Rain-Attenuation Peak Frequency in the Terahertz Band

1. Problem Statement

As 6G wireless networks move toward the Terahertz (THz) band (0.1–10 THz), accurate characterization of rain-induced channel degradation is critical. While existing models (e.g., ITU-R P.838-3) describe rain attenuation as a power-law function of rainfall rate ($R$) and frequency, they fail to capture the spectral structure of the loss.

• The Gap: Rain attenuation in the THz band is frequency-selective, exhibiting a "peak" where loss is maximized. Previous studies qualitatively noted a peak near 100 GHz but did not treat the peak frequency ($f_{peak}$) as a distinct, dynamic spectral descriptor.
• The Question: How does the dominant loss region (the peak frequency) evolve as rainfall conditions change? Does it remain fixed, or does it migrate? What physical parameters govern this migration?

2. Methodology

The authors employed a systematic Mie-scattering theory framework to investigate rain attenuation, moving beyond empirical curve-fitting to analyze the underlying microphysics.

• Models Used:
  • Laboratory Case: A separable Gaussian Drop-Size Distribution (DSD) based on controlled nozzle experiments, where the spectral shape is fixed, and only total drop concentration varies with rainfall rate.
  • Outdoor Empirical Cases: Seven representative DSD models covering three families:
    • Exponential: Marshall-Palmer (M-P), Joss-Waldvogel (J-W), Joss-Drizzle (J-D), Joss-Thunderstorm (J-T).
    • Gamma (Atlas-Ulbrich): Hail, Sleet, Snow.
    • Best Model: A complex empirical fit.
• Simulation Parameters:
  • Frequency range: 50–1000 GHz.
  • Rainfall rates ($R$): 1–50 mm/h.
  • Temperatures ($T$): 1°C, 10°C, 20°C, 27°C, 30°C.
  • Calculations included total loss, absorption, and scattering components, excluding atmospheric gas absorption to isolate rain effects.
• Analytical Approach:
  • Derived the relationship between $f_{peak}$ and DSD characteristic scales ($D_c$).
  • Tested three empirical fitting functions for the $f_{peak}$ vs. $R$ relationship: Asymptotic Power-Law (APL), Hill function, and simple Power-Law (PL).
  • Analyzed the impact of the complex refractive index of water (dielectric dispersion) versus DSD structural changes.

3. Key Contributions

1. Identification of Peak Migration: The paper establishes that unlike fixed-shape laboratory conditions, outdoor empirical DSD models exhibit a monotonic migration of the attenuation peak toward lower frequencies as rainfall rate increases.
2. Physical Mechanism Isolation: The study proves that this migration is governed primarily by the rainfall-dependent DSD characteristic scale (specifically the growth of drop sizes), rather than by the total drop concentration or the frequency dispersion of water's dielectric properties (at a fixed temperature).
3. Empirical Parameterization: The authors derived a robust Asymptotic Power-Law (APL) expression to describe the peak frequency migration:
   $$f_{peak}(R) = f_{\infty} + B R^{-p}$$
   Where $f_{\infty}$ is the asymptotic peak frequency at high rainfall, $B$ is an amplitude coefficient, and $p$ is the decay exponent.
4. Cross-Model Descriptor: The number median diameter ($D_{50}$) is identified as a physically transparent, cross-model indicator that linearly correlates with the inverse of the peak frequency ($f_{peak} \propto 1/D_{50}$).

4. Key Results

• Laboratory vs. Outdoor Behavior:
  • Laboratory (Fixed Shape): As $R$ increases, attenuation magnitude increases, but the spectral shape and $f_{peak}$ (fixed at ~56 GHz for the specific Gaussian model) remain unchanged.
  • Outdoor (Variable Shape): As $R$ increases, the DSD shape changes (droplets get larger). Consequently, $f_{peak}$ shifts monotonically downward (e.g., from ~70 GHz to ~30 GHz depending on the model).
• Component Analysis:
  • At low rainfall rates, the scattering component dominates the peak frequency, which shifts rapidly downward.
  • At higher rates, the absorption component becomes dominant, and the total-loss peak settles into an intermediate position.
• Fitting Performance:
  • The Asymptotic Power-Law (APL) model provided the best fit ($R^2 > 0.999$) across all eight DSD models, outperforming the Hill and simple Power-Law models.
  • The simple Power-Law (without an asymptotic offset) failed to capture the flattening of the peak frequency at high rainfall rates.
• Temperature Effects:
  • Temperature acts as a vertical shift to the $f_{peak}$ curve (higher $T$ $\rightarrow$ higher $f_{peak}$) but does not alter the migration trend (the downward slope with increasing $R$).
  • Dielectric dispersion of water was found to have a negligible effect on the location of the peak at a fixed temperature.

5. Significance and Implications

• System Design: For wideband THz links, relying on attenuation at a single carrier frequency is insufficient. The "worst-case" frequency is not static; it shifts as rain intensity changes.
• Adaptive Communications: The findings support the development of weather-aware frequency planning and adaptive subband selection. Systems should dynamically avoid the shifting peak frequency region rather than a fixed band.
• Modeling Accuracy: The proposed APL model and the use of $D_{50}$ as a descriptor offer a more physically grounded approach to predicting THz link availability compared to standard ITU-R models, which assume a static spectral shape.
• Future Research: This work provides a foundation for robust link design in 6G, emphasizing that rain acts not just as an additive loss but as a spectral reshaping mechanism.

In summary, this paper fundamentally shifts the understanding of rain attenuation in the THz band from a static magnitude problem to a dynamic spectral evolution problem, providing the mathematical tools to predict and mitigate frequency-selective fading caused by rain.