Improving (Microwave) Full Polarization (4-Stokes) Simulation Capability in CRTMv3

[BenjaminTJohnson]

Overview

This master issue tracks the implementation of robust full polarization (4-Stokes: I, Q, U, V) simulation capability for the microwave (MW) spectral region in CRTM v3.

Currently, CRTM v3 has infrastructure for 4-Stokes simulation (MAX_N_STOKES=4, Options%n_Stokes), but MW is fundamentally implemented as a 2-component model (V/H polarizations only). This issue coordinates the work needed to enable operational full-polarization MW simulation.

Current State

• UV/VIS: Full Stokes capable
• IR: Full Stokes capable
• MW: Limited to 2-component (V/H) only

What Works

• n_Stokes parameter exists and can be set to 4
• RT solvers (ADA, AMOM) have some multi-Stokes infrastructure
• Zeeman LBL code exists (but disconnected)

What's Missing

• Surface emissivity models output only 2 components (V/H)
• No oriented particle scattering (6 elements sufficient for random orientation)
• No Faraday rotation
• Zeeman not integrated with operational RT
• No cross-polarization coupling at surface

---

Priority Items

High Priority

1. Extend MW Surface Emissivity/Reflection Models to Full 4x4 Matrix

Files: src/SfcOptics/CRTM_MW_Water_SfcOptics.f90, CRTM_FastemX.f90, MW land/ice/snow modules

Current: Only diagonal elements populated

SfcOptics%Emissivity(i,1)  ! V only
SfcOptics%Emissivity(i,2)  ! H only
SfcOptics%Reflectivity(i,1,i,1)  ! V-V only
SfcOptics%Reflectivity(i,2,i,2)  ! H-H only
! All off-diagonal = 0

Required:

• Compute Q, U, V Stokes emissivity components
• Populate full 4x4 reflection matrix including off-diagonal terms
• Add cross-polarization coupling from surface roughness/foam
• Update FASTEM and all MW surface models

---

2. Add Oriented Particle Scattering Capability

Files: src/AtmScatter/CRTM_CloudScatter.f90, src/RTSolution/Common_RTSolution.f90, CloudCoeff

Current:

• 6 phase elements sufficient for randomly oriented symmetric particles (spheres, spheroids)
• RT solver assumes symmetry-based parameterization (α₁, α₂, α₃, α₄, β₁, β₂)
• No support for preferentially oriented particles

Note: The existing 6-element structure correctly handles the full 4×4 Mueller matrix for randomly oriented particles due to symmetry constraints. Expansion to 16 elements is NOT needed for this case.

Required for oriented particles (ice crystals, canted raindrops):

• Add orientation distribution parameters to cloud/aerosol definitions
• Implement orientation-dependent scattering calculations
• Add canting angle distribution support
• Create new coefficient data for oriented particle optical properties
• Modify RT solver to handle asymmetric scattering matrices when needed

---

3. Integrate Zeeman LBL Code into Operational RT Solver (long term goal is to create generic Zeeman capability that applies to all relevant channels and sensors, rather than the current specific-sensor based approach.)

Files: src/Zeeman/src_lbl/O2_zeeman_lbl.f90, src/AtmAbsorption/ODZeeman/, src/RTSolution/

Current: Zeeman LBL computes off-diagonal transmittance (tau12_Re, tau12_Im) but is disconnected from main RT

Required:

• Create interface between Zeeman module and CRTM forward model
• Propagate off-diagonal transmission matrix through RT equations
• Add magnetic field vector as input to operational CRTM calls
• Enable Zeeman-induced cross-polarization in RTSolution

---

4. Implement Ionospheric Faraday Rotation Module

Files: New module needed in src/

Current: Not implemented - no ionosphere module exists

Required:

• Create CRTM_Faraday_Module.f90
• Add TEC (Total Electron Content) as input parameter
• Compute rotation angle from magnetic field and electron density
• Apply polarization rotation matrix in RT solution
• Add gyrofrequency calculations

---

Medium Priority

5. Implement Rain/Ice Depolarization Physics

Files: src/AtmScatter/, RT solvers

Current: No differential extinction or depolarization from precipitation

Required:

• Add rain-induced depolarization model
• Implement cross-polarization extinction
• Add raindrop shape (oblate spheroid) effects
• Model ice particle depolarization

---

6. Add Circular Polarization Generation Mechanisms

Files: Surface modules, RT solvers

Current: No mechanism to generate Stokes V

Required:

• Identify physical sources of circular polarization in MW
• Implement V-Stokes generation from relevant processes
• Add circular polarization coupling in reflection matrix

---

Architecture Changes

7. RT Solver Updates

Files: src/RTSolution/UWisc_SOI/SOI_CRTM_Forward_Module.f90, src/RTSolution/CRTM_RTSolution.f90

Current: SOI hardcoded to SFCOPTICS_N_STOKES = 2

Required:

• Remove hardcoded 2-Stokes limitation in SOI
• Ensure all RT solvers support n_Stokes=4 for MW
• Update surface boundary conditions for full 4x4 matrix
• Validate ADA/AMOM/SOI with 4-Stokes MW

---

8. Testing and Validation

Files: test/mains/

Current: No regression tests for MW full-polarization

Required:

• Add unit tests for 4-Stokes MW surface emissivity
• Add regression tests for Zeeman polarization
• Add validation against WindSat/SSMIS polarimetric observations
• Create benchmark cases for Faraday rotation

---

Sub-Issues

This master issue will be broken into the following sub-issues:

Priority	Sub-Issue	Components
High	#xxx	MW Surface 4x4 Reflection Matrix
High	#xxx	Oriented Particle Scattering
High	#xxx	Zeeman Integration with RT Solver
High	#xxx	Faraday Rotation Module
Medium	#xxx	Rain/Ice Depolarization
Medium	#xxx	Circular Polarization Mechanisms
Arch	#xxx	RT Solver 4-Stokes Support
Test	#xxx	MW Full-Pol Validation Suite

---

References

• WindSat polarimetric radiometer documentation
• SSMIS polarimetric channels
• Meissner & Wentz ocean emissivity model
• Rosenkranz O2 Zeeman absorption model

Related Sensors

Full-pol MW capability would benefit simulation of:

• WindSat (all 4 Stokes)
• SSMIS (V, H, plus some polarimetric)
• Future polarimetric MW radiometers

---

Labels

enhancement, MW, polarization, multi-issue

[BenjaminTJohnson]

Partially addressed in the v3.2.0 updates -- will continue to iterate on this.