L2d products
#1
It is now possible to derive L2d products out of L2b products using the skimulator.
Using the same parameter files as the one to compute L2b, run the following command:

Code:
skimul2d [yourparameterfile]

The following keys should be added to the parameter file:

Code:
# -----------------------#
# L2D computation
# -----------------------#
# Length resolution to select neighbors (in km):
resol_spatial_l2d = 50
# Temporal resolution to select neighbors (in days):
resol_temporal_l2d = 8
# Grid resolution for l2d (lat, lon) grid (in degrees):
posting_l2d = (0.1, 0.1)
# Time domain: (start_time, end_time, dtime) in days:
time_domain = (5, 25, 1)
# Spatial domain (lon_min, lon_max, lat_min, lat_max):
spatial_domain = [lonmin, lonmax, latmin, latmax]
# List of variables to be interpolated on the grid:
list_input_var_l2d = {'ucur': ['ucur', 'cur', 0], 'vcur': ['vcur', 'cur', 0]}
#2
A new option has been added to the parameter file.
You can add to the name of the l2d file a specific attribute to refer your test case

Code:
# config_l2d:
config_l2d = 'my_test'

The spatial and temporal filtering are now a function of the latitude.
A fit in cos(lat) has been implemented so that the default filtering length is 90km, 10 days at the Equator, 40km, 14 days at 60º.
The parameters resol_temporal_l2d and resol_spatial_l2d are now multiplication factors of the default filtering length
In the parameter file:

Code:
# Factor for the spatial length resolution to select neighbors (in km):
# if this factor is 1, length is 90km at the Equator, 40km at 60º.
resol_spatial_l2d = 1
# Factor for the temporal resolution to select neighbors (in days):
# if this factor is 1, length is 10 days at the Equator, 14 days at 60º.
resol_temporal_l2d = 1


Forum Jump: