[Paper Review] Atmospheric dispersion and the implications for phase calibration

Emily I. Curtis, Bojan Nikolic+2 more|ArXiv.org|Dec 15, 2009

Calibration and Measurement Techniques3 citations

TL;DR

This paper investigates the impact of atmospheric dispersion on ALMA's phase calibration, particularly the frequency-dependent wet dispersive path delay. Using the atm transmission code, it finds that dispersive delays become significant (≥5%) at high ALMA bands (≥160 GHz), especially Band 8 (385–500 GHz), where they reach 20–55% of non-dispersive delays. The study shows that ground temperature and lapse rate are the dominant atmospheric factors affecting dispersion, and ancillary calibration instruments can constrain these uncertainties to ~1–2%, enabling accurate phase transfer and WVR-based corrections.

ABSTRACT

The success of any ALMA phase-calibration strategy, which incorporates phase transfer, depends on a good understanding of how the atmospheric path delay changes with frequency (e.g. Holdaway & Pardo 2001). We explore how the wet dispersive path delay varies for realistic atmospheric conditions at the ALMA site using the ATM transmission code. We find the wet dispersive path delay becomes a significant fraction (>5 per cent) of the non-dispersive delay for the high-frequency ALMA bands (>160 GHz, Bands 5 to 10). Additionally, the variation in dispersive path delay across ALMA's 4-GHz contiguous bandwidth is not significant except in Bands 9 and 10. The ratio of dispersive path delay to total column of water vapour does not vary significantly for typical amounts of water vapour, water vapour scale heights and ground pressures above Chajnantor. However, the temperature profile and particularly the ground-level temperature are more important. Given the likely constraints from ALMA's ancillary calibration devices, the uncertainty on the dispersive-path scaling will be around 2 per cent in the worst case and should contribute about 1 per cent overall to the wet path fluctuations at the highest frequencies.

Motivation & Objective

• To quantify the magnitude of wet dispersive path delay in ALMA’s submillimeter bands under realistic atmospheric conditions at the Chajnantor site.
• To assess how variations in atmospheric parameters—water vapor, scale height, ground pressure, and temperature profile—affect dispersive path delay.
• To evaluate the constraints on dispersive path scaling achievable through ancillary calibration instruments, particularly temperature profilers and WVRs.
• To determine the implications of dispersion for ALMA’s phase-transfer and water-vapor radiometry calibration techniques.
• To support the development of accurate frequency-dependent phase correction models for high-frequency ALMA observations.

Proposed method

• Employed the atm transmission code to model atmospheric path delays across ALMA’s full frequency range (84–720 GHz), focusing on wet dispersive components.
• Calculated the dispersive path delay ratio ΔSν/Δc as a function of frequency, using standard atmospheric profiles from the ALMA Observatory Site (AOS) model.
• Varied key atmospheric parameters—water vapor content, scale height, ground pressure, and temperature profile—within realistic ranges to assess sensitivity.
• Used temperature profiler measurements to estimate constraints on the dispersive path scaling, particularly focusing on ground-level temperature and lapse rate (ΓT).
• Evaluated the impact of these variations on phase calibration accuracy, especially for phase-transfer techniques and WVR-based corrections.
• Compared results across ALMA bands, identifying critical frequencies where dispersion dominates (e.g., Bands 5–10, especially Band 8).

Experimental results

Research questions

• RQ1 How significant is the wet dispersive path delay relative to the non-dispersive path delay in ALMA’s high-frequency bands (≥160 GHz)?
• RQ2 How does the dispersive path delay vary across ALMA’s 4-GHz instantaneous bandwidth, and does this variation affect channel-by-channel calibration?
• RQ3 Which atmospheric parameters—water vapor, scale height, ground pressure, or temperature profile—most strongly influence the dispersive path delay?
• RQ4 To what extent can ancillary calibration instruments (e.g., temperature profilers) constrain uncertainties in the dispersive path scaling?
• RQ5 What are the implications of dispersive path delay for ALMA’s phase-transfer calibration strategy and WVR-based phase correction?

Key findings

• In ALMA Bands 5–10 (≥160 GHz), the wet dispersive path delay becomes a significant fraction (≥5%) of the non-dispersive delay, reaching 20–55% in Band 8 (385–500 GHz).
• The variation in dispersive path delay across a 4-GHz bandwidth is typically 2–5%, rising to ~13% in Band 9, indicating the need for channel-by-channel phase correction.
• Water vapor amount and scale height have minimal impact on ΔSν/Δc, with changes <2% across typical ranges.
• Ground-level temperature variations (10–90% of Chajnantor’s distribution) cause 7–9% changes in ΔSν/Δc, and diurnal variations (±5 K) induce ±5% changes at frequencies >345 GHz.
• The lapse rate (ΓT) affects ΔSν/Δc by ~2–3% for typical variations, and temperature profiler measurements with ±1 K accuracy can constrain uncertainty to 0.5–1%.
• With ground temperature and lapse rate measurements, ancillary calibration instruments can constrain dispersive path scaling to 1–2% uncertainty, sufficient for ALMA’s calibration budget.