The Viewing Geometry#
The viewing geometry can be thought of as specifying what output quantities we need from the model. These can be things like radiances at the top of the atmosphere, ground, or anywhere inbetween. Or they can be fluxes that are automatically integrated by the model.
Radiance Output#
We specify radiance output through a list of single positions and look vectors. These are stored inside the
sasktran2.ViewingGeometry object.
import sasktran2 as sk
viewing_geo = sk.ViewingGeometry()
Rather than specifying look directions and positions directly, they are specified through objects that we
colloquially refer to as Viewing Policies. One example of a viewing policy is the sasktran2.TangentAltitudeSolar
object, which defines a position and viewing direction based off an observer altitude, as well as various parameters
that are defined at the tangent point of the measurement. For example,
ray = sk.TangentAltitudeSolar(tangent_altitude_m=10000,
relative_azimuth=0,
observer_altitude_m=200000,
cos_sza=0.6)
To add it to our overall container, we can do,
viewing_geo.add_ray(ray)
And we can look at all of the rays in our container
viewing_geo.observer_rays
[Tangent Viewing Ray: tangentaltitude: 10000, relative_azimuth_angle: 0, observeraltitude: 200000, cos_sza: 0.6]
Viewing Policies#
Defines a viewing ray from the observer altitude, and tangent point parameters. |
|
|
Defines a viewing ray that is looking at the ground from angles defined at the ground location. Note that |
Defines a viewing ray that is defined at a location defined from the solar angles. |
Working with Real Measurements#
The Solar Geometry Handler#
sasktran2 contains utilities to compute the solar angles automatically for a given location and time
using the astropy package. To use this functionality you must have astropy installed.
For example, to calculate the (solar zenith angle, solar azimuth angle) at a given location we can do,
import pandas as pd
solar = sk.solar.SolarGeometryHandlerAstropy()
print(solar.target_solar_angles(
latitude=20,
longitude=-100,
altitude=0,
time=pd.Timestamp("2024-11-12 20:00:00")
))
(np.float64(44.66518717640884), np.float64(213.30155059850907))
Solar azimuth angles are always measured from true north, with 90 degrees pointing in the east direction.
Note
The solar handler specifies angles in degrees, whereas most other aspects of SASKTRAN2 use radians.
Converting from ECEF Coordinates#
In the original sasktran1 package, all rays were specified as a triplet of
(observer position, local look vector, time) in Earth Centered Earth Fixed coordinates.
Here we have pre-computed the position and look vector of a satellite directly above Saskatoon, Canada, and we will specify a local time of noon (18 UTC).
import numpy as np
observer = np.array([-1158730.59368676, -3875262.18406142, 5170772.5134034 ])
look_vector = np.array([ 0.17579194, 0.58791911, -0.78958743])
time = pd.Timestamp("2024-11-12 18:00:00")
We can then convert the ray to one sasktran2 recognizes,
ray = sk.viewinggeo.ecef_to_sasktran2_ray(
observer=observer,
look_vector=look_vector,
time=time,
solar_handler=solar
)
print(ray)
Ground Viewing Ray: cos_sza: 0.3258368629858771, relative_azimuth_angle: 2.9184752483795724, cos_viewing_zenith: -0.9999999978430802, observer_altitude_m: 200000.00000000442
Note here we re-used our solar handler from the previous section.
Flux Output#
Fluxes can be internally calcluated in sasktran2 by specifying a location to calculate
the fluxes at with the sasktran2.ViewingGeometry.add_flux_observer() method.
Note
Currently flux output is only supported when using the DiscreteOrdinates source, in PlaneParallel mode, i.e.
config = sk.Config()
config.single_scatter_source = sk.SingleScatterSource.DiscreteOrdinates
config.multiple_scatter_source = sk.MultipleScatterSource.DiscreteOrdinates
model_geometry = sk.Geometry1D(...,
geometry_type=sk.GeometryType.PlaneParallel)
Flux Observer Policies#
|
Defines a flux observer that is defined at a location defined from the solar angles. |