Pith. sign in

REVIEW 4 major objections 4 minor 29 references

A 4.3-million-parameter neural network can forecast Mars weather for a full Martian sol, reaching a one-hour validation RMSE of about 0.004 in normalized units and a 25-hour RMSE of about 0.031 without divergence.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-02 11:12 UTC pith:AURURXZQ

load-bearing objection A promising but under-validated first Mars weather emulator using HEALPix-aware convolutions; the core skill claim is plausible but rests on missing baselines and an uncontrolled comparison. the 4 major comments →

arxiv 2607.19370 v1 pith:AURURXZQ submitted 2026-06-14 physics.ao-ph

Aircast-Mars: A Mars Foundation Model for Global Weather Forecasting with HEALPix-Aware Convolutions

Manmeet Singh , Saptarishi Dhanuka , Naveen Sudharsan , Houman Owhadi , Krista M. Soderlund , Alphan Altinok This is my paper
classification physics.ao-ph
keywords Mars weather forecastingdeep learning emulatorHEALPix convolutionplanetary atmosphereautoregressive rolloutEMARS reanalysisU-Netfoundation model
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Deep learning has produced fast weather emulators for Earth, but Mars is a different regime: a thin CO2 atmosphere, no oceans, and rapid thermal tides. Aircast-Mars adapts an equal-area spherical-convolution architecture to the Martian atmosphere, training on the EMARS reanalysis to predict hourly temperature and winds. The paper's central claim is that a compact, geometry-respecting network with only 4.3 million parameters captures synoptic-scale Martian dynamics: it reaches a validation MSE of 1.58e-5, and when unrolled autoregressively it stays stable for 25 hours with RMSE growing from about 0.004 to 0.031. This matters because a 0.5-second single-GPU forecast turns thousand-member ensembles and onboard forecasting into practical options for mission support. The authors are careful to note that skill is measured against reanalysis, not independent observations, and that the diurnal cycle may be under-represented.

Core claim

The central discovery is that a 2D U-Net operating on the 12-face HEALPix sphere, with inter-face padding that carries neighboring face values across convolution boundaries, learns a skillful one-hour propagation operator for Mars temperature and winds. On the EMARS v1.0 reanalysis (28 levels, three variables), it achieves a best validation MSE of 1.58e-5 in normalized units, corresponding to roughly 0.59 K at t+1h and 4.59 K at t+25h for temperature. Recursive rollout for 25 hours produces monotonically growing but non-divergent RMSE that saturates toward the model's learned climatology after about 12 hours. Against a 3D U-Net baseline that treats the sphere's faces as an unconnected third

What carries the argument

The load-bearing mechanism is the HEALPix inter-face padding scheme. HEALPix is a spherical pixelization with 12 equal-area curvilinear faces; before each 3x3 convolution, the model folds the faces into a batch dimension and pads each face by borrowing strips from its neighbors according to the HEALPix adjacency graph, applying the correct rotations for polar faces and a 50/50 blend at tri-face corners. This removes the artificial discontinuities that zero- or reflect-padding would create at face boundaries. The rest of the network uses residual convolutional blocks with capped GELU activations, the cap preventing unbounded intermediate values that would destabilize long autoregressive rollo

Load-bearing premise

All skill numbers are measured against the EMARS v1.0 reanalysis, which assimilates sparse spacecraft temperature retrievals and may under-represent the true diurnal cycle; if EMARS carries model or assimilation artifacts, the reported MSE/RMSE values overstate real-world forecast skill.

What would settle it

A direct comparison of Aircast-Mars one-hour and 25-hour forecasts against a persistence baseline on the same hold-out set, or against independent surface meteorological measurements over the same dates, would settle the claim: the model must beat persistence at short lead times and stay within reanalysis-level error on independent observations.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • A single GPU can produce a 25-hour Mars weather rollout in about 12 seconds, enabling thousands of perturbed initial conditions for landing-risk assessment in minutes rather than weeks.
  • The order-of-magnitude gap over the 3D U-Net suggests that other geometry-agnostic deep emulators for planetary atmospheres could be improved by respecting the target topology.
  • The compact 17MB checkpoint opens the door to near-real-time forecasting on mission hardware.
  • The stable saturation of RMSE after ~12 hours implies the model is a reliable short-range propagator but not a mid-range climate simulator, setting an expectation for future recurrent or multi-step architectures.
  • The same three-variable, single-step training recipe can be extended with dust and surface-pressure channels, which the paper identifies as the critical next step.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The claimed superiority over the 3D U-Net is not from a controlled ablation: the baseline trained at lower resolution, for fewer epochs, and with a different learning rate. A fair comparison holding resolution, epochs, and optimizer constant is needed to isolate the contribution of inter-face padding; until then, the order-of-magnitude claim should be read as an indication, not a measurement.
  • Because EMARS assimilates temperature retrievals from a limited set of spacecraft overpass local times, the reanalysis's diurnal cycle may be smoother than Mars's actual 60-80 K surface swings. If so, the model's stable 25-hour rollout could be partly a consequence of that smoothing, and validation against in-situ rover measurements at fixed local times would reveal it.
  • The t+1h error of ~0.004 is not yet benchmarked against a persistence forecast. The paper states that comparison is deferred; if persistence matches or beats the model at one hour, the learned operator may be closer to an interpolation than to dynamics.
  • A strong stress test for the same architecture would be a global dust storm, where radiative heating changes the thermal profile within hours; the current model has no dust channel, so its next natural testbed is whether adding dust optical depth as a channel preserves rollout stability or requires an explicit regime-switching mechanism.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper presents Aircast-Mars, a 4.3M-parameter HEALPix-aware 2D U-Net trained on EMARS v1.0 hourly temperature, zonal wind, and meridional wind fields regridded to Nside=64 (~110 km). The model is trained to predict the next hourly state. The manuscript reports a best validation MSE of 1.58e-5, stable autoregressive rollouts for one Martian sol (RMSE growing from ~0.004 at t+1h to ~0.031 at t+25h), an order-of-magnitude improvement over a 30M-parameter 3D U-Net baseline, and a 0.5-second per-step GPU inference time. The central claim is that geometry-respecting HEALPix-aware convolutions capture synoptic-scale Martian atmospheric dynamics with high efficiency.

Significance. If the claims are substantiated, Aircast-Mars would be a useful fast surrogate for Mars weather emulation, with plausible applications to mission planning, ensemble generation, and onboard forecasting. The paper is honest about several limitations, including normalized units, reanalysis bias, and the absence of aerosol channels. However, the headline skill claims are currently not supported by controlled comparisons or standard baselines. The 3D U-Net comparison is explicitly uncontrolled, no persistence or climatology baseline is reported, and the data-split protocol is not specified. The code and data are not yet available, so the numerical results cannot be independently verified from the manuscript alone.

major comments (4)
  1. [Section 3.1, Table 1] The central 'geometry matters' claim rests on an uncontrolled comparison. The 3D U-Net was trained at Nside=8 with 20 epochs and lr=1e-3, while the HEALPix U-Net ran at Nside=64 with 50 epochs and lr=1e-4. The HEALPix U-Net at Nside=16 achieves only 1.4e-4, identical to the 3D U-Net, so the reported order-of-magnitude gain could be due to resolution, training length, or optimizer settings rather than HEALPix-aware padding. The manuscript acknowledges this in Section 3.1, but Section 5 still lists 'Geometry matters' as a key finding. This overstates the evidence. A controlled comparison (matching resolution, epochs, and learning rate, or ablating inter-face padding) is needed before the architecture claim can be accepted.
  2. [Section 2.3] The data-split protocol is underspecified. The text says only '80/20 train/validation split' without stating whether it is temporal or random. If the split is random, then neighboring hourly states from the same sols are likely in both training and validation sets, allowing the network to memorize rather than learn dynamics. This would inflate the reported validation MSE and undermine both the absolute accuracy and the rollout results. The authors must specify the split criterion (e.g., contiguous blocks, by Mars year, or by sol) and justify that no temporal leakage occurs.
  3. [Section 3.2] The claim that the network has 'learned a meaningful dynamical propagation operator rather than a trivial identity mapping' is not supported because no persistence or climatology baseline is provided. In fact, the text states that such comparisons are 'deferred to future work.' For an atmosphere with high hourly temporal correlation, a persistence forecast could plausibly achieve RMSE near 0.004 in normalized units. Without that baseline, the absolute RMSE cannot distinguish learned dynamics from self-similarity. This is a load-bearing omission for the central claim and should be addressed before the paper can be considered complete.
  4. [Section 3.3] The rollout stability result is based on 'a single initial condition.' The RMSE growth curve in Figure 2 is therefore anecdotal; a different start date, season, or dust condition could yield substantially different behavior. The paper should present statistics over multiple initial conditions (e.g., 10–20 cases across Mars seasons and years), with mean and spread of RMSE as a function of lead time. Without this, the monotonic growth and no-divergence conclusion cannot be generalized.
minor comments (4)
  1. [References] The abstract and Section 2.2.1 attribute the 'DLESyM framework' to reference [1], but reference [1] is Watt-Meyer et al., which describes the ACE model, not DLESyM. Please clarify the correct citation or define DLESyM explicitly.
  2. [Figure 2 caption] The figure caption says '24 forecast hours' while the text (Sections 3.3 and 5) says 25 hours (one Martian sol). Please align these numbers.
  3. [Section 4.2] The physical-unit translation states that wind components have a 'typical range of ~350 m/s.' Given that Mars wind speeds are generally reported up to ~30 m/s, this range seems implausibly large and needs justification (e.g., if it is the full reanalysis range including extreme events, that should be stated).
  4. [General] There are several typographical artifacts (e.g., 'T able 1', 'F old/Unfold', 'C ¸ic ¸ek') that should be corrected in a final version.

Circularity Check

0 steps flagged

No circularity: the model is trained on EMARS and evaluated on held-out EMARS states; no fitted parameter is renamed as a prediction and no load-bearing self-citation chain exists.

full rationale

The derivation chain is a standard supervised-learning setup: EMARS reanalysis fields are preprocessed (staggered-grid interpolation, normalization, HEALPix regridding), a HEALPix-aware 2D U-Net is trained to minimize MSE against the next-hour EMARS state, and skill is then reported on a held-out portion of the same reanalysis using both one-step and autoregressive rollouts. The target quantity (the t+1h state) is not baked into the model definition or into any fitted parameter; it is the supervised label, and the reported MSE/RMSE are measured on validation data that were not used for training. No equation in the paper reduces to an identity or to a fit renamed as a prediction. The DLESyM/ACE-inspired architecture is attributed to an external reference [1], and the authors of the present paper do not overlap with that reference's authors; there is no uniqueness theorem imported from prior work by the same authors. The comparison to the 3D U-Net is explicitly acknowledged as not strictly controlled (Section 3.1), which weakens the architecture claim but is a statement about experimental design, not circularity. The absence of a persistence or climatology baseline is also explicitly acknowledged and deferred to future work (Section 3.2); this is a missing baseline, not a circular step. The reanalysis-bias caveat (Section 4.2) and the aerosol limitation similarly limit external validity without making the derivation self-referential. The central scientific content — that a 4.3M-parameter geometry-aware network achieves the reported validation MSE and stable 25-hour rollout on held-out EMARS data — stands independently of its inputs; it could in principle be falsified by a different training/evaluation protocol. Therefore, no significant circularity is present and the score is 0.

Axiom & Free-Parameter Ledger

5 free parameters · 6 axioms · 0 invented entities

The central claim rests on EMARS fidelity as ground truth, on the sufficiency of T/U/V channels, on HEALPix regridding preserving dynamics, and on the usual ML assumption that training distribution covers evaluation distribution. Free parameters are the network weights and normalization statistics fitted to EMARS, plus hand-chosen architecture/training hyperparameters. No new physical entities are introduced, but the model itself is a fitted surrogate rather than a first-principles derivation.

free parameters (5)
  • Neural network weights (4.3M) = Trained to minimize MSE on EMARS
    All predictive skill of the model comes from these fitted weights; they are not derived from physics.
  • Min-max normalization statistics = Global per-variable min/max from EMARS
    Normalization changes the effective weighting of variables and defines the normalized RMSE units in which all metrics are reported.
  • Architecture hyperparameters = Widths 64/128/256, 3 encoder levels, 2x2 pooling, 3x3 kernels, Capped GELU cap (value not stated)
    Chosen by hand; the cap value directly affects the rollout stability claim.
  • Training hyperparameters = lr 1e-4, batch size 1, 50 epochs, 80/20 split (random vs chronological not stated)
    The comparison to the 3D U-Net uses different lr/epochs/resolution, so the performance difference is confounded.
  • HEALPix corner blending ratio = 50/50
    Hand-chosen edge blending at tri-face vertices; no ablation is performed.
axioms (6)
  • domain assumption EMARS v1.0 reanalysis is an accurate representation of Martian atmospheric state at hourly resolution
    Sections 2.1 and 4.2; all training and evaluation uses EMARS, and the paper itself notes diurnal-cycle biases from sparse overpass local times.
  • domain assumption Temperature, zonal wind, and meridional wind are sufficient prognostic variables for short-term Mars forecasting
    Section 2.1; dust and surface pressure are omitted, so sol-scale regime transitions cannot be represented.
  • domain assumption Bilinear regridding from lat-lon to HEALPix preserves the dynamics needed for forecasting
    Section 2.1; no sensitivity to the interpolation scheme is reported.
  • domain assumption Min-max normalization to [0,1] and MSE loss produce a meaningful learning target
    Section 2.1; normalized metrics are hard to interpret without physical-unit or baseline comparisons.
  • domain assumption Standard supervised training on hourly samples learns a useful dynamical operator rather than interpolation or autocorrelation structure
    Section 2.3; no persistence baseline is provided, so one-hour skill could partly reflect temporal autocorrelation.
  • standard math HEALPix inter-face padding as implemented correctly realizes spherical adjacency
    Section 2.2.1; core to the architecture claim, but no verification or tests are provided.

pith-pipeline@v1.3.0-alltime-deepseek · 9769 in / 10780 out tokens · 110907 ms · 2026-08-02T11:12:30.818322+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of Aircast-Mars: A Mars Foundation Model for Global Weather Forecasting with HEALPix-Aware Convolutions." pith.science (2026). https://pith.science/paper/AURURXZQ

@misc{pith2026260719370,
  author       = {Pith},
  title        = {Pith review of: Aircast-Mars: A Mars Foundation Model for Global Weather Forecasting with HEALPix-Aware Convolutions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AURURXZQ}},
  note         = {Machine review of arXiv:2607.19370}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Foundation models for planetary atmospheres promise fast, lightweight surrogates of expensive general circulation models (GCMs) for mission planning and scientific inquiry. Here we present Aircast-Mars, a deep-learning weather prediction system for Mars trained on the Ensemble Mars Atmosphere Reanalysis System (EMARS) v1.0. We regrid temperature, zonal wind, and meridional wind fields across 28 vertical levels onto a hierarchical equal-area isolatitude pixelization (HEALPix) mesh at Nside = 64 (~110 km resolution) and train a HEALPix-aware 2D U-Net inspired by the DLESyM architecture to predict the next hourly atmospheric state. The model employs custom inter-face padding that respects the topology of the 12-face HEALPix sphere and modern ConvNeXt residual blocks with capped Gaussian Error Linear Unit (GELU) activations. While containing 4.3 million trainable parameters, a compact size compared to terrestrial weather foundation models, the network achieves a best validation Mean Squared Error (MSE) of 1.58e-5 in normalized units. Recursive autoregressive rollouts remain stable and physically coherent for 25 hours (one Martian sol), with Root Mean Square Error (RMSE) growing monotonically from ~0.004 at t + 1 h to ~0.031 at t + 25 h without divergence. Compared to a baseline 3D U-Net, the HEALPix-aware architecture reduces validation loss by more than an order of magnitude while using fewer parameters. The model generates a one-hour forecast in approximately 0.5 seconds on a single GPU, offering several orders-of-magnitude speedup over traditional numerical GCMs. These results demonstrate that parsimonious, geometry-respecting neural architectures can capture synoptic-scale Martian atmospheric dynamics and provide a foundation for planetary-scale weather forecasting.

Figures

Figures reproduced from arXiv: 2607.19370 by Alphan Altinok, Houman Owhadi, Krista M. Soderlund, Manmeet Singh, Naveen Sudharsan, Saptarishi Dhanuka.

Figure 1
Figure 1. Figure 1: Single-step prediction at HEALPix Level 6 for a single equatorial face (Face 4). Left: EMARS ground truth surface temperature. Right: DLESyM HEALPix U-Net prediction. Showing the local face patch highlights the model’s ability to capture fine￾scale structures. probabilistic scoring rules such as CRPS [23]. 3.3 Recursive Rollout Stability A critical test for any weather emulator is whether it remains stable… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

29 extracted references · 4 linked inside Pith

  1. [1]

    Watt-Meyer, O., Dresdner, G., McGibbon, J., et al. (2024). ACE: A fast, skillful learned global atmospheric model for climate predic- tion.arXiv preprint arXiv:2310.02074

    Pith/arXiv arXiv 2024

  2. [2]

    and Gimpel, K

    Hendrycks, D. and Gimpel, K. (2016). Gaus- sian error linear units (GELUs).arXiv preprint arXiv:1606.08415

    Pith/arXiv arXiv 2016

  3. [3]

    Liu, Z., Mao, H., Wu, C.-Y., Feichtenhofer, C., Darrell, T., and Xie, S. (2022). A Con- vNet for the 2020s. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 11976–11986

    2022

  4. [4]

    Ronneberger, O., Fischer, P., and Brox, T. (2015). U-Net: Convolutional networks for biomedical image segmentation. InInterna- tional Conference on Medical Image Comput- ing and Computer-Assisted Intervention, pp. 234–241. Springer

    2015

  5. [5]

    S., Brox, T., and Ronneberger, O

    C ¸i¸ cek,¨O., Abdulkadir, A., Lienkamp, S. S., Brox, T., and Ronneberger, O. (2016). 3D U-Net: Learning dense volumetric segmenta- tion from sparse annotation. InInternational Conference on Medical Image Computing and Computer-Assisted Intervention, pp. 424–432. Springer

    2016

  6. [6]

    L., Lewis, S

    Read, P. L., Lewis, S. R., and Mulholland, D. P. (2015). The physics of Martian weather and climate: A review.Reports on Progress in Physics, 78(12):125901

    2015

  7. [7]

    M., Clancy, R

    Haberle, R. M., Clancy, R. T., Forget, F., Smith, M. D., and Zurek, R. W., eds. (2017). The Atmosphere and Climate of Mars. Cam- bridge University Press

    2017

  8. [8]

    J., Gillespie, H

    Greybush, S. J., Gillespie, H. E., and Wil- son, R. J. (2019). Transient eddies in the TES/MCS Ensemble Mars Atmosphere Re- analysis System (EMARS).Icarus, 317:158– 181

    2019

  9. [9]

    J., Wilson, R

    Greybush, S. J., Wilson, R. J., Hoffman, R. N., Hoffman, M. J., Miyoshi, T., Ide, K., McConnochie, T., and Kalnay, E. (2012). Ensemble Kalman filter data assimilation of Thermal Emission Spectrometer temperature retrievals into a Mars GCM.Journal of Geo- physical Research: Planets, 117(E11)

    2012

  10. [10]

    R., Moreno, R

    Karlbauer, M., Cresswell-Clay, N., Durran, D. R., Moreno, R. A., Kurth, T., Bonev, B., Brenowitz, N., and Butz, M. V. (2024). Ad- vancing parsimonious deep learning weather prediction using the HEALPix mesh.Jour- nal of Advances in Modeling Earth Systems, 16(8):e2023MS004021. Preprint 9

    2024

  11. [11]

    M., Hivon, E., Banday, A

    G´ orski, K. M., Hivon, E., Banday, A. J., Wan- delt, B. D., Hansen, F. K., Reinecke, M., and Bartelmann, M. (2005). HEALPix: A frame- work for high-resolution discretization and fast analysis of data distributed on the sphere. The Astrophysical Journal, 622(2):759

    2005

  12. [12]

    Bi, K., Xie, L., Zhang, H., Chen, X., Gu, X., and Tian, Q. (2023). Pangu-Weather: A 3D high-resolution model for fast and accurate global weather forecast.Nature, 619(7970):533–538

    2023

  13. [13]

    Lam, R., Sanchez-Gonzalez, A., Willson, M., et al. (2023). GraphCast: Learning skill- ful medium-range global weather forecasting. Science, 382(6677):1416–1421

    2023

  14. [14]

    Pathak, J., Subramanian, S., Harrington, P., et al. (2022). FourCastNet: A global data- driven high-resolution weather forecasting model.arXiv preprint arXiv:2202.11214

    Pith/arXiv arXiv 2022

  15. [15]

    Montabone, L., Forget, F., Millour, E., et al. (2015). Eight-year climatology of dust optical depth on Mars.Icarus, 251:65–95

    2015

  16. [16]

    A., Murphy, J

    Kahre, M. A., Murphy, J. R., Newman, C. E., Wilson, R. J., Cantor, B. A., Lemmon, M. T., and Wolff, M. J. (2017). The Mars dust cycle. In Haberle, R. M. et al., eds.,The Atmosphere and Climate of Mars, pp. 295–337. Cambridge University Press

    2017

  17. [17]

    Bonev, B., Kurth, T., Grossman, C., et al. (2023). Spherical Fourier neural operators: Learning stable dynamics on the sphere. In International Conference on Machine Learn- ing, pp. 2806–2823. PMLR

    2023

  18. [18]

    Paszke, A., Gross, S., Massa, F., et al. (2019). PyTorch: An imperative style, high- performance deep learning library. InAd- vances in Neural Information Processing Sys- tems, vol. 32

    2019

  19. [19]

    Kingma, D. P. and Ba, J. (2015). Adam: A method for stochastic optimization. InPro- ceedings of the 3rd International Conference on Learning Representations (ICLR)

    2015

  20. [20]

    Battalio, J. M. and Wang, H. (2021). The Mars dust activity database (MDAD): A com- prehensive statistical study of dust storm se- quences.Icarus, 354:114059

    2021

  21. [21]

    E., Bertrand, T., Battalio, J

    Newman, C. E., Bertrand, T., Battalio, J. M., et al. (2021). Multi-model meteorological and aeolian predictions for Mars 2020 and the Jezero crater region.Space Science Reviews, 217(1):20

    2021

  22. [22]

    Kochkov, D., Yuval, J., Langmore, I., et al. (2024). Neural general circulation models for weather and climate.Nature, 632(8027):1060– 1066

    2024

  23. [23]

    and Raftery, A

    Gneiting, T. and Raftery, A. E. (2007). Strictly proper scoring rules, prediction, and estimation.Journal of the American Statisti- cal Association, 102(477):359–378

    2007

  24. [24]

    Keisler, R. (2022). Forecasting global weather with graph neural networks.arXiv preprint arXiv:2202.07575

    Pith/arXiv arXiv 2022

  25. [25]

    Leovy, C. B. and Zurek, R. W. (1979). Ther- mal tides and Martian dust storms: Direct evidence for coupling.Journal of Geophysical Research: Solid Earth, 84(B6):2956–2968

    1979

  26. [26]

    L., Rosenbaum, J

    Wang, J. L., Rosenbaum, J. J., Prasad, A. N., Raad, R. R., Putman, E. J., Harrington, A. D., et al. (2025). Potential health im- pacts, treatments, and countermeasures of Martian dust on future human space explo- ration.GeoHealth, 9:e2024GH001213

    2025

  27. [27]

    S., Kerschmann, R

    Levine, J. S., Kerschmann, R. L., and Win- terhalter, D. (Eds.). (2018).Dust in the At- mosphere of Mars and its Impact on Human Exploration. Cambridge Scholars Publishing

    2018

  28. [28]

    Hapgood, M. (2019). The Impact of Space Weather on Human Missions to Mars: The Need for Good Engineering and Good Fore- casts. In: Szocik, K. (Eds)The Human Fac- tor in a Mission to Mars. Space and Society. Springer, Cham

    2019

  29. [29]

    Hersbach, H., Bell, B., Berrisford, P., et al. (2020). The ERA5 global reanalysis.Quar- terly Journal of the Royal Meteorological So- ciety, 146(730):1999–2049

    2020