Methods - cesm-hawc

This page describes how cesm-hawc turns a WACCM column into a simulated ALI observation. Module names refer to the cesm_hawc package.

Column extraction¶

waccm.WACCMAtmosphere selects the model column nearest the requested latitude and longitude. Longitudes are normalized to 0–360° first, matching the CAM grid, so −180–180 inputs select the correct column.

Pressure and altitude¶

Pressure on model levels comes from the hybrid coefficients,

p_k = a_k p_0 + b_k p_s ,

(1)

with $p_0$ from P0 (100 000 Pa if absent). Heights come from hydrostatic integration of the hypsometric equation using virtual temperature $T_v = T\,[1 + (R_v/R_d - 1)\,q]$ :

\Delta z_k = \frac{R_d T_{v,k}}{g_0} \ln\frac{p_{k-1/2}}{p_{k+1/2}} .

(2)

Interface pressures are the geometric means of adjacent levels, with the surface pressure at the bottom and half the top-level pressure at the model top. Each level sits at the midpoint of its layer.

Interpolation to the output grid¶

Profiles are interpolated linearly in altitude onto the configured grid (default 0–65 km every 1 km). Pressure, gas mixing ratios, air density and aerosol number are interpolated in log space. Values beyond the model’s highest or lowest level are held at the end values.

Water vapour¶

WACCM carries tropospheric humidity Q and a stratospheric chemistry tracer H2O. They are blended with a cosine taper in log pressure, centred at 100 hPa and spread over a factor of 10 in pressure (about 316 to 32 hPa): Q (converted to a mixing ratio) below, H2O above. The blended water vapour is saved with each column but is not passed to the simulator.

Sulfate aerosol¶

Two MAM4 modes are used: accumulation (_a1) and coarse (_a3). The Aitken mode (_a2) is not used.

Mode	Constituent	Geometric standard deviation $\sigma_g$
Accumulation, `_a1`	`aerosol_accum`	1.6
Coarse, `_a3`	`aerosol_coarse`	1.2 (Mills et al., 2016)

Number and median radius¶

For each level, the number concentration $N$ is the mode’s number mixing ratio (num_a1, num_a3) times the air density, which is computed from the level’s pressure and temperature. $N$ is floored at 1 m⁻³.

The lognormal median radius $r_m$ is half of CESM’s wet median diameter for the mode (dgnumwet1, dgnumwet3). These include the water the particles take up, so they describe the hydrated sulfuric-acid droplets the Mie calculations below assume, and they are the sizes CESM’s own aerosol optics use. $r_m$ is floored at 1 nm.

If a history file has no wet diameters, $r_m$ is instead derived from the mode’s dry sulfate mass mixing ratio (so4_a1, so4_a3) and number, using the lognormal mass moment

M = N \cdot \tfrac{4}{3}\pi \rho \, r_m^3 \exp\!\left(\tfrac{9}{2}\ln^2\sigma_g\right),

(3)

with $\rho = 1600$ kg m⁻³. This ignores water uptake, so it underestimates the wet radius. cesm-hawc warns when it falls back to this method, and records which method was used in saved files (radius_source_a1, radius_source_a3).

Extinction¶

Each mode gets its own Mie database (sasktran2), built for a lognormal size distribution with that mode’s $\sigma_g$ and the H₂SO₄ refractive index:

Parameter	Value
Wavelengths	470, 525, 745, 869, 1020, 1230, 1450, 1500, 1560, 1750, 2000, 2250, 2500 nm
Median radii	10–590 nm in 10 nm steps
Single-scattering albedo	clamped below 1 (to 0.99999), as in `aliprocessing`

The retrieval uses its own database, with wavelengths from 470 to 1650 nm (aliprocessing 2026.10). When the retrieval runs, cesm-hawc rejects simulated wavelengths outside that range.

Extinction is $N \times \sigma_\text{ext}(r_m, \lambda)$ , where $\sigma_\text{ext}$ is the distribution-weighted total cross-section from that database. Median radii are clipped to 10–590 nm. Wet coarse-mode radii can exceed 590 nm near and below the tropopause, where their extinction is therefore underestimated. Where $r_m$ is below 10 nm the extinction is set to zero. Truth extinction at other wavelengths is taken at the nearest database wavelength, so configured wavelengths should be on the list above.

Each mode enters the simulator as a sasktran2 ExtinctionScatterer defined by its 745 nm extinction and median radius. The radiative transfer model then uses the same mode-matched database for extinction and phase function at every wavelength.

Simulated atmosphere¶

hawcsimulator’s default atmosphere contains Rayleigh scattering, MIPAS standard ozone, solar irradiance and a Lambertian surface with albedo 0.3. cesm-hawc replaces the MIPAS ozone with WACCM ozone and adds WACCM NO₂ and the two sulfate modes. No other WACCM fields enter the radiative transfer.

The L2 retrieval knows nothing about this simulated atmosphere. It retrieves a single lognormal aerosol mode with $\sigma_g = 1.6$ (aliprocessing), and holds ozone close to the MIPAS climatology.

Instrument model¶

cesm-hawc uses the ideal_dolp_imager configuration of hawcsimulator’s IdealALISimulator: a filter-based imaging polarimeter that measures three polarization angles and recovers radiance and degree of linear polarization, with tangent altitudes from −0.5 to 50 km.

Instrument noise¶

Noise comes from hawcsimulator.noise.ALINoiseModel with straylight_fraction = 0 (the upstream default is 0.02). The CLI doesn’t seed it, so repeated runs draw different noise; from Python, pass a seed to cesm_hawc.noise.default_noise_model for reproducible noise.

Orbit sampling¶

In orbit mode, observations come from real orbit files but are placed on the simulated model dates:

The dates of the case’s history files are sorted, and the n-th date uses orbit day n mod D, where D is the number of days the orbit set spans. Gaps in the history files shift this pairing.
From that orbit day’s files, observations are taken every obs_cadence_s seconds at center_pixel. The starting offset is rotated from file to file, so that successive orbits sample different points along the track instead of repeating the same latitudes.
Each observation keeps its real time of day, tangent point and satellite position, but its date is replaced with the simulated date. Solar angles are computed from the resulting time and position, and night-side observations are skipped.

Sulfate burden¶

WACCMAtmosphere.sulfate_column_burden integrates the dry sulfate mass concentration of both modes (the so4 mass mixing ratios times air density) between 15 and 35 km, giving the burden in mg m⁻². fixed mode reports it in each file’s summary.txt.

References¶

Liu, X., et al. (2016). Description and evaluation of a new four-mode version of the Modal Aerosol Module (MAM4) within version 5.3 of the Community Atmosphere Model. Geosci. Model Dev., 9, 505–522.
Mills, M. J., et al. (2016). Global volcanic aerosol properties derived from emissions, 1990–2014, using CESM1(WACCM). J. Geophys. Res. Atmos., 121, 2332–2348.