyrad’s documentation!¶
pyrad¶
Pyrad is a simple (one-dimensional), pure python, all-sky atmospheric radiation package.
Gas absorption coefficients can be calculated by:
from pyrad.optics.gas import Gas
from pyrad.utils.grids import UniformGrid1D
import matplotlib.pyplot as plt
from pyrad.lbl.hitran import Voigt
grid = UniformGrid1D(600., 900., 0.01)
gas = Gas(formula="CO2", line_profile=Voigt())
k = gas.absorption_coefficient(temperature=299.7, pressure=101300.,
volume_mixing_ratio=.02595108, spectral_grid=grid.points)
plt.plot(grid.points, k)
plt.title("CO2 Absorption Spectrum")
plt.xlabel("wavenumber [cm-1]")
plt.ylabel("absorption coefficient [m2].")
plt.savefig("gas-optics.png")
By default, the above code will download the necessary molecular line and total partition function data from the web. This can take a significant amount of time, especially if the Gas objects are created often. To retify this, I recommend creating local SQLite databases, then re-using when creating Gas objects:
from pyrad.lbl.tips import TotalPartitionFunction
from pyrad.lbl.hitran import Hitran, Voigt
from pyrad.optics.gas import Gas
for x in ["H2O", "CO2", "O3"]:
Hitran(x, Voigt()).create_database("hitran.sqlite")
TotalPartitionFunction(x).create_database("tips-2017.sqlite")
gas = Gas("H2O", hitran_database="hitran.sqlite", tips_database="tips-2017.sqlite")
gas = Gas("CO2", hitran_database="hitran.sqlite", tips_database="tips-2017.sqlite")
gas = Gas("O3", hitran_database="hitran.sqlite", tips_database="tips-2017.sqlite")
Clouds are generated in a stochastic fashion (typically found in GCMs):
import matplotlib.pyplot as plt
from numpy import array, zeros
from pyrad.optics.clouds.stochastic import overlap_parameter, TotalWaterPDF
altitude = array([1., 2., 3.])
cloud_fraction = array([.9, .8, .75])
liquid_water_content = array([.3, .3, .3])
ice_water_content = array([.2, .2, .2])
num_subcolumns = 10
lwc, iwc = zeros((num_subcolumns, altitude.size)), zeros((num_subcolumns, altitude.size))
overlap = overlap_parameter(altitude, scale_length=2.)
for i in range(num_subcolumns):
# Repeat sampling in each column.
lwc[i, :], iwc[i, :] = TotalWaterPDF().sample_condensate(cloud_fraction, liquid_water_content,
ice_water_content, overlap=overlap)
for condensate, name in zip([lwc, iwc], ["Liquid", "Ice"]):
plt.clf()
plt.imshow(condensate.transpose(), cmap="Greys", interpolation="none")
plt.title("{} water content".format(name))
plt.xlabel("sub-column")
plt.xticks([], [])
plt.ylabel("layer")
plt.yticks([], [])
colorbar = plt.colorbar()
colorbar.set_label("g m-2")
plt.savefig("{}-stochastic-clouds.png".format(name).lower())
and their optics are calculated using standard look-up table parameterizations:
from os.path import join
from pyrad.optics.clouds.ice import IceCloudOptics
from pyrad.optics.clouds.liquid import LiquidCloudOptics
from pyrad.utils.grids import UniformGrid1D
ice_clouds = IceCloudOptics(join("pyrad_data", "clouds", "chou_suarez.nc"))
ice_cloud_optics = ice_clouds.optics(iwc=0.2, ice_particle_size=10,
spectral_grid=UniformGrid1D(600., 900., .01),
mode="longwave")
liquid_clouds = LiquidCloudOptics(join("pyrad_data", "clouds", "hu_stamnes.nc"))
liquid_cloud_optics = liquid_clouds.optics(lwc=0.2, liquid_droplet_radius=10.,
spectral_grid=UniformGrid1D(500., 800., .1))
Aerosol optics are also calculated using a GCM parameterization:
from os.path import join
from pyrad.optics.aerosols import AerosolOptics
from pyrad.utils.grids import UniformGrid1D
sulfate = AerosolOptics(join("pyrad_data", "aerosols", "sulfate_optics.nc"))
aerosol_optics = sulfate.optics(concentration=0.5, grid=UniformGrid1D(1., 500., 0.1),
relative_humidity=50, mixture=50)