pith. sign in

arxiv: 2605.19812 · v1 · pith:V2SSXAGGnew · submitted 2026-05-19 · 💻 cs.LG · cs.AI· stat.AP· stat.ML

FLUXtrapolation: A benchmark on extrapolating ecosystem fluxes

Anya Fries , Jacob A Nelson , Martin Jung , Markus Reichstein , Jonas Peters This is my paper

Pith reviewed 2026-05-20 07:16 UTC · model grok-4.3

classification 💻 cs.LG cs.AIstat.APstat.ML
keywords ecosystem fluxesdistribution shiftbenchmarkmachine learningflux upscalingdomain generalizationtail errorscarbon cycle
0
0 comments X

The pith

FLUXtrapolation creates a benchmark that tests machine learning extrapolation of ecosystem fluxes under realistic shifts in climate, space, and conditions.

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

This paper introduces FLUXtrapolation, a benchmark for evaluating how well models extrapolate ecosystem fluxes to locations without direct measurements. Fluxes of carbon, water, and energy are observed only at sparse towers, so global estimates require training on seen sites and predicting under changes in covariates and unobserved drivers. The benchmark draws on domain expertise to define temporal, spatial, and temperature-based extrapolation tasks and measures performance on held-out domains, multiple time scales, and tail errors. A pilot study shows that standard median hourly RMSE keeps baselines close together, while the new evaluations separate them. The work therefore supplies a concrete testbed whose solutions would improve global flux estimates for cycle studies.

Core claim

FLUXtrapolation supplies a benchmark that quantifies shifts in both the covariate distribution P(X) and the conditional distribution P(Y|X) for flux upscaling. It organizes extrapolation around temporal, spatial, and temperature scenarios and scores models across held-out domains, temporal aggregations, and tail errors, exposing performance gaps that median hourly RMSE conceals.

What carries the argument

The FLUXtrapolation benchmark, which structures progressively harder extrapolation scenarios around temporal, spatial, and temperature criteria and evaluates with tail-focused and multi-scale metrics.

If this is right

  • Models that succeed on the benchmark would produce more reliable global maps of carbon, water, and energy fluxes.
  • The emphasis on tail errors would encourage methods that handle extreme flux events accurately.
  • Progress on the benchmark would directly advance scientific estimates of ecosystem contributions to climate cycles.
  • The multi-scale evaluation would push models toward consistent performance across hourly to seasonal resolutions.

Where Pith is reading between the lines

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

  • The benchmark structure could be adapted to test generalization in other sparse environmental sensing domains such as soil moisture or biodiversity mapping.
  • Prioritizing robustness to the benchmark's shifts might improve downstream applications like carbon accounting or drought forecasting.
  • The observed separation under tail metrics suggests that future models should incorporate explicit handling of rare but high-impact flux conditions.

Load-bearing premise

The chosen temporal, spatial, and temperature-based extrapolation scenarios together with the tail and multi-scale metrics accurately reflect the covariate and conditional shifts that arise when upscaling fluxes from sparse towers to global scales.

What would settle it

A larger study in which standard median hourly RMSE separates model performances as clearly as the benchmark's tail and multi-scale metrics, or empirical evidence that real flux upscaling errors do not track the defined shift types.

Figures

Figures reproduced from arXiv: 2605.19812 by Anya Fries, Jacob A Nelson, Jonas Peters, Markus Reichstein, Martin Jung.

Figure 1
Figure 1. Figure 1: Flux tower data and the upscaling challenge. (a) An eddy covariance tower, which measures [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the three extrapolation scenarios and the corresponding distribution shifts. Top row: train–test splits. Middle row: we quantify covariate shift in PX by the balanced accuracy of a domain classifier distinguishing training from test inputs in the full covariate space; the plots illustrate this shift through NDWI density plots. Bottom row: we quantify conditional shift in PY |X by the percentage… view at source ↗
Figure 3
Figure 3. Figure 3: Cumulative distribution of weekly site-level RMSE for temperature-based extrapolation. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Overview of tower locations in space and time. Points colored corresponding to aggregated [PITH_FULL_IMAGE:figures/full_fig_p022_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Autocorrelation functions for ET, GPP, and NEE at site AU-Cum, shown at hourly, daily, [PITH_FULL_IMAGE:figures/full_fig_p024_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Model-based illustration of conditional shift in the VPD–ET relationship across extrapola [PITH_FULL_IMAGE:figures/full_fig_p026_6.png] view at source ↗
read the original abstract

We introduce FLUXtrapolation, a benchmark for extrapolating ecosystem fluxes under progressively harder distribution shifts. Ecosystem fluxes are central to understanding the carbon, water, and energy cycles, yet they can only be measured directly at sparsely located measurement towers. Producing global flux estimates therefore requires training models on observed sites using globally available covariates and predicting in unobserved regions, that is, upscaling. Flux upscaling is a challenging domain generalization problem that is affected by a shift in covariate distribution across climates, ecosystem types, and environmental conditions, as well as by conditional shift: important drivers remain unobserved at global scale. We provide a quantitative analysis of both these shifts in $P_X$ and $P_{Y\mid X}$. FLUXtrapolation is designed based on domain expertise on flux upscaling: it defines temporal, spatial, and temperature-based extrapolation scenarios and evaluates performance across held-out domains, temporal aggregations, and tail errors. In a pilot study, we find that baselines perform similarly under median hourly RMSE, but separate under the proposed tail-focused and multi-scale evaluation. FLUXtrapolation therefore poses a realistic and thus relevant challenge for machine learning methods under distribution shift; at the same time, progress on this benchmark would directly support the scientific goal of improving flux upscaling.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper introduces FLUXtrapolation, a benchmark for extrapolating ecosystem fluxes under progressively harder distribution shifts. It defines temporal, spatial, and temperature-based extrapolation scenarios using domain expertise on flux upscaling, provides quantitative analysis of shifts in both P_X (covariate) and P_{Y|X} (conditional), and evaluates performance across held-out domains, temporal aggregations, and tail errors. A pilot study finds that baselines perform similarly under median hourly RMSE but separate under the proposed tail-focused and multi-scale evaluation.

Significance. If the benchmark's scenarios and metrics are shown to induce and measure meaningful conditional shifts beyond covariate shifts, this work would provide a valuable, domain-grounded testbed for ML methods addressing distribution shift. Progress here would directly aid scientific flux upscaling for carbon, water, and energy cycle studies. The pilot differentiation on tail and multi-scale metrics is a positive indicator of practical relevance, though verification requires the full methods and shift analysis.

major comments (1)
  1. [Benchmark design and quantitative shift analysis sections] The central claim that the temporal, spatial, and temperature-based scenarios capture conditional shifts (P_{Y|X}) in addition to covariate shifts (P_X) is load-bearing for the benchmark's stated relevance to real-world flux upscaling. The quantitative analysis of these shifts (described in the abstract and presumably detailed in the benchmark design or results sections) must demonstrate statistically detectable changes in the conditional distribution of fluxes given observed global covariates for the held-out domains; otherwise the benchmark primarily tests covariate-shift generalization rather than the harder conditional-shift problem emphasized.
minor comments (1)
  1. [Abstract] The abstract summarizes the pilot study findings but provides no methods details, data sources, or error analysis; these should be expanded in the main text to allow verification that the reported separation under tail-focused metrics is robust.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and for emphasizing the importance of rigorously establishing conditional shifts in the benchmark. We address the major comment below and clarify the existing analysis while offering to strengthen its presentation.

read point-by-point responses

  1. Referee: [Benchmark design and quantitative shift analysis sections] The central claim that the temporal, spatial, and temperature-based scenarios capture conditional shifts (P_{Y|X}) in addition to covariate shifts (P_X) is load-bearing for the benchmark's stated relevance to real-world flux upscaling. The quantitative analysis of these shifts (described in the abstract and presumably detailed in the benchmark design or results sections) must demonstrate statistically detectable changes in the conditional distribution of fluxes given observed global covariates for the held-out domains; otherwise the benchmark primarily tests covariate-shift generalization rather than the harder conditional-shift problem emphasized.

    Authors: We agree that demonstrating statistically detectable changes in P_{Y|X} is central to the benchmark's relevance. The manuscript already includes a quantitative analysis of both P_X and P_{Y|X} shifts in the benchmark design section. We quantify conditional shifts by comparing the distribution of observed fluxes conditioned on the same global covariates (e.g., meteorology and land cover) across held-out domains in each scenario. Differences arise from domain-specific unobserved drivers, which we measure via divergence metrics and direct comparison of conditional quantiles and tail behavior. These analyses show that, even when covariate overlap exists, the conditional flux distributions differ significantly between training and held-out domains, particularly in the temperature-based and spatial extrapolation settings. We can expand this section with additional formal statistical tests (e.g., conditional Kolmogorov-Smirnov or permutation tests) in a revision to make the detectability more explicit. revision: partial

Circularity Check

0 steps flagged

No significant circularity: benchmark design is self-contained

full rationale

This paper introduces an evaluation benchmark rather than deriving a mathematical result or prediction from first principles. The temporal, spatial, and temperature-based extrapolation scenarios are defined using domain expertise on flux upscaling, with quantitative analysis provided for shifts in P_X and P_{Y|X}. No equations, fitted parameters, or self-citations reduce by construction to the target claims; the central premise that the benchmark poses a realistic challenge rests on the explicit scenario definitions and held-out domain evaluations rather than any self-referential loop. This is the expected outcome for a benchmark paper whose value is assessed against external data and ML baselines.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No mathematical derivations, free parameters, or new entities are described in the abstract; the work rests on domain expertise for scenario design.

pith-pipeline@v0.9.0 · 5770 in / 1079 out tokens · 58541 ms · 2026-05-20T07:16:42.124284+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

  • IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    We provide a quantitative analysis of both these shifts in PX and PY|X. FLUXtrapolation is designed based on domain expertise on flux upscaling: it defines temporal, spatial, and temperature-based extrapolation scenarios and evaluates performance across held-out domains, temporal aggregations, and tail errors.

  • IndisputableMonolith/Foundation/BranchSelection.lean branch_selection unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    We measure covariate shift by the balanced classifier accuracy distinguishing training from test inputs... Conditional shift we quantify by the increase in RMSE from training to test after accounting for differences in PX.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

241 extracted references · 241 canonical work pages · 3 internal anchors

  1. [1]

    Arain, M. A. (2025a). AmeriFlux FLUXNET-1F CA-TP3 Ontario - Turkey Point 1974 Plantation White Pine. doi:10.17190/AMF/1881566

    work page doi:10.17190/amf/1881566 1974

  2. [2]

    Arain, M. A. (2025b). AmeriFlux FLUXNET-1F CA-TPD Ontario - Turkey Point Mature Deciduous. doi:10.17190/AMF/1881567

    work page doi:10.17190/amf/1881567

  3. [3]

    Arain, M. A. (2026). AmeriFlux FLUXNET-1F CA-TP1 Ontario - Turkey Point 2002 Plantation White Pine. doi:10.17190/AMF/2006962

    work page doi:10.17190/amf/2006962 2026

  4. [4]

    and Baldocchi, D

    Arias-Ortiz, A. and Baldocchi, D. (2026). AmeriFlux FLUXNET-1F US-Dmg Dutch Slough Marsh Gilbert Tract. doi:10.17190/AMF/2469451

    work page doi:10.17190/amf/2469451 2026

  5. [5]

    Badiou, P. (2025). AmeriFlux FLUXNET-1F CA-EM1 Newdale Manitoba. doi:10.17190/AMF/2469436

    work page doi:10.17190/amf/2469436 2025

  6. [6]

    and Griffis, T

    Baker, J. and Griffis, T. (2025a). AmeriFlux FLUXNET-1F US-Ro5 Rosemount I18_South. doi:10.17190/AMF/1818371

    work page doi:10.17190/amf/1818371

  7. [7]

    and Griffis, T

    Baker, J. and Griffis, T. (2025b). AmeriFlux FLUXNET-1F US-Ro6 Rosemount I18_North. doi:10.17190/AMF/1881590

    work page doi:10.17190/amf/1881590

  8. [8]

    and Griffis, T

    Baker, J. and Griffis, T. (2026a). AmeriFlux FLUXNET-1F US-Ro2 Rosemount- C7. doi:10.17190/AMF/2204876

    work page doi:10.17190/amf/2204876

  9. [9]

    and Griffis, T

    Baker, J. and Griffis, T. (2026b). AmeriFlux FLUXNET-1F US-Ro4 Rosemount Prairie. doi:10.17190/AMF/1881589

    work page doi:10.17190/amf/1881589

  10. [10]

    Baker, J., Griffis, T., and Griffis, T. (2026). AmeriFlux FLUXNET-1F US-Ro1 Rosemount- G21. doi:10.17190/AMF/1881588

    work page doi:10.17190/amf/1881588 2026

  11. [11]

    and Windham-Myers, L

    Bergamaschi, B. and Windham-Myers, L. (2026). AmeriFlux FLUXNET-1F US-Srr Suisun marsh - Rush Ranch. doi:10.17190/AMF/2229389

    work page doi:10.17190/amf/2229389 2026

  12. [12]

    Bernhofer, C., Eichelmann, U., Gruenwald, T., Hehn, M., Mauder, M., Moderow, U., and Prasse, H. (2024a). Fluxnet Product from Grillenburg, 2016-12-31–2023-12-31. hdl:11676/reoAFD2eeGXxukbP7TpR6r6a

    work page 2016

  13. [13]

    Bernhofer, C., Eichelmann, U., Gruenwald, T., Hehn, M., Mauder, M., Moderow, U., and Prasse, H. (2024b). Fluxnet Product from Hetzdorf, 2009-12-31–2023-12-31. hdl:11676/yek1dgsd5zQboXAjB7HQ90FT

    work page 2009

  14. [14]

    Bernhofer, C., Eichelmann, U., Gruenwald, T., Hehn, M., Mauder, M., Moderow, U., and Prasse, H. (2024c). Fluxnet Product from Klingenberg, 2017-12-31–2023-12-31. hdl:11676/6Vfw-uhvsQGdw33rNTM9ZFjz

    work page 2017

  15. [15]

    Bernhofer, C., Eichelmann, U., Gruenwald, T., Hehn, M., and Moderow, U. (2024d). Fluxnet Product from Tharandt, 1995-12-31–2020-12-31. hdl:11676/TlbhTRRRRfSt8juxf_D7p-LM

    work page 1995

  16. [16]

    and Gruenwald, T

    Bernhofer, C. and Gruenwald, T. (2024). Fluxnet Product from Oberbarenburg, 2007-12-31–2023-12-31. hdl:11676/T5rNmm4p68XZYu1VivME6it6. 10

    work page 2024

  17. [17]

    Besnard, S., Koirala, S., Santoro, M., Weber, U., Nelson, J., Gütter, J., Herault, B., Kassi, J., N’Guessan, A., Neigh, C., Poulter, B., Zhang, T., and Carvalhais, N. (2021). Mapping global forest age from forest inventories, biomass and climate data.Earth System Science Data, 13(10):4881–4896

    work page 2021

  18. [18]

    Bickel, S., Brückner, M., and Scheffer, T. (2009). Discriminative learning under covariate shift.Journal of Machine Learning Research, 10(75):2137–2155

    work page 2009

  19. [19]

    Billesbach, D., Kueppers, L., Torn, M., and Biraud, S. (2026). AmeriFlux FLUXNET-1F US-A32 ARM-SGP Medford hay pasture. doi:10.17190/AMF/1881568

    work page doi:10.17190/amf/1881568 2026

  20. [20]

    Biraud, S., Fischer, M., Chan, S., and Torn, M. (2025). AmeriFlux FLUXNET-1F US-ARM ARM Southern Great Plains site- Lamont. doi:10.17190/AMF/1854366

    work page doi:10.17190/amf/1854366 2025

  21. [21]

    D., Monson, R

    Blanken, P. D., Monson, R. K., Burns, S. P., Bowling, D. R., and Turnipseed, A. A. (2026). AmeriFlux FLUXNET-1F US-NR1 Niwot Ridge Forest (LTER NWT1). doi:10.17190/AMF/1871141

    work page doi:10.17190/amf/1871141 2026

  22. [22]

    Bohrer, G. (2025). AmeriFlux FLUXNET-1F US-ORv Olentangy River Wetland Research Park. doi:10.17190/AMF/1832164

    work page doi:10.17190/amf/1832164 2025

  23. [23]

    Bowling, D., Kannenberg, S., and Anderegg, W. (2025). AmeriFlux FLUXNET-1F US-CdM Cedar Mesa. doi:10.17190/AMF/2469445

    work page doi:10.17190/amf/2469445 2025

  24. [24]

    Bracho, R., Celis, G., Rodenhizer, H., See, C., and Schuur, E. A. (2026). AmeriFlux FLUXNET-1F US-EML Eight Mile Lake Permafrost thaw gradient, Healy Alaska. doi:10.17190/AMF/2007170

    work page doi:10.17190/amf/2007170 2026

  25. [25]

    and Martin, T

    Bracho, R. and Martin, T. A. (2026). AmeriFlux FLUXNET-1F US-SP1 Slashpine-Austin Cary- 65yrs nat regen. doi:10.17190/AMF/2469470

    work page doi:10.17190/amf/2469470 2026

  26. [26]

    AlphaEarth Foundations: An embedding field model for accurate and efficient global mapping from sparse label data

    Brown, C. F., Kazmierski, M. R., Pasquarella, V . J., Rucklidge, W. J., Samsikova, M., Zhang, C., Shelhamer, E., Lahera, E., Wiles, O., Ilyushchenko, S., Gorelick, N., Zhang, L. L., Alj, S., Schechter, E., Askay, S., Guinan, O., Moore, R., Boukouvalas, A., and Kohli, P. (2025). AlphaEarth foundations: An embedding field model for accurate and efficient gl...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  27. [27]

    Brunsell, N. (2026a). AmeriFlux FLUXNET-1F US-KFS Kansas Field Station. doi:10.17190/AMF/1881585

    work page doi:10.17190/amf/1881585

  28. [28]

    Brunsell, N. (2026b). AmeriFlux FLUXNET-1F US-Kon Konza Prairie LTER (KNZ). doi:10.17190/AMF/2316062

    work page doi:10.17190/amf/2316062

  29. [29]

    Buchmann, N., Hortnagl, L., Feigenwinter, I., Wang, Y ., Zeeman, M., Fuchs, K., Baur, T., and Meier, P. (2024). Fluxnet Product from Chamau, 2004-12-31–2023-12-31. hdl:11676/u86sqUQo57KIYjIgZjpQBMvM

    work page 2024

  30. [30]

    Buysse, P., Loubet, B., Depuydt, J., Durand, B., and Gueudet, J. (2024). Fluxnet Product from Grignon, 2003-12-31–2023-12-31. hdl:11676/sRvNe2sOkU4VL4ipBWhOIUtL

    work page 2024

  31. [31]

    T., Namkoong, H., and Yadlowsky, S

    Cai, T. T., Namkoong, H., and Yadlowsky, S. (2026). Diagnosing model performance under distribution shift.Operations Research, 74(2):898–916

    work page 2026

  32. [32]

    Calvet, J., Canut-Rocafort, G., Etienne, J., Maurel, W., and Paci, A. (2024). Fluxnet Product from Toulouse, 2017-12-31–2023-12-31. hdl:11676/jVgAFrA7fY40mDmOFN_ypHUl

    work page 2024

  33. [33]

    D., Oikawa, P., Sturtevant, C., Szutu, D., Verfaillie, J., and Baldocchi, D

    Chamberlain, S. D., Oikawa, P., Sturtevant, C., Szutu, D., Verfaillie, J., and Baldocchi, D. (2026). AmeriFlux FLUXNET-1F US-Tw3 Twitchell Alfalfa. doi:10.17190/AMF/1881594

    work page doi:10.17190/amf/1881594 2026

  34. [34]

    and Knox, S

    Christen, A. and Knox, S. (2025). AmeriFlux FLUXNET-1F CA-DBB Delta Burns Bog. doi:10.17190/AMF/1881565

    work page doi:10.17190/amf/1881565 2025

  35. [35]

    Crill, P., Friborg, T., Jammet, M., Jansen, J., and Jansen, J. (2024). Fluxnet Product from Stordalen grassland, 2011-12-31–2019-12-31. hdl:11676/9r8BeQrUuOcdRyO3VV7k0Jxd

    work page 2024

  36. [36]

    K., Hotz, V

    Crump, R. K., Hotz, V . J., Imbens, G. W., and Mitnik, O. A. (2009). Dealing with limited overlap in estimation of average treatment effects.Biometrika, 96(1):187–199

    work page 2009

  37. [37]

    Czerný, R., Kowalska, N., and Pavelka, M. (2024). Fluxnet Product from Rajec, 2011-12-31–2022-12-31. hdl:11676/u_UX7v3F5AlRBVNYcBR-dUbl

    work page 2024

  38. [38]

    Czerný, R., Pavelka, M., and Šigut, L. (2024). Fluxnet Product from Stitna, 2009-12-31–2022-12-31. hdl:11676/KTb73PkwJI6As0hGJkhZH3ob. 11

    work page 2024

  39. [39]

    Desai, A. (2024a). AmeriFlux FLUXNET-1F US-Pnp Lake Mendota, Picnic Point Site. doi:10.17190/AMF/2229386

    work page doi:10.17190/amf/2229386

  40. [40]

    Desai, A. (2024b). AmeriFlux FLUXNET-1F US-Syv Sylvania Wilderness Area. doi:10.17190/AMF/2204879

    work page doi:10.17190/amf/2204879

  41. [41]

    Desai, A. (2026). AmeriFlux FLUXNET-1F US-CS2 Tri county school Pine Forest. doi:10.17190/AMF/1881577

    work page doi:10.17190/amf/1881577 2026

  42. [42]

    and Desai, A

    Duff, A. and Desai, A. (2026). AmeriFlux FLUXNET-1F US-DFC US Dairy Forage Research Center, Prairie du Sac. doi:10.17190/AMF/2006975

    work page doi:10.17190/amf/2006975 2026

  43. [43]

    Dusek, J., Kivalov, S., Pavelka, M., Šigut, L., Jocher, G., Czerný, R., Staník, K., Stellner, S., and Trusina, J. (2024). Fluxnet Product from Trebon, 2005-12-31–2023-12-31. hdl:11676/tyVjLxtNIu3yMPR4sGZMgAiB

    work page 2024

  44. [44]

    Ecosystem Thematic Centre, Acosta, M., Czerný, R., and Fischer, M. (2021a). Fluxnet Product from Kresin u Pacova, 2013-12-31–2020-12-31. hdl:11676/30hwGNHPHXM83-yXYam0yOeL

    work page 2013

  45. [45]

    Ecosystem Thematic Centre, Alekseychik, P., Kolari, P., Laakso, H., Levula, J., Li, X., Peltola, O., Taipale, R., and Tuittila, E. (2021b). Fluxnet Product from Siikaneva, 2015-12-31–2020-12-31. hdl:11676/J8Irl-jihmkPrA8WVpMuron2

    work page 2015

  46. [46]

    Ecosystem Thematic Centre, Arriga, N., De Meulder, T., and Roland, M. (2021c). Fluxnet Product from Maasmechelen, 2015-12-31–2020-12-31. hdl:11676/od36WNQog-z8pzPaDo0zrY8f

    work page 2015

  47. [47]

    Ecosystem Thematic Centre, Arriga, N., Goded, I., and Manca, G. (2022a). Fluxnet Product from San Rossore 2, 2012-12-31–2020-12-31. hdl:11676/E_mBXGeQxTWyuXxYl_uETTOO

    work page 2012

  48. [48]

    Ecosystem Thematic Centre, Belelli Marchesini, L., Gianelle, D., Marcolla, B., and Sottocornola, M. (2021d). Fluxnet Product from Monte Bondone, 2002-12-31–2020-12-31. hdl:11676/Clb3PaX0IgdqDvXBocguQW_A

    work page 2002

  49. [49]

    Ecosystem Thematic Centre, Belelli Marchesini, L., Gianelle, D., and Sottocornola, M. (2021e). Fluxnet Product from Lavarone, 2002-12-31–2020-12-31. hdl:11676/sJA2oTew4l0lyyrBW9U_j4Ro

    work page 2002

  50. [50]

    Ecosystem Thematic Centre, Brümmer, C., Delorme, J., Lucas-Moffat, A., and Schrader, F. (2021f). Fluxnet Product from Gebesee, 2000-12-31–2020-12-31. hdl:11676/aKmlHV66TYdH3AN9JY_q_3vX

    work page 2000

  51. [51]

    Ecosystem Thematic Centre, Brut, A., Ceschia, E., and Tallec, T. (2021g). Fluxnet Product from Lamasquere, 2004-12-31–2020-12-31. hdl:11676/38J08LOrwy_T-OoiO_nmtB00

    work page 2004

  52. [52]

    Ecosystem Thematic Centre, Carrara, A., Hertel, M., Kolle, O., Lopez Jimenez, R., Migliavacca, M., Perez-Priego, O., Reichstein, M., and El-Madany, T. (2021h). Fluxnet Product from Albuera, 2014-12-31–2020-12-31. hdl:11676/RiUT0H52x88isyYd6USXpKrO

    work page 2014

  53. [53]

    Ecosystem Thematic Centre, Carrara, A., Hertel, M., Kolle, O., Lopez Jimenez, R., Migliavacca, M., Perez-Priego, O., Reichstein, M., and El-Madany, T. (2021i). Fluxnet Product from Majadas del Tietar South, 2013-12-31–2020-12-31. hdl:11676/dDlpnhS3XKyZjB22MUzP_nAm

    work page 2013

  54. [54]

    Ecosystem Thematic Centre, Chipeaux, C., and Loustau, D. (2022b). Fluxnet Product from Bilos, 2013-12-31–2020-12-31. hdl:11676/k6zm9zfIMosYX217qXf6LMrg

    work page 2013

  55. [55]

    Ecosystem Thematic Centre, Chopin, H., De Ligne, A., Heinesch, B., Longdoz, B., Manise, T., and Thyrion, T. (2021j). Fluxnet Product from Lonzee, 2003-12-31–2020-12-31. hdl:11676/ql2ZkJ2Xx4a4yOyG3cd5lsBS

    work page 2003

  56. [56]

    Ecosystem Thematic Centre, Chopin, H., De Ligne, A., Heinesch, B., Longdoz, B., Manise, T., Thyrion, T., and Vincke, C. (2021k). Fluxnet Product from Vielsalm, 1995-12-31–2020-12-31. hdl:11676/OxWwBoGJ1YtWy2dro1LC8xZz

    work page 1995

  57. [57]

    Ecosystem Thematic Centre, Conte, A., Fares, S., Moretti, V ., and Sorgi, T. (2021l). Fluxnet Product from Castelporziano2, 2011-12-31–2020-12-31. hdl:11676/EFaLhqd0HuztDhggz3bqZdSZ

    work page 2011

  58. [58]

    Ecosystem Thematic Centre, Czerný, R., Havrankova, K., and Kowalska, N. (2021m). Fluxnet Product from Lanzhot, 2014-12-31–2020-12-31. hdl:11676/i1iVdqxE3UvLrKAaa52UXjXD

    work page 2014

  59. [59]

    Ecosystem Thematic Centre, Czerný, R., Kowalski, A., Moya Jimenez, M., Perez-Priego, O., Poveda, F., and Sanchez-Canete, E. (2021n). Fluxnet Product from Llano de los Juanes, 2003-12-31–2020-12-31. hdl:11676/Cr-JDShDipHoRjx4A1u2Ztg_. 12

    work page 2003

  60. [60]

    Ecosystem Thematic Centre, de la Motte, L., De Ligne, A., Heinesch, B., and Manise, T. (2021o). Fluxnet Product from Dorinne, 2010-12-31–2020-12-31. hdl:11676/-BXpIJo3yHnyWtpFJDuhspRC

    work page 2010

  61. [61]

    Ecosystem Thematic Centre, De Meulder, T., Janssens, I., Neirynck, J., Roland, M., and Segers, J. (2021p). Fluxnet Product from Brasschaat, 1995-12-31–2020-12-31. hdl:11676/KAwldGoiEpQUhNLhtpQiuunn

    work page 1995

  62. [62]

    Ecosystem Thematic Centre, De Simon, G., Nilsson, M., Peichl, M., and Pilegaard, K. (2021q). Fluxnet Product from Rosinedal-3, 2013-12-31–2020-12-31. hdl:11676/eB7V oOGljshGM3SRVDt-hSZH

    work page 2013

  63. [63]

    Ecosystem Thematic Centre, Di Tommasi, P., Famulari, D., Magliulo, V ., Tedeschi, A., and Vitale, L. (2021r). Fluxnet Product from Borgo Cioffi, 2003-12-31–2020-12-31. hdl:11676/HV3Vdu3wizlI1WDelOFo6tXL

    work page 2003

  64. [64]

    Ecosystem Thematic Centre, Dienstbach, L., Gimper, S., and Rebmann, C. (2021s). Fluxnet Product from Hohes Holz, 2014-12-31–2020-12-31. hdl:11676/R48rbzDkSXrn3xSbwvA-DN_u

    work page 2014

  65. [65]

    Ecosystem Thematic Centre, Guibal, F., Marloie, O., and Simioni, G. (2021t). Fluxnet Product from Font-Blanche, 2007-12-31–2020-12-31. hdl:11676/VpfIV-qkAaa2pK3pEOjpKVb6

    work page 2007

  66. [66]

    Ecosystem Thematic Centre, Heimsch, L., Kulmala, L., and Lohila, A. (2021u). Fluxnet Product from Qvidja, 2017-12-31–2020-12-31. hdl:11676/hLSa6ewsorj59DxvHG5VXIlY

    work page 2017

  67. [67]

    Ecosystem Thematic Centre, Heliasz, M., and Holst, J. (2021v). Fluxnet Product from Hyltemossa, 2014-12-31–2020-12-31. hdl:11676/C0pDM8bNR8HX3WifYkc_i96a

    work page 2014

  68. [68]

    Ecosystem Thematic Centre, Holst, J., Lagergren, F., Lankreijer, H., Lindroth, A., and Molder, M. (2021w). Fluxnet Product from Norunda, 2013-12-31–2020-12-31. hdl:11676/qwGaRkeqNlg1x7MUZlxseg2S

    work page 2013

  69. [69]

    Ecosystem Thematic Centre, Holst, J., Marklund, P., Nilsson, M., Ottosson-Löfvenius, M., and Peichl, M. (2021x). Fluxnet Product from Svartberget, 2013-12-31–2020-12-31. hdl:11676/Ln8lpBVmO5CezuX971idRffI

    work page 2013

  70. [70]

    Ecosystem Thematic Centre, Ibrom, A., and Pilegaard, K. (2022c). Fluxnet Product from Soroe, 1995-12-31–2020-12-31. hdl:11676/WTIy4tghcSXSceDBBm05uPAQ

    work page 1995

  71. [71]

    Ecosystem Thematic Centre, Kolari, P., Mammarella, I., and Vesala, T. (2021y). Fluxnet Product from Hyytiala, 1995-12-31–2020-12-31. hdl:11676/4F2-9d7QV9A0SlL2pIaRxsJP

    work page 1995

  72. [72]

    Ecosystem Thematic Centre and Kruijt, B. (2019). Fluxnet Product from Loobos, 1995-12-31–2018-12-31. hdl:11676/VXyr5llD0kKeovM4lV7D82Eu

    work page 2019

  73. [73]

    Ecosystem Thematic Centre, López Ballesteros, A., Moya Jimenez, M., Perez-Priego, O., Poveda, F., and Sanchez-Canete, E. (2021z). Fluxnet Product from Aguamarga, 2005-12-31–2020-12-31. hdl:11676/p7eV oJ0wLsL9apY2Kzd0kRhe

    work page 2005

  74. [74]

    Ecosystem Thematic Centre and Saunders, M. (2022). Fluxnet Product from Clara Raised Bog, 2020. hdl:11676/hMaaT1-nmS8xQEDZmf_BuOxc

    work page 2022

  75. [75]

    Ecosystem Thematic Centre and Weslien, P. (2020). Fluxnet Product from Lanna, 2013-12-31–2018-12-31. hdl:11676/OPqCBWpZiv5GNg0zTW6v0IBn

    work page 2020

  76. [76]

    and Stecher, M

    Ecosystem Thematic Centre, Montagnani, L. and Stecher, M. (2021). Fluxnet Product from Renon, 1998-12-31–2020-12-31. hdl:11676/eQhkwmXzwyLatT8qqktzYxzh

    work page 2021

  77. [77]

    Ecosystem Thematic Centre, Muller, J., Rotenberg, E., Schwartz, E., and Yakir, D. (2021). Fluxnet Product from Yatir, 1999-12-31–2020-12-31. hdl:11676/ahX7qSkxLQu0GF5MRtdv-f4Y

    work page 2021

  78. [78]

    Eichelmann, E., Shortt, R., Knox, S., Rey Sanchez, C., Valach, A., Sturtevant, C., Szutu, D., Verfaillie, J., and Baldocchi, D. (2026). AmeriFlux FLUXNET-1F US-Tw4 Twitchell East End Wetland. doi:10.17190/AMF/2204881

    work page doi:10.17190/amf/2204881 2026

  79. [79]

    Euskirchen, E. (2025a). AmeriFlux FLUXNET-1F US-BZF Bonanza Creek Rich Fen. doi:10.17190/AMF/1881570

    work page doi:10.17190/amf/1881570

  80. [80]

    Euskirchen, E. (2025b). AmeriFlux FLUXNET-1F US-BZo Bonanza Creek Old Thermokarst Bog. doi:10.17190/AMF/1881571

    work page doi:10.17190/amf/1881571

Showing first 80 references.