Dr. :
I computing rayleigh radiance using a code that is similar to "rayrad.pro" https://oceancolor.gsfc.nasa.gov (/cgi/idllibrary.cgi?dir=atmocorr). However, the result is different with seadas product "Lr". the scatter in the following attachment file shows difference at eight bands. I find，after analysis in ENVI，the difference increase from center to edge of the scan line. the figure and their description and my code are as follows. Does anyone know the reason?
# * coding: utf8 *
import numpy as np
from netCDF4 import Dataset
from scipy import interpolate
import general
def rayleigh(rayleigh_lut_path=None, sza=None, vza=None, saa=None, vaa=None, windspeed=None, pressure=None, F0=None):
"""
Args:
rayleigh_lut_path ():
sza (): solar zenith
vza (): sensor zenith
vaa (): sensor azimuth
saa (): solar azimuth
windspeed (): wind speed
pressure (): atmospheric pressure
Returns:
Lr_i
"""
reaa=vaa180saa
reaa[reaa<180]=reaa[reaa<180]+360
reaa[reaa > 180] = reaa[reaa > 180]  360
rayleigh_lut = general.get_filelist(rayleigh_lut_path, 'rayleigh', 'iqu.hdf')
windspeed[windspeed > 30] = 30
sza[sza > 88] = 88
mu0 = np.cos(sza / 180 * np.pi)
mu = np.cos(vza / 180 * np.pi)
airmass = 1 / mu0 + 1 / mu
Taur = np.zeros(shape=(rayleigh_lut.__len__()))
Lr_i = np.zeros(shape=(sza.shape[0], sza.shape[1], rayleigh_lut.__len__()))
for i in range(rayleigh_lut.__len__()):
rayDtset = Dataset(rayleigh_lut[i])
taur = rayDtset.variables['taur'][:]
depol = rayDtset.variables['depol'][:]
senz = rayDtset.variables['senz'][:]
solz = rayDtset.variables['solz'][:]
wind = rayDtset.variables['wind'][:]
# sigma = rayDtset.variables['sigma'][:]
i_ray = rayDtset.variables['i_ray'][:]
# q_ray = rayDtset.variables['q_ray'][:]
# u_ray = rayDtset.variables['u_ray'][:]
Taur[i] = taur
Norder0 = np.zeros_like(sza)
Norder1 = np.ones_like(vza)
Norder2 = Norder0 + 2
ray_i0 = interpolate.interpn(
(wind.reshape(1), solz.reshape(1), np.arange(3).reshape(1), senz.reshape(1)), i_ray,
np.stack([windspeed, sza, Norder0, vza], axis=2))
ray_i1 = interpolate.interpn(
(wind.reshape(1), solz.reshape(1), np.arange(3).reshape(1), senz.reshape(1)), i_ray,
np.stack([windspeed, sza, Norder1, vza], axis=2))
ray_i2 = interpolate.interpn(
(wind.reshape(1), solz.reshape(1), np.arange(3).reshape(1), senz.reshape(1)), i_ray,
np.stack([windspeed, sza, Norder2, vza], axis=2))
ray_i = ray_i0 + ray_i1 * np.cos(reaa / 180 * np.pi) + ray_i2 * np.cos(2 * reaa / 180 * np.pi)
# from Wang menghua ;is from Seadas
p0 = 1013.25 # hpa
x = ((0.6543  1.608 * taur) + (0.8192  1.2541 * taur) * np.log(airmass)) * taur * airmass
fac = ((1.0  np.exp(x * pressure / p0)) / (1.0  np.exp(x)))
Lr_i[:, :, i] = ray_i * fac * F0[i]
return Lr_i
rayleigh radiance computing problem
rayleigh radiance computing problem
 Attachments

 the scatter between my Lr and seadas Lr
 Lr_Lr_ac2_300dpi.png (169.04 KiB) Not viewed yet

 the difference of spectrum at left of scan line between my Lr and seadas Lr
 left.png (58.28 KiB) Not viewed yet

 the difference of spectrum at center of scan line between my Lr and seadas Lr
 center.png (60.38 KiB) Not viewed yet
Tags:
Re: rayleigh radiance computing problem
Hi, the Rayleigh table coefficients i_ray have the dimensions of wave mean square slope *sigma* and not *windspeed*. Before the interpolation convert windspeed to sigma using sigma = sqrt(0.00534*WS).