A Unified Radiative Transfer Model: Microwave to Infrared

1
A Unified Radiative Transfer Model Microwave to
Infrared

• Tom Greenwald
• CIMSS

2
Motivation

• Why use one RT model across the thermal spectrum
  for multiple scattering calculations?
• Physics is the same
• Consistency in calculations
• Reduces complexity in applications such as
  multi-sensor retrievals of atmospheric/cloud
  properties and direct radiance data assimilation
• This presentation will examine monochromatic
  adjoint sensitivities and contrast them in clear
  sky and cloudy conditions

3
Description of Modeling System

• Multiple scattering RT model Successive Order of
  Interaction (SOI) method (Heidinger et al. 2005
  ODell et al. 2005)
• Fast and accurate (lt 1 K over most conditions)
• Can accommodate any number of cloud layers
• Adaptive variable number of streams (angular
  resolution). Other fast models do not have this
  flexibility since angular resolution is fixed
• Incorporated into CRTM and tested
• LBL gas absorption models
• Infrared LBLRTM v8.4
• Microwave Liebe MPM v. 1989

4
Description of Modeling System contd.

• Cloud single-scattering properties
• Infrared Tables of properties for mixture of
  habits (Baum et al. 2005) and individual habits
  (bullet rosettes, columns, plates, aggregates)
  based on rigorous techniques (Yang) water
  spheres based on Anomalous Diffraction Theory
• Microwave Tables of properties for bullet
  rosettes, plates, columns and ice spheres based
  on DDA calculations (Evans)
• Adjoint models built for all forward models
  except LBLRTM

5
Experimental Setup

• Tropical mean temperature/humidity profiles
• Cloud properties

6
Experimental Setup contd.

• Surface properties
• Microwave emissivity Fresnel equation
  permittivity model from Stogryn et al. (1995)
• Infrared emissivity Fixed 0.98
• Geometry
• Zenith angle 0o

7
Forward Results
8
Adjoint Sensitivity Analysis
Jacobian
Adjoint forcing

• Definition of adjoint
• Forward vs. adjoint sensitivity Adjoint is
  much faster and gives more accurate derivatives
• Non-dimensional relative sensitivity (Zou et
  al. 1993)

9
Non-dimensional Tb Sensitivity to Atmospheric
Temperature (Thermal Source only)
Clear sky
Cloudy
10
Non-dimensional Tb Sensitivity to Water Vapor
Mixing Ratio
Clear sky
Cloudy
11
Absolute Tb Sensitivity to Water Vapor Mixing
Ratio (K/g/kg)
Clear sky
Cloudy
12
Non-dimensional Tb Sensitivity to LWP and IWP
IWP
LWP
13
Summary

• An example of the capabilities of a unified RT
  modeling system was given and illustrated the
  complementary nature of information on the
  atmosphere and clouds in the microwave and IR
• Additional near-term work
• Compare various techniques for handling gas
  absorption and scattering across spectral bands
  (e.g., k-distribution method and Optimal Spectral
  Sampling)
• Explore and further develop approximate physical
  methods (e.g., Anomalous Diffraction Theory) for
  fast computation of scattering properties