Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

An Ultra-Wideband Study of Vegetation Impact on Upper Midband / FR3 Communication

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims vegetation loss in urban 6–18 GHz links grows linearly with foliage depth and with frequency, giving a model Lveg(f,dveg)=alpha(f)dveg, with alpha from 1.26 dB/m at 6–7 GHz to 1.79 dB/m at 17–18 GHz.

desk verdict First 6-18 GHz UWB vegetation-loss dataset with a plausible qualitative trend, but the slope table rests on a single clear-air reference and an origin-constrained fit. read the letter →

arxiv 2412.17864 v1 pith:DT45B4MI submitted 2024-12-20 eess.SP cs.SYeess.SY

Naveed A. Abbasi , Tathagat Pal , Kelvin Arana , Vikram Vasudevan , Jorge Gomez-Ponce , Young-Han Nam , Charlie Zhang , Andreas F. Molisch This is my paper · ORCID
classification eess.SPcs.SYeess.SY
keywords vegetationlossuppermid-bandFR36-18GHzfoliageattenuationurbanpropagationchannelmeasurementlinkbudget
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper reports ultra-wideband measurements of vegetation-induced signal loss in the 6–18 GHz upper mid-band in an urban street-canyon setting. It claims that the excess loss grows linearly with the depth of foliage crossed by the line of sight, with a slope that rises from about 1.26 dB/m at 6–7 GHz to 1.79 dB/m at 17–18 GHz. If correct, this gives network designers a simple frequency-dependent foliage term for FR3 link budgets, replacing extrapolations based on measurements below 3 GHz and above 30 GHz. The paper also contributes a practical ellipse-based method for computing vegetation depth from tree geometry, which makes the model directly usable in planning tools.

What carries the argument

The central machinery is a three-step chain: an ellipse-based geometric model of each tree silhouette gives the vegetation depth dveg along the line of sight; directional power delay profiles are used to extract the best-aligned line-of-sight component PLoS(f;d), cleaned by a 12 dB noise threshold and delay gating; and the excess loss Lveg = PFriis - PLoS is fit by a linear model through the origin, Lveg(f,dveg) = alpha(f)dveg. The ellipse representation is load-bearing because it turns on-site measurements of trunk height, foliage height, and width into a solvable intersection problem for arbitrary path geometry.

What would settle it

Repeat the same six links with a double-directional sounder that resolves azimuth at both ends and with a wider delay gate; if the best-aligned line-of-sight power changes by more than the quoted alpha-min-to-alpha-max spread, the fitted alpha values include unresolvable scattering rather than vegetation attenuation alone. Alternatively, place a foliage-free reference receiver at the same distance as each vegetation point to verify that the Friis and over-the-air calibration account for all non-vegetation urban losses.

Watch

Extended reading notes

Core claim

Using six receiver positions with modeled vegetation depths from 0 to about 28 meters, the paper isolates the best-aligned line-of-sight component in directional power delay profiles, subtracts the Friis free-space power at the same distance and frequency, and treats the remainder as vegetation loss. A linear regression forced through the origin yields an attenuation coefficient alpha(f) for each 1 GHz sub-band; the fitted slopes rise from 1.26 dB/m at 6–7 GHz to 1.79 dB/m at 17–18 GHz, with a min–max spread that widens with frequency. The paper concludes that vegetation loss in FR3 is frequency-dependent and approximately linear in foliage depth, and presents this as the first measured dataset covering the entire 6–18 GHz range in an urban environment.

Load-bearing premise

The model assumes that the strongest directional line-of-sight component picked out of the measured power delay profiles is exactly the signal that passed through the elliptical foliage depth, so that subtracting the Friis free-space power leaves only vegetation loss.

Editorial extensions

If this is right

  • At the upper end of FR3, each meter of tree canopy along the path adds roughly 1.8 dB of loss, which can dominate the link budget at cell-edge distances.
  • The frequency trend means operators can assign foliage-sensitive links to the lower part of FR3, where the per-meter vegetation penalty is smaller.
  • The linear depth model makes vegetation loss a tractable planning input: compute the ellipse-based depth for a path and multiply by alpha(f).
  • The scatter around the fitted lines implies that a single alpha per GHz band is a first-order model, and additional measurements are needed before it generalizes to other vegetation types and densities.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the linear-depth law holds across tree species, foliage-aware FR3 planning could rely on crown silhouettes alone, avoiding detailed leaf-level ray tracing.
  • Measuring the same links with full double-directional resolution would separate true vegetation attenuation from scattering and canopy diffraction, tightening the reported alpha intervals.
  • Comparing these alphas against existing vegetation attenuation standards would identify which frequency bands need revised assumptions, since those standards currently interpolate across the 6–18 GHz gap.
  • Because the reference path has no foliage, any non-vegetation urban loss not captured by Friis would be attributed to the trees; repeating with a foliage-free path at each distance would bound this error.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. This paper reports ultra-wideband (6-18 GHz) directional channel measurements in an urban street canyon, comparing one clear line-of-sight (LoS1) and six vegetated receiver positions. The authors compute vegetation depth by intersecting the Tx-Rx line with ellipse models of measured tree silhouettes, isolate the best-aligned LoS component from directional power delay profiles, and define excess vegetation loss relative to Friis free-space path loss. They fit origin-constrained linear models L_veg(f, d_veg) = α(f) d_veg in 1 GHz sub-bands, reporting α rising from 1.26 dB/m at 6-7 GHz to 1.79 dB/m at 17-18 GHz, and conclude that excess loss increases with vegetation depth and frequency.

Significance. If the quantitative slopes are reliable, this is the first measured 6-18 GHz urban vegetation attenuation dataset and a useful input for FR3 link budgets. The manuscript's strengths are the broad 12 GHz bandwidth, the daily over-the-air calibration, the use of directional PDPs to isolate the direct path, and a clearly described geometric method for vegetation depth. The qualitative claim—excess loss grows with depth and frequency—is directly visible in the measured data and does not depend on the fitted model. However, the fitted α(f) values are in-sample parameters extracted from only six vegetated points and one no-vegetation reference, and the paper provides no confidence intervals or external validation; they should be presented as descriptive estimates rather than validated model coefficients.

major comments (3)
  1. [Section III, Eq. (4); Table I] The excess loss is computed relative to theoretical Friis free-space path loss, and the only obstruction-free reference is LoS1 at 64.5 m. If the true clear-air path loss in this street canyon departs from the free-space exponent of 2 (ground reflection, building/waveguide effects), then P_Friis - P_LoS contains a distance-dependent term that is not vegetation loss. Since d_veg is partly correlated with d in Table I, the origin-constrained fit in Eq. (5) absorbs that baseline error into α(f). The paper provides no confidence intervals, no second clear-air distance, and no test of the free-space baseline; the reported 1.26-1.79 dB/m slopes are thus as sensitive to the baseline assumption as to foliage. Please provide a sensitivity analysis, for example by re-fitting with a free intercept or under alternative path-loss exponents.
  2. [Section III, Eq. (3); Fig. 3] The direct-path power P_LoS is extracted from the best-aligned directional PDP using a 12 dB noise threshold, and the authors explicitly acknowledge residual small-scale fading from unresolvable MPCs and canopy diffraction. This means the single LoS1 reference is not an exactly measured zero point: Fig. 3a shows a 0.57 dB offset from Friis at 6.5 GHz. Any bias of several dB in P_LoS, which the paper's own discussion allows, propagates through the origin-constrained fit into every α(f) in Table II. Please quantify this sensitivity, for example by perturbing P_LoS by a plausible baseline error and re-fitting, or by reporting the fluctuation of P_LoS across repeated measurements.
  3. [Table II; Section IV-B] The columns α_min and α_max are never defined. It is unclear whether they are confidence bounds of the slope estimate, the observed range of α across frequency points within each 1 GHz sub-band, or something else. With six vegetated points per band, the fit has five degrees of freedom, and no R², residual plot, or per-band figure is provided for most bands. Without a definition and fit statistics, the central quantitative result is not reproducible. Please define these columns precisely and add standard errors or bootstrap confidence intervals for α.
minor comments (4)
  1. [Title; Section II-C] The text contains 'V egetation' with an extra space in the title and in the Section II-C heading; please fix this typographical error.
  2. [Section IV-B] The statement that 'at both frequencies, for a vegetation depth of 0, the excess loss is also nearly zero' should be reworded: the zero-depth point is the single LoS1 measurement with a 0.57 dB offset at 6.5 GHz, and the zero intercept is imposed by Eq. (5), not independently measured.
  3. [Table I] Please add estimated uncertainties to the vegetation depth values; the elliptical-silhouette method and the rod/GoPro surveying introduce measurement and modeling error that propagates into the slopes of Table II.
  4. [Fig. 4] Fig. 4 displays only the 6-7 GHz and 17-18 GHz bands; because Table II reports all twelve bands, a compact summary of all fits or a table of residuals would strengthen the presentation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the vegetation-loss model is an explicit empirical fit, not a prediction obtained from its own inputs.

full rationale

The paper's central claim is an empirical characterization: it measures LoS component power at six vegetated points and one clear point, computes excess loss relative to Friis, estimates vegetation depth from on-site tree geometry, and fits Lveg(f,dveg)=alpha(f)*dveg. This is explicit curve fitting, not a derivation that pretends to predict its own data. The slope alpha(f) is admittedly fitted from the same measurements, and the paper does not claim external predictive validation; it repeatedly notes scatter and the need for additional measurements. The origin constraint is an a priori physical assumption (zero vegetation depth implies zero vegetation loss), not a fitted parameter that is later called a prediction. The main measurement-processing choices, such as using the best-aligned directional PDP to isolate the direct path, are documented assumptions and potential sources of bias, but they are not circular. Self-citations to the authors' sounder papers and noise-processing paper describe equipment and established methods, and no load-bearing uniqueness theorem or prior result is imported to make the conclusion forced. The observed qualitative trend, increasing excess loss with vegetation depth and frequency, is directly visible in the measured data and does not depend on the fitted model. No equation is defined in terms of the conclusion it is supposed to support, and no fitted parameter is renamed as a prediction. The quantitative slopes carry uncertainty from the Friis baseline and the single clear-LoS reference, but that is a correctness or bias concern, not circularity.

Assumptions & free parameters 13 free parameters · 5 assumptions · 0 invented entities

Central claim relies on geometric tree modeling, LoS component isolation, a Friis baseline, and an origin-constrained linear fit. No invented entities are introduced. The free parameters are the 12 fitted per-band slopes plus a hand-set noise threshold. The main burden is that the model's quantitative output is fitted to the same measurements used to establish the trend.

free parameters (13)
  • alpha (6-7 GHz) = 1.26 dB/m
    Eq 5 linear fit through origin for the 6-7 GHz band.
  • alpha (7-8 GHz) = 1.50 dB/m
    Eq 5 linear fit through origin for the 7-8 GHz band.
  • alpha (8-9 GHz) = 1.55 dB/m
    Eq 5 linear fit through origin for the 8-9 GHz band.
  • alpha (9-10 GHz) = 1.57 dB/m
    Eq 5 linear fit through origin for the 9-10 GHz band.
  • alpha (10-11 GHz) = 1.68 dB/m
    Eq 5 linear fit through origin for the 10-11 GHz band.
  • alpha (11-12 GHz) = 1.80 dB/m
    Eq 5 linear fit through origin for the 11-12 GHz band.
  • alpha (12-13 GHz) = 1.80 dB/m
    Eq 5 linear fit through origin for the 12-13 GHz band.
  • alpha (13-14 GHz) = 1.69 dB/m
    Eq 5 linear fit through origin for the 13-14 GHz band.
  • alpha (14-15 GHz) = 1.80 dB/m
    Eq 5 linear fit through origin for the 14-15 GHz band.
  • alpha (15-16 GHz) = 1.81 dB/m
    Eq 5 linear fit through origin for the 15-16 GHz band.
  • alpha (16-17 GHz) = 1.76 dB/m
    Eq 5 linear fit through origin for the 16-17 GHz band.
  • alpha (17-18 GHz) = 1.79 dB/m
    Eq 5 linear fit through origin for the 17-18 GHz band.
  • Noise threshold P_lambda = 12 dB above noise floor
    Hand-chosen threshold in Eq 3 that determines which delay bins are kept, and therefore affects the extracted LoS power.
assumptions (5)
  • domain assumption Elliptical tree silhouettes estimated from rod and GoPro field measurements accurately represent the foliage traversed by the LoS ray.
    Section II-C computes dveg by intersecting the LoS line with fitted ellipses; the real canopy has irregular branches and leaves, so this geometric depth is an approximation.
  • domain assumption The best-aligned directional LoS component PLoS(f;d) isolates the through-vegetation direct path, with residual unresolvable MPCs behaving as small-scale fading.
    Section III relies on this to define excess vegetation loss; if unresolvable components contaminate the LoS delay bin, Lveg is biased.
  • domain assumption Friis free-space received power at distance d is the proper lossless baseline for the urban LoS path.
    Eq 4 defines Lveg = P_Friis - P_LoS; any distance or path-loss modeling error is absorbed into vegetation loss.
  • ad hoc to paper Excess vegetation loss is linear in vegetation depth and zero at zero depth.
    Eq 5 and Section IV-B impose an origin-constrained linear fit; this is a modeling convenience, not derived from propagation theory.
  • domain assumption OTA calibration at 44 m, with no vegetation in the first Fresnel zone, removes system and antenna effects for all measurement distances.
    Section II-A uses one calibration point per day; antenna gain pattern and frequency response are assumed stable across the 64.5-126 m links.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An Ultra-Wideband Study of Vegetation Impact on Upper Midband / FR3 Communication." pith.science (2026). https://pith.science/paper/DT45B4MI

@misc{pith2026241217864,
  author       = {Pith},
  title        = {Pith review of: An Ultra-Wideband Study of Vegetation Impact on Upper Midband / FR3 Communication},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DT45B4MI}},
  note         = {Machine review of arXiv:2412.17864}
}
read the original abstract

Growing demand for high data rates is driving interest in the upper mid-band (FR 3) spectrum (6-24 GHz). While some propagation measurements exist in literature, the impact of vegetation on link performance remains under-explored. This study examines vegetation-induced losses in an urban scenario across 6-18 GHz. A simple method for calculating vegetation depth is introduced, along with a model that quantifies additional attenuation based on vegetation depth and frequency, divided into 1 GHz sub-bands. We see that excess vegetation loss increases with vegetation depth and higher frequencies. These findings provide insights for designing reliable, foliage-aware communication networks in FR 3.

Figures

Figures reproduced from arXiv: 2412.17864 by the authors.

Figure 2
Figure 2. Elliptical modeling of tree silhouettes for vegetation depth calculation. (74.7 m) through V eg6 (126 m) introduced varying vegetation depths. It is important to note that longer distances do not al￾ways correspond to greater vegetation depth, as the vegetation a signal traverses depends on the specific scenario geometry. For instance, the LoS to some farther Rx points may go over the canopies of many of the trees a… view at source ↗
Figure 1
Figure 1. Measurement site. The setup aligns the Tx and Rx horn antennas to point to each other along the line-of-sight (LoS), establishing a coordinate system in which the LoS multipath component (MPC) has a co-elevation of 0 ◦ at Tx and Rx. The Rx was rotated in the azimuthal range of −60◦ to 60◦ in 10◦ steps and in the elevation range of −30◦ to 30◦ in 10◦ steps to identify potential errors from misalignment. However, no s… view at source ↗
Figure 3
Figure 3. Sample PDPs between 6-7 GHz. to fluctuations in the observed results. Such MPCs might arise from (unresolvable) scattering, diffraction around the canopy, and ground reflections of MPCs passing through the canopy; they will arise not only in our measurements but in all real￾world urban scenarios with vegetation. The excess vegetation loss is then calculated as the differ￾ence between PLoS(f; d), and the theoretical … view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Vegetation depth vs excess loss for different frequencies. calculated as τ × c, where c = 3 × 108 m/s, providing a clear representation of the propagation paths. In both cases, the strongest component, corresponding to the LoS, appears at a delay matching the physical …

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Ultra-Wideband Double-Directional Channel Measurements and Statistical Modeling in Urban Microcellular Environments for the Upper-Midband/FR3

    eess.SY 2024-12 conditional novelty 6.0 of 10

    An 8 GHz-bandwidth double-directional channel measurement campaign in an urban microcell at 6-14 GHz produces statistical models for path loss, delay spread, and angular spread, and highlights vegetation-obstructed Lo...

Reference graph

Works this paper leans on

12 extracted references · 6 canonical work pages · cited by 1 Pith paper

  1. [1]

    Cellular wireless networks in the upper mid-band,

    S. Kang, M. Mezzavilla, S. Rangan, A. Madanayake, S. B. V e nkatakr- ishnan, G. Hellbourg, M. Ghosh, H. Rahmani, and A. Dhananjay , “Cellular wireless networks in the upper mid-band,” IEEE Open Journal of the Communications Society , 2024

    work page 2024

  2. [2]

    Ultra-wide-band propagation channels,

    A. F. Molisch, “Ultra-wide-band propagation channels, ” Proceedings of the IEEE , vol. 97, no. 2, pp. 353–371, 2009

    2009

  3. [3]

    Rel-19 ran1 agreements - 7-24 ghz channel modelin g,

    RAN 1, “Rel-19 ran1 agreements - 7-24 ghz channel modelin g,” 3GPP , Tech. Rep., 2025

    work page 2025

  4. [4]

    Wideband penetration loss through building materials and partitions at 6.75 ghz in fr1 (c) and 16.95 ghz in the fr3 up per mid-band spectrum,

    D. Shakya, M. Ying, T. S. Rappaport, H. Poddar, P . Ma, Y . Wa ng, and I. Al-Wazani, “Wideband penetration loss through building materials and partitions at 6.75 ghz in fr1 (c) and 16.95 ghz in the fr3 up per mid-band spectrum,” arXiv preprint arXiv:2405.01362 , 2024

    arXiv 2024

  5. [5]

    An empirical propagation pre diction model for urban street canyon environments at 6, 10, and 18 gh z,

    S.-S. Oh, J.-W. Choi, H.-C. Lee, Y .-C. Lee, B.-L. Cho, I.- Y . Lee, J.-H. Lim, J.-I. Lee, and S. W. Park, “An empirical propagation pre diction model for urban street canyon environments at 6, 10, and 18 gh z,” Microwave and Optical Technology Letters , vol. 61, no. 6, pp. 1574– 1578, 2019

    work page 2019

  6. [6]

    Sub- 6 ghz to mmwave for 5g-advanced and beyond: Channel measurem ents, characteristics and impact on system performance,

    H. Miao, J. Zhang, P . Tang, L. Tian, X. Zhao, B. Guo, and G. L iu, “Sub- 6 ghz to mmwave for 5g-advanced and beyond: Channel measurem ents, characteristics and impact on system performance,” IEEE Journal on Selected Areas in Communications, vol. 41, no. 6, pp. 1945–1960, 2023

    work page 1945

  7. [7]

    Urban outdoor propagation measurements and channel models at 6.75 ghz fr1 (c) and 16.95 ghz fr3 upper mid-band spectrum for 5g and 6g,

    D. Shakya, M. Ying, T. S. Rappaport, P . Ma, I. Al-Wazani, Y . Wu, Y . Wang, D. Calin, H. Poddar, A. Bazzi et al. , “Urban outdoor propagation measurements and channel models at 6.75 ghz fr1 (c) and 16.95 ghz fr3 upper mid-band spectrum for 5g and 6g,” arXiv preprint arXiv:2410.17539, 2024

    arXiv

  8. [8]

    Ultra-wideband double-directionally resolved channel m easurements of line-of-sight microcellular scenarios in the upper mid- band,

    N. A. Abbasi, K. Arana, J. Gomez-Ponce, T. Pal, V . V asudev an, A. Bist, O. G. Serbetci, Y . H. Nam, C. Zhang, and A. F. Molisch, “Ultra-wideband double-directionally resolved channel m easurements of line-of-sight microcellular scenarios in the upper mid- band,” 2024. [Online]. Available: https://arxiv.org/abs/2412.12306

    arXiv 2024

Show all 12 references
  1. [9]

    Mill imeter-wave propagation in vegetation: Experiments and theory,

    F. K. Schwering, E. J. Violette, and R. H. Espeland, “Mill imeter-wave propagation in vegetation: Experiments and theory,” IEEE Transactions on Geoscience and Remote Sensing , vol. 26, no. 3, pp. 355–367, 1988

    1988

  2. [10]

    Recommendation ITU-R P .833-10 attenuati on in vegeta- tion,

    I. T. Union, “Recommendation ITU-R P .833-10 attenuati on in vegeta- tion,” ITU, Tech. Rep., 2021

    2021

  3. [11]

    Impact of noisy measurements wi th fourier-based evaluation on condensed channel parameters ,

    J. Gomez-Ponce, N. A. Abbasi, Z. Cheng, S. Abu-Surra, G. Xu, J. Zhang, and A. F. Molisch, “Impact of noisy measurements wi th fourier-based evaluation on condensed channel parameters ,” IEEE Transactions on Wireless Communications , 2023

    2023

  4. [12]

    A. F. Molisch, Wireless communications, 3rd ed. IEEE Press - John Wiley & Sons, 2023

    2023

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.