Pith. sign in

REVIEW 3 major objections 5 minor 39 references

Resolution Revolution: A Physics-Guided Deep Learning Framework for Spatiotemporal Temperature Reconstruction

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a single physics-guided convolutional network can reconstruct daily, all-weather land surface temperature at 30 m and 2 km with 2–3 K accuracy.

desk verdict Credible integration of known components into an end-to-end model, but the all-weather claim rests on a time-invariant residual that the paper never tests under cloudy conditions. read the letter →

arxiv 2507.09872 v1 pith:FHIOBAWF submitted 2025-07-14 eess.IV cs.CV

Shengjie Liu , Lu Zhang , Siqin Wang This is my paper · ORCID
classification eess.IVcs.CV
keywords landsurfacetemperaturereconstructionphysics-guideddeeplearningannualcycleERA5reanalysisconvolutionalneuralnetworkall-weatherLSTGOES-16
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 tries to establish that one end-to-end deep learning model, not a hand-crafted multi-stage pipeline, can produce seamless daily land surface temperature at very high spatial resolution. The model combines three additive components: a pixel-wise annual temperature cycle, a linear scaling of coarse ERA5 reanalysis temperature to capture daily fluctuations, and a U-Net-style convolutional network that learns residual spatial structure from surface reflectance. This design keeps the physically meaningful temporal behavior explicit while letting the neural network only fill in what the physics terms miss. If true, it closes the gap between high-resolution but cloud-gapped satellite observations and coarse but continuous reanalysis data, which would benefit climate-health studies, urban heat analysis, and ecological monitoring.

What carries the argument

The load-bearing mechanism is the additive decomposition of Eq. (6): reconstructed temperature equals a pixel-wise annual temperature cycle (a cosine with learnable mean, amplitude, and phase) plus a per-pixel linear scaling of coarse ERA5 temperature plus a residual field learned by a U-Net-style CNN from surface reflectance. The ATC and ERA5 branches supply physically plausible seasonal and daily variability, so the CNN only learns spatial patterns that those branches cannot express. This is what enables end-to-end training with heavily missing observations: the masked L1 loss penalizes only observed pixels, preventing the network from overfitting to sparse clear-sky dates.

What would settle it

Choose a cloud-prone area with several in situ stations, train on clear-sky Landsat or GOES-16 pairs, and compare reconstructed cloudy-day temperatures against stations on days when the coarse ERA5 field places a temperature gradient opposite to the fine-scale pattern; if the model's error jumps on those days while clear-sky error stays low, the time-invariant reflectance residual is the failing component.

Watch

Extended reading notes

Core claim

The central discovery is that temperature reconstruction can be reformulated as a single additive model embedded inside a CNN, rather than separate temporal and spatial modeling steps. The reconstruction is the sum of a learned ATC cosine per pixel, a per-pixel linear amplification of resampled ERA5 temperature, and a residual surface produced by a U-Net from surface reflectance. Because the loss is masked to observed pixels only, the whole model trains end-to-end without manual thresholds. On the four datasets, the method reaches hold-out MAE near 1.8 K and RMSE near 2.5–2.7 K for GOES-16, in situ RMSE of 3.44 K and 4.42 K for 30 m Landsat reconstructions at two sites, and, on GOES-16, better agreement with in situ measurements than the cloud-free satellite observations themselves.

Load-bearing premise

The load-bearing premise is that the residual correction network, which takes only a surface reflectance image and no time information, can correct the ERA5-scaled field equally well on cloudy days and across all seasons; if fine-scale temperature patterns under clouds are not a stationary function of reflectance, the cloudy-day fields will simply inherit ERA5's coarse spatial shape.

Editorial extensions

If this is right

  • Daily 30 m land surface temperature can be reconstructed for every day of the year from only tens of clear-sky Landsat scenes, with in situ RMSE of 3.44 K at the PSU site and 4.42 K at the BON site.
  • Hourly all-weather LST at 2 km can be reconstructed with hold-out MAE near 1.8 K, roughly half the error of the ATC-only model on the same test split.
  • Cloudy-day reconstructions can agree better with ground-station measurements than direct cloud-free satellite observations, because the model estimates temperature where the sensor sees clouds.
  • End-to-end single-model training replaces hand-crafted two-stage approaches, removing manual thresholding and the separate temporal/spatial modeling steps used by earlier methods.

Reading between the lines

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

  • A global deployment would need to prove that the reflectance-only residual branch generalizes beyond the two mid-latitude U.S. sites tested; snow, desert, irrigation, and complex terrain may produce time-varying reflectance-to-temperature relationships that the current architecture cannot express.
  • A directly testable extension is to make the CNN residual time-dependent by adding an auxiliary surface-state band such as snow cover or vegetation index, which would address the stationary-residual assumption without changing the physics branches.
  • The same additive architecture could be re-targeted to hourly 30 m temperature by replacing the annual cosine with a diurnal cycle, using GOES-16's hourly observations as training signal and Landsat for fine spatial detail.
  • The result that reconstructions beat cloud-free observations against in situ data suggests that evaluations of all-weather LST products should be reported separately for cloudy and clear days, since the error mechanisms differ.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes a physics-guided deep learning framework for reconstructing daily, all-weather land surface temperature at high spatial resolution (30 m for Landsat, 2 km for GOES-16). The model combines three additive components: a pixel-wise annual temperature cycle (ATC), a pixel-wise linear scaling of ERA5 reanalysis temperature to capture daily fluctuations, and a U-Net-based CNN that maps surface reflectance to a spatiotemporal residual field. Training uses an L1 loss evaluated only on observed (cloud-free) pixels. The method is evaluated on four datasets (two Landsat sites and two GOES-16 regions) with hold-out validation for GOES-16 and in situ validation against SURFRAD stations for all datasets, reporting 2–3 K accuracy and better agreement with in situ data than cloud-free satellite observations. The central claim is that this is the first end-to-end, single deep learning model for seamless all-weather spatiotemporal temperature reconstruction.

Significance. If the all-weather claim holds, the framework addresses a real and important gap: the trade-off between spatial and temporal resolution in satellite thermal remote sensing, with direct applications in climate, health, and urban studies. The architectural decomposition into physically interpretable components (ATC, ERA5 linear term, residual CNN) is a sensible way to inject physical structure while retaining end-to-end trainability, and the paper is generally clearly written. The evaluation is substantial: four datasets, two sensors, hold-out and in situ validation, and quantitative comparison with two existing methods. A notable strength is that the proposed method achieves lower RMSE against in situ measurements than the actual cloud-free satellite observations, an interesting and non-obvious result. However, the central all-weather contribution hinges on the assumption that a time-invariant, reflectance-only residual can correct cloudy-day temperatures, an assumption that is not directly tested; the headline 'first end-to-end' novelty claim also needs more support.

major comments (3)
  1. [§3.3, Eq. (10), Eq. (8), Eq. (11), Fig. 1] The residual term M_resid in Eq. (10) is a function only of the surface reflectance tensor X and has no time or cloud-cover input; the U-Net in Fig. 1 receives a single annual-mean reflectance image. All day-to-day variability beyond the annual cosine in Eq. (7) is therefore carried by the pixel-wise linear term w_ij * Tc(i,j,t) in Eq. (8). Because the loss in Eq. (11) is evaluated only at observed clear-sky pixels, both w_ij and the static residual are fitted under clear-sky conditions and cannot, by construction, adapt to cloudy conditions. The in situ evaluations in Tables 1 and 2 and Fig. 4 aggregate all days, so they do not isolate a cloudy-day test; on cloudy days the reconstruction is essentially the ERA5 temporal shape plus a static spatial correction. The all-weather claim therefore presumes that the relationship between fine-scale LST and coarse ERA5 temperature is stationary across cloud states and seasons, which is not demonstrated. I recommend reporting in situ error stratified by cloud cover, or adding a time/cloud-state input to the residual branch, or explicitly limiting the all-weather claim to the demonstrated clear-sky-corrected ERA5 extrapolation.
  2. [Tables 1 and 2] The comparison with prior methods is not apples-to-apples: the proposed method is evaluated on 362/359 days at PSU/BON, whereas the 4-step ATC competitor is evaluated on 48/20 days and GEC-SEB on 229/222 days. The reported RMSE improvements therefore conflate method performance with test-set composition. To support the claim of state-of-the-art reconstruction, the authors should report metrics on the intersection of days available to all methods, or clearly state that the competitors cannot produce daily all-weather outputs and compare only on the day sets each method supports. Without this, the accuracy advantage is not yet established.
  3. [Conclusion, §1] The conclusion claims 'achieving, for the first time, end-to-end temperature reconstruction based on a single deep learning model.' This is overstated as written: the model is a hybrid of an analytically parameterized ATC, a linear ERA5 scaling, and a CNN residual, with the ATC parameters obtained via optimization rather than through end-to-end backpropagation as part of a single differentiable model. No details are given on how gradients flow into the ATC branch. Please clarify what 'end-to-end' means in this architecture and temper the novelty claim, or provide evidence that prior two-step approaches cannot be trained jointly in the same manner.
minor comments (5)
  1. [§4.3, Eq. (11)] The text 'As we are optimizing the F1 loss (MAE)' should read 'L1 loss'; F1 is a classification metric and this is a typo.
  2. [§4.1] The years of the two GOES-16 datasets are inconsistent: the first dataset is stated as 2022, while the second dataset is not given a year; Fig. 4 shows N=364, which suggests 2023. Please clarify the years for both GOES-16 datasets.
  3. [Tables 1 and 2] The competitor rows list RMSE and bias but not MAE; since MAE is the primary metric used elsewhere, please report MAE for the competing methods as well, or explain its omission.
  4. [Fig. 4] The 'Valid Cloud-free Observations' plots are compared with the reconstruction, but the sample sizes differ (e.g., N=145 vs N=364 at PSU); a fairer comparison would restrict both to the same days, or report the comparison over the common subset.
  5. [§2.2, reference [22]] The prior work of Liu et al. [22] is described as achieving 1.48–2.11 K accuracy for daily 30 m reconstruction, while the proposed method achieves 2.75–3.43 K MAE on the Landsat sites; the discussion should directly address how the proposed method improves on that baseline beyond the architectural differences.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: all fitted components are trained with a masked loss and evaluated on held-out and in situ data, so the claimed reconstruction does not reduce to its inputs by construction.

full rationale

The paper's derivation chain is self-contained with respect to circularity. The reconstructed field is the sum of an ATC fitted from the incomplete target temperature tensor (Eq. 7), a per-pixel linear amplification of resampled ERA5 (Eq. 8), and a U-Net residual learned from surface reflectance (Eq. 10); the loss (Eq. 11) is evaluated only on observed pixels. Hold-out GOES-16 validation (20% of valid observed data) and independent SURFRAD in situ comparisons are used, so the reported metrics are not computed by re-inserting the fitted values into the target. No step exhibits a definitional reduction (e.g., a fitted parameter renamed as a prediction, or a result equivalent to its input by construction). The only self-citations (refs 20, 21, 22) appear as related work and do not carry the central claim. The limitation that M_resid has no time or cloud-state argument is a generalization or identifiability concern about all-weather claims, not a circularity pattern: the model does not define its target in terms of its outputs.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central reconstruction rests on four modeling assumptions: a single-harmonic ATC, a per-pixel linear link to ERA5, a time-invariant reflectance-based residual, and generalization from clear-sky training to all-weather use. No new physical entities are introduced; the fitted parameters are standard per-pixel coefficients and neural network weights.

free parameters (3)
  • ATC parameters (a_ij, b_ij, phi_ij) per pixel = fitted from incomplete satellite observations
    The annual mean, amplitude, and phase of the cosine in Eq. 7 are learned per pixel from the observed temperature values used in training.
  • ERA5 linear scale w_ij per pixel = learned via gradient descent
    Eq. 8 multiplies the resampled ERA5 temperature by a pixel-wise weight; this weight is the core 'amplification' term and is fitted to match satellite LST on clear-sky observations.
  • U-Net weights (fCNN) = trained
    Standard neural network parameters in Eq. 10, learned on reflectance features.
assumptions (5)
  • domain assumption ATC single-cosine with period T=365 days captures the annual thermal trend (Eq. 7).
    The seasonal cycle is represented by one harmonic; multi-harmonic or weather-driven deviations are left to the residual and ERA5 terms.
  • domain assumption Fine-scale LST is a linear function of resampled coarse ERA5 temperature at each pixel (Eq. 8).
    The daily fluctuation branch uses only a per-pixel scaling without offset or nonlinearity; this assumes a stationary linear relationship between ERA5 and LST.
  • ad hoc to paper The spatiotemporal residual depends only on surface reflectance X and has no explicit time index (Eq. 10, Fig. 1).
    The U-Net input is a single reflectance image, so the residual surface is constant across time; day-specific deviations beyond the ATC and ERA5 scaling cannot be captured.
  • domain assumption The clear-sky-trained mapping generalizes to cloudy and all-weather conditions.
    Loss (Eq. 11) uses only observed cloud-free pixels; all-weather in situ validation tests this, but the model never sees cloudy-sky LST in training.
  • standard math Missing values are ignorable; the L1 loss over observed entries is a sufficient training signal (Eq. 11).
    The mask M excludes NaN entries; this assumes the observed pixels are representative for optimizing the reconstruction on unobserved days.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Resolution Revolution: A Physics-Guided Deep Learning Framework for Spatiotemporal Temperature Reconstruction." pith.science (2026). https://pith.science/paper/FHIOBAWF

@misc{pith2026250709872,
  author       = {Pith},
  title        = {Pith review of: Resolution Revolution: A Physics-Guided Deep Learning Framework for Spatiotemporal Temperature Reconstruction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FHIOBAWF}},
  note         = {Machine review of arXiv:2507.09872}
}
read the original abstract

Central to Earth observation is the trade-off between spatial and temporal resolution. For temperature, this is especially critical because real-world applications require high spatiotemporal resolution data. Current technology allows for hourly temperature observations at 2 km, but only every 16 days at 100 m, a gap further exacerbated by cloud cover. Earth system models offer continuous hourly temperature data, but at a much coarser spatial resolution (9-31 km). Here, we present a physics-guided deep learning framework for temperature data reconstruction that integrates these two data sources. The proposed framework uses a convolutional neural network that incorporates the annual temperature cycle and includes a linear term to amplify the coarse Earth system model output into fine-scale temperature values observed from satellites. We evaluated this framework using data from two satellites, GOES-16 (2 km, hourly) and Landsat (100 m, every 16 days), and demonstrated effective temperature reconstruction with hold-out and in situ data across four datasets. This physics-guided deep learning framework opens new possibilities for generating high-resolution temperature data across spatial and temporal scales, under all weather conditions and globally.

Figures

Figures reproduced from arXiv: 2507.09872 by the authors.

Figure 1
Figure 1. Overview of the proposed method. The annual trend and daily fluctuation are derived from physics-guided models (ATC and [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Comparison of reconstructed LST of Landsat (30 m) at PSU and BON sites from four models (ATC, Na¨ [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Reconstructed daily temperature maps at 30 m resolution from Landsat data. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Comparison of GOES-16 LST reconstructions from ATC, Na¨ [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Reconstructed temperature maps of the GOES-16 data (East Coast including PSU). [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 37 canonical work pages

  1. [1]

    A new ther- mal fusion method to downscale land surface temperature to finer spatial resolution using sentinel-msi and landsat-oli/tirs imagery

    Hicham Bahi, Lahouari Bounoua, Anas Sabri, Abderrazak Bannari, Anass Malah, and Hassan Rhinane. A new ther- mal fusion method to downscale land surface temperature to finer spatial resolution using sentinel-msi and landsat-oli/tirs imagery. Remote Sensing Applications: Society and Envi- ronment, 37:101519, 2025. 2

    work page 2025

  2. [2]

    Compar- ison of diurnal variation of land surface temperature from goes-16 abi and modis instruments

    Christopher Beale, Hamid Norouzi, Zahra Sharifnezhadaz- izi, Abdou Rachid Bah, Peng Yu, Yunyue Yu, Reginald Blake, Anna Vaculik, and Jorge Gonzalez-Cruz. Compar- ison of diurnal variation of land surface temperature from goes-16 abi and modis instruments. IEEE Geoscience and Remote Sensing Letters, 17(4):572–576, 2019. 5

    work page 2019

  3. [3]

    Reconstructing historical climate fields with deep learning

    Nils Bochow, Anna Poltronieri, Martin Rypdal, and Niklas Boers. Reconstructing historical climate fields with deep learning. Science Advances, 11(14):eadp0558, 2025. 2

    work page 2025

  4. [4]

    Estimating the optimal broadband emissivity spectral range for calculating surface longwave net radiation

    Jie Cheng, Shunlin Liang, Yunjun Yao, and Xiaotong Zhang. Estimating the optimal broadband emissivity spectral range for calculating surface longwave net radiation. IEEE Geo- science and Remote Sensing Letters , 10(2):401–405, 2012. 5

    work page 2012

  5. [5]

    Reconstruction of hourly all-weather land surface temperature by integrating reanaly- sis data and thermal infrared data from geostationary satel- lites (rtg)

    Lirong Ding, Ji Zhou, Zhao-Liang Li, Jin Ma, Chunxiang Shi, Shuai Sun, and Ziwei Wang. Reconstruction of hourly all-weather land surface temperature by integrating reanaly- sis data and thermal infrared data from geostationary satel- lites (rtg). IEEE Transactions on Geoscience and Remote Sensing, 60:1–17, 2022. 1

    work page 2022

  6. [6]

    Wenhui Du, Zhao-Liang Li, Zhihao Qin, Jinlong Fan, Xi- angyang Liu, Chunliang Zhao, and Kun Cao. Reconstruction of cloudy land surface temperature by combining surface en- ergy balance theory and solar-cloud-satellite geometry.IEEE Transactions on Geoscience and Remote Sensing, 2025. 5, 7

    work page 2025

  7. [7]

    Representation of hetero- geneity effects in earth system modeling: Experience from land surface modeling

    Filippo Giorgi and Roni Avissar. Representation of hetero- geneity effects in earth system modeling: Experience from land surface modeling. Reviews of Geophysics, 35(4):413– 437, 1997. 1

    work page 1997

  8. [8]

    Global surface temperature change

    James Hansen, Reto Ruedy, Mki Sato, and Ken Lo. Global surface temperature change. Reviews of Geophysics, 48(4),

    work page

Show all 39 references
  1. [9]

    Spatial-temporal super-resolution of satellite im- agery via conditional pixel synthesis

    Yutong He, Dingjie Wang, Nicholas Lai, William Zhang, Chenlin Meng, Marshall Burke, David Lobell, and Stefano Ermon. Spatial-temporal super-resolution of satellite im- agery via conditional pixel synthesis. Advances in Neural Information Processing Systems, 34:27903–27915, 2021. 2

    2021

  2. [10]

    Denoising and inpainting of sea surface temperature image with adversarial physical model loss

    Nobuyuki Hirahara, Motoharu Sonogashira, Hidekazu Kasa- hara, and Masaaki Iiyama. Denoising and inpainting of sea surface temperature image with adversarial physical model loss. In Asian Conference on Pattern Recognition , pages 339–352. Springer, 2019. 2

    2019

  3. [11]

    The next landsat satellite: The landsat data continuity mission

    James R Irons, John L Dwyer, and Julia A Barsi. The next landsat satellite: The landsat data continuity mission. Re- mote Sensing of Environment, 122:11–21, 2012. 1

    2012

  4. [12]

    Advances in methodology and generation of all-weather land surface temperature products from polar-orbiting and geostationary satellites: A comprehensive review

    Aolin Jia, Shunlin Liang, Dongdong Wang, Kanishka Mallick, Shugui Zhou, Tian Hu, and Shuo Xu. Advances in methodology and generation of all-weather land surface temperature products from polar-orbiting and geostationary satellites: A comprehensive review. IEEE Geoscience and R...

    2024

  5. [13]

    Artificial intelligence reconstructs missing climate informa- tion

    Christopher Kadow, David Matthew Hall, and Uwe Ulbrich. Artificial intelligence reconstructs missing climate informa- tion. Nature Geoscience, 13(6):408–413, 2020. 2

    2020

  6. [14]

    Spatial and temporal dis- tribution of clouds observed by modis onboard the terra and aqua satellites

    Michael D King, Steven Platnick, W Paul Menzel, Steven A Ackerman, and Paul A Hubanks. Spatial and temporal dis- tribution of clouds observed by modis onboard the terra and aqua satellites. IEEE transactions on geoscience and remote sensing, 51(7):3826–3852, 2013. 1

    2013

  7. [15]

    Uncertainty estimation method and landsat 7 global validation for the landsat surface temperature product

    Kelly G Laraby and John R Schott. Uncertainty estimation method and landsat 7 global validation for the landsat surface temperature product. Remote Sensing of Environment, 216: 472–481, 2018. 5

    2018

  8. [16]

    Lfsr: Low-resolution filling then super-resolution re- construction framework for gapless all-weather modis-like land surface temperature generation

    Chan Li, Penghai Wu, Si-Bo Duan, Yixuan Jia, Shuai Sun, Chunxiang Shi, Zhixiang Yin, Huifang Li, and Huanfeng Shen. Lfsr: Low-resolution filling then super-resolution re- construction framework for gapless all-weather modis-like land surface temperature generation. Remote Sens...

    2025

  9. [17]

    Evaluation of machine learning algorithms in spatial down- scaling of modis land surface temperature

    Wan Li, Li Ni, Zhao-liang Li, Si-Bo Duan, and Hua Wu. Evaluation of machine learning algorithms in spatial down- scaling of modis land surface temperature. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(7):2299–2307, 2019. 1

    2019

  10. [18]

    Climatenerf: Extreme weather synthesis in neural radiance field

    Yuan Li, Zhi-Hao Lin, David Forsyth, Jia-Bin Huang, and Shenlong Wang. Climatenerf: Extreme weather synthesis in neural radiance field. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3227– 3238, 2023. 2

    2023

  11. [19]

    Satellite remote sensing of global land sur- face temperature: Definition, methods, products, and appli- cations

    Zhao-Liang Li, Hua Wu, Si-Bo Duan, Wei Zhao, Huazhong Ren, Xiangyang Liu, Pei Leng, Ronglin Tang, Xin Ye, Jin- shun Zhu, et al. Satellite remote sensing of global land sur- face temperature: Definition, methods, products, and appli- cations. Reviews of Geophysics, 61(1), 2023. 1, 2

    2023

  12. [20]

    Deep feature gaussian processes for single-scene aerosol optical depth reconstruction

    Shengjie Liu and Lu Zhang. Deep feature gaussian processes for single-scene aerosol optical depth reconstruction. IEEE Geoscience and Remote Sensing Letters, 2024. 2

    2024

  13. [21]

    Spatial vari- ability of diurnal temperature range and its associations with local climate zone, neighborhood environment and mortality in los angeles

    Shengjie Liu, An-Min Wu, and Hung Chak Ho. Spatial vari- ability of diurnal temperature range and its associations with local climate zone, neighborhood environment and mortality in los angeles. Urban Climate, 49:101526, 2023. 1

    2023

  14. [22]

    Daily land surface temperature reconstruction in landsat cross-track areas us- ing deep ensemble learning with uncertainty quantification

    Shengjie Liu, Siqin Wang, and Lu Zhang. Daily land surface temperature reconstruction in landsat cross-track areas us- ing deep ensemble learning with uncertainty quantification. arXiv preprint arXiv:2502.14433, 2025. 1, 2

    2025

  15. [23]

    Optically enhanced super- resolution of sea surface temperature using deep learning

    David T Lloyd, Aaron Abela, Reuben A Farrugia, Anthony Galea, and Gianluca Valentino. Optically enhanced super- resolution of sea surface temperature using deep learning. IEEE Transactions on Geoscience and Remote Sensing , 60: 1–14, 2021. 2

    2021

  16. [24]

    Generation of modis-like land surface tem- peratures under all-weather conditions based on a data fu- sion approach

    Di Long, La Yan, Liangliang Bai, Caijin Zhang, Xueying Li, Huimin Lei, Hanbo Yang, Fuqiang Tian, Chao Zeng, Xiany- ong Meng, et al. Generation of modis-like land surface tem- peratures under all-weather conditions based on a data fu- sion approach. Remote Sensing of Environmen...

  17. [25]

    Pangaea: A global and inclusive benchmark for geospatial foundation models

    Valerio Marsocci, Yuru Jia, Georges Le Bellier, David Kerekes, Liang Zeng, Sebastian Hafner, Sebastian Gerard, Eric Brune, Ritu Yadav, Ali Shibli, et al. Pangaea: A global and inclusive benchmark for geospatial foundation models. arXiv preprint arXiv:2412.04204, 2024. 8

    2024 arXiv

  18. [26]

    Deep learning to represent subgrid processes in climate mod- els

    Stephan Rasp, Michael S Pritchard, and Pierre Gentine. Deep learning to represent subgrid processes in climate mod- els. Proceedings of the National Academy of Sciences , 115 (39):9684–9689, 2018. 2

    2018

  19. [27]

    Im- proving land surface temperature estimation in cloud cover scenarios using graph-based propagation

    Iain Rolland, Sivasakthy Selvakumaran, Shaikh Fairul Edros Ahmad Shaikh, Perrine Hamel, and Andrea Marinoni. Im- proving land surface temperature estimation in cloud cover scenarios using graph-based propagation. Geophysical Re- search Letters, 51(23):e2024GL108263, 2024. 2

    2024

  20. [28]

    T. J. Schmit, S. S. Lindstrom, J. J. Gerth, and M. M. Gun- shor. Applications of the 16 spectral bands on the advanced baseline imager (abi). Journal of Operational Meteorology, 6(4):33–46, 2018. 5

    2018

  21. [29]

    Evaluation of modis land surface tem- perature data to estimate air temperature in different ecosys- tems over africa

    Christelle Vancutsem, Pietro Ceccato, Tufa Dinku, and Stephen J Connor. Evaluation of modis land surface tem- perature data to estimate air temperature in different ecosys- tems over africa. Remote Sensing of Environment , 114(2): 449–465, 2010. 7

    2010

  22. [30]

    Environmental cooling provided by urban trees under extreme heat and cold waves in us cities

    Chenghao Wang, Zhi-Hua Wang, Chuyuan Wang, and Soe W Myint. Environmental cooling provided by urban trees under extreme heat and cold waves in us cities. Remote Sensing of Environment, 227:28–43, 2019. 1

    2019

  23. [31]

    Artificial intelligence achieves easy-to-adapt nonlinear global temper- ature reconstructions using minimal local data

    Martin Wegmann and Fernando Jaume-Santero. Artificial intelligence achieves easy-to-adapt nonlinear global temper- ature reconstructions using minimal local data. Communica- tions Earth & Environment, 4(1):217, 2023. 2

    2023

  24. [32]

    Satellite observations and malaria: new opportunities for research and applications

    Michael C Wimberly, Kirsten M de Beurs, Tatiana V Lo- boda, and William K Pan. Satellite observations and malaria: new opportunities for research and applications. Trends in Parasitology, 37(6):525–537, 2021. 1

    2021

  25. [33]

    Spatially continuous and high- resolution land surface temperature product generation: A review of reconstruction and spatiotemporal fusion tech- niques

    Penghai Wu, Zhixiang Yin, Chao Zeng, Si-Bo Duan, Frank-Michael G ¨ottsche, Xiaoshuang Ma, Xinghua Li, Hui Yang, and Huanfeng Shen. Spatially continuous and high- resolution land surface temperature product generation: A review of reconstruction and spatiotemporal fusion tech- ...

    2021

  26. [34]

    Generative image inpainting with con- textual attention

    Jiahui Yu, Zhe Lin, Jimei Yang, Xiaohui Shen, Xin Lu, and Thomas S Huang. Generative image inpainting with con- textual attention. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages 5505– 5514, 2018. 2

    2018

  27. [35]

    A global seamless 1 km resolution daily land surface temperature dataset (2003–2020)

    Tao Zhang, Yuyu Zhou, Zhengyuan Zhu, Xiaoma Li, and Ghassem R Asrar. A global seamless 1 km resolution daily land surface temperature dataset (2003–2020). Earth System Science Data Discussions, 2021:1–16, 2021. 2

    2003

  28. [36]

    A practical reanalysis data and thermal infrared re- mote sensing data merging (rtm) method for reconstruction of a 1-km all-weather land surface temperature

    Xiaodong Zhang, Ji Zhou, Shunlin Liang, and Dongdong Wang. A practical reanalysis data and thermal infrared re- mote sensing data merging (rtm) method for reconstruction of a 1-km all-weather land surface temperature. Remote Sensing of Environment, 260:112437, 2021. 1

    2021

  29. [37]

    Hourly mapping of sur- face air temperature by blending geostationary datasets from the two-satellite system of goes-r series

    Zhenwei Zhang and Qingyun Du. Hourly mapping of sur- face air temperature by blending geostationary datasets from the two-satellite system of goes-r series. ISPRS Journal of Photogrammetry and Remote Sensing , 183:111–128, 2022. 1

    2022

  30. [38]

    Reconstruction of land surface temperature under cloudy conditions from landsat 8 data using annual temperature cycle model

    Xiaolin Zhu, Si-Bo Duan, Zhao-Liang Li, Penghai Wu, Hua Wu, Wei Zhao, and Yonggang Qian. Reconstruction of land surface temperature under cloudy conditions from landsat 8 data using annual temperature cycle model. Remote Sensing of Environment, 281:113261, 2022. 2, 5, 7

    2022

  31. [39]

    On the founda- tions of earth and climate foundation models

    Xiao Xiang Zhu, Zhitong Xiong, Yi Wang, Adam J Stewart, Konrad Heidler, Yuanyuan Wang, Zhenghang Yuan, Thomas Dujardin, Qingsong Xu, and Yilei Shi. On the founda- tions of earth and climate foundation models. arXiv preprint arXiv:2405.04285, 2024. 8 10

    2024 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.