Optical PNT through the atmosphere

The optical-atmosphere scenario kind answers how often an optical link to the ground works, why it fails, what margin or technique would help, and what a clock does while it is down.

Kshana is the analysis side only. It does not run an optical-ground-station experiment, and it names no measurement partner. The kind is MODELLED: its numerics are checked against an independent program, but no measured availability record has been compared, and every weather, technique and clock input is an assumption. It is an advisory analysis, not a certified availability, integrity or navigation product.

kshana scenarios/optical-atmosphere-network.toml
kshana scenarios/optical-atmosphere-techniques.toml

Both bundled scenarios carry ILLUSTRATIVE cloud climates. They are inputs, not measurements and not any dataset's values.

What is computed#

1. Clear-air loss at the wavelength and elevation: aerosol extinction from the Kim, McArthur and Korevaar visibility law, exponential in height, plus Rayleigh scattering and a stated gas-absorption input. This replaces the single lumped atmospheric_loss_db of optical_linkbudget when an atmosphere is supplied. Without one, the link budget is bit-identical to 0.36 (pre-registered bar A5).
2. Turbulence. A Hufnagel–Valley Cn² profile (HV 5/7 by default) integrated along the slant path gives the plane- or spherical-wave Rytov variance, uplink or downlink. From it come the Andrews–Phillips scintillation index, the gamma-gamma parameters, the weak-turbulence aperture-averaging factor (Churnside) and the gamma-gamma fade probability and fade margin. Aperture averaging is applied by moment matching: the scintillation index is scaled by the averaging factor and both log variances are rescaled to give it exactly.
3. Cloud. A site's cloud climate is the ITU-R P.840 form: a probability of cloud and, given cloud, a log-normal liquid-water column. P.840 supplies the amount of cloud. The attenuation is the geometric-optics optical depth τ = 3L/(2 ρ_w r_e), because P.840's own attenuation law is a microwave law and does not apply at 0.85 to 1.55 µm, where droplets are far larger than the wavelength.
4. Availability of a site: the link is up while the cloud and turbulence fades together stay inside the margin left after the clear-air loss. Cloud and turbulence are taken as independent. The report gives the availability, the clear-sky availability, the ceiling no margin can lift (an optical link cannot pass thick cloud), and the extra margin needed for each target availability, or null if the target is above the ceiling.
5. Outages. Durations are log-normal with a stated mean and spread. The report gives the outages per year, the duration at the median, 90th and 99th percentile, and the probability and share of outage time beyond 1, 6, 24 and 72 hours.
6. Site diversity. The best N sites combined with independent weather (1 − Π(1−a_i)) and with distance-correlated weather: pair correlation exp(−d/d₀), combined in a one-factor Gaussian copula. At d₀ → 0 it is the independent union, and at full correlation the best single site.
7. Techniques. Each [[technique]] is a what-if: a new wavelength (the clear-sky budget, the clear-air loss and the aperture averaging are re-evaluated), a new receive aperture, a stated fractional scintillation reduction (adaptive optics) and extra margin. The result is the availability gain over the baseline at the same site. Kshana does not model an adaptive-optics loop, so the gains are inputs.
8. Timing and ranging through an outage. Each outage duration is mapped through the closed-form coast variance of the clock holdover model, for the clocks listed in [clock]. The report gives the timing and ranging error at stated outage durations, the holdover to the timing threshold, the share of outages inside it, and the timing availability A + U·E[min(D, T_h)]/E[D]. The class power-law floors are the holdover module's synthesised ones, so a tight threshold is governed by that assumed floor, as holdover.rs states.

Scenario fields#

All top-level fields are optional. A bare kind = "optical-atmosphere" runs three illustrative sites. See the two bundled files for every key: wavelength_nm, elevation_deg, availability_targets_pct, [link], [atmosphere], [network], [clock], [[site]], [[technique]].

Verification status#

The oracle checks the implementation of the stated formulas, not the atmosphere, so no row is VALIDATED. ITU-R validation values are not used and are kept out of the repository.

Data#

No weather dataset is bundled. Cloud-climate inputs for a real site would come from a public dataset (Copernicus ERA5, NASA MODIS or ISCCP); each dataset's licence has to be read and recorded before use, and raw data fetched at test time with a pinned SHA-256. The one climatology already in the repository is the Cavazzani et al. clear-night table used by optical_availability.

References#

• Andrews and Phillips, Laser Beam Propagation through Random Media, 2nd ed. (Rytov variance, scintillation index, gamma-gamma parameters).
• Kim, McArthur and Korevaar, Comparison of laser beam propagation at 785 nm and 1550 nm in fog and haze for optical wireless communications, Proc. SPIE 4214 (2001) (visibility law).
• Churnside, Aperture averaging of optical scintillations in the turbulent atmosphere, Appl. Opt. 30 (1991) (aperture-averaging factor).
• ITU-R P.840 (cloud liquid-water statistics), P.1814 and P.1817 (free-space-optics methods).