the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Sep 2026
Noah-MP-ML² v1.0: Development of a multilayer canopy land surface model with improved representation of vadose zone soil water dynamics
Abstract. Land surface models (LSMs) simulate surface processes and land–atmosphere exchanges of energy, water, and carbon. However, many LSMs still rely on simplified parameterizations of surface-layer radiative transfer, canopy turbulence, and vadose-zone hydrology. This paper describes Noah-MP-ML², a multilayer canopy extension of the community Noah-MP LSM developed to address these limitations. Noah-MP-ML² implements a roughness-sublayer scheme and the Medlyn stomatal-conductance parameterization to represent canopy turbulence and photosynthetic processes. The implementation of the Speedy Algorithm for Radiative Transfer through Cloud Sides (SPARTACUS)-Surface scheme enables consistent N-stream shortwave and longwave radiative transfer across multiple spectral bands while accounting for lateral radiative exchange within heterogeneous vegetation canopies. A multilayer implicit solution is used to calculate canopy scalar concentrations and the temperatures of sunlit and shaded foliage. Soil-water movement is represented using a mass-conserving, mixed-form, one-dimensional Richards-equation solver with van Genuchten water-retention relationships. Noah-MP-ML² was evaluated against observations from five forest flux-tower sites spanning different vegetation types and climatic regimes. The results demonstrate improved representation of surface energy partitioning, turbulence through friction velocity, and soil-water dynamics relative to the community Noah-MP LSM.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Geoscientific Model Development.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(9237 KB) - Metadata XML
-
Supplement
(6471 KB) - BibTeX
- EndNote
Status: open (until 19 Nov 2026)
- RC1: 'Comment on egusphere-2026-4875', Anonymous Referee #1, 20 Oct 2026 reply
Model code and software
Noah-MP-ML²: Multilayer canopy land surface model with improved representation of vadose zone soil water dynamics Harsh G. Kamath, Gordon B. Bonan, Zong-Liang Yang, Ronnie Abolafia-Rosenzweig, Cenlin He, Ashley M. Matheny, and Dev Niyogi https://doi.org/10.5281/zenodo.21908465
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 561 | 228 | 62 | 851 | 72 | 66 | 59 |
- HTML: 561
- PDF: 228
- XML: 62
- Total: 851
- Supplement: 72
- BibTeX: 66
- EndNote: 59
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents a substantial extension of Noah-MP, bringing together multilayer canopy exchange, SPARTACUS radiative transfer, and a more detailed representation of soil-water movement. The manuscript is well written and the development is strong controbution for the modelling community. My concerns are that the experimental design does not adequately separate the contributions of the individual developments from differences in prescribed parameters, implementation details, and calibration. Consequently, several claims of improvement are broader than the evidence supports.
Specific comments:
1. The abstract states that the model improves surface-layer energy partitioning (p. 1, lines 11–13). However, Table 3 shows that single-layer Noah-MP with Medlyn has higher latent heat flux Taylor scores than ML²-20 at all five sites. At the same time, single-layer Medlyn has higher sensible heat flux skill at four of the five sites. The difference at DE-HAI is particularly substantial: 0.86 for single-layer Medlyn compared with 0.63 for ML²-20. The discussion acknowledges that single-layer Medlyn mostly outperformed ML²-20 (p. 31, lines 590–592). ML²-20 does improve heat-flux skill relative to the original community configuration at some sites. The distinction is therefore between improvement over that particular baseline and an additional benefit from the multilayer formulation. The abstract and conclusions should make this distinction explicit. Friction velocity provides the clearest evidence of improvement, with substantial gains at four sites and essentially unchanged performance at US-MMS relative to community Noah-MP.
2. Another issue I am concerned is that the experiments do not isolate the contributions of canopy layering and the other model changes. The single-layer Ball–Berry/Medlyn and MOST/RSL comparisons in Table 2 are useful. However, the transition to the multilayer model simultaneously changes radiative transfer, canopy discretization, and important details of physiology and turbulence. The transition from ML²-4 to ML²-20 additionally changes soil discretization, hydraulic relationships, infiltration, root distribution, and the weighting of soil-moisture stress. I would like to see a comparison using the same Medlyn formulation, convergence criteria, humidity treatment, and prescribed physiological parameters in both canopy configurations. Additional targeted experiments, such as ML² with Ball–Berry and ML² retaining the original radiation treatment, would help identify which developments explain the changes in performance. Similarly, the hydrological comparison should distinguish the effects of vertical resolution from those of hydraulic parameters, infiltration, and root-water-uptake assumptions.
3. Please clarify whether all experiments use the site-specific canopy heights and LAI values listed in Table 1. Section 5 treats canopy height, LAI, and leaf geometry as prescribed static attributes (p. 22, lines 436–440), whereas Appendix B states that canopy and surface properties for the single-layer experiments were obtained from the Noah-MP lookup table (p. 34, lines 658–665). Table 1 gives canopy heights of 21, 34, 18, 27, and 41 m, but the manuscript does not clearly identify the actual heights used in each single-layer experiment or any overrides to the default tables. If the multilayer model uses observed site heights while the baseline uses generic vegetation-class values, part of the improvement in friction velocity could arise from better specified canopy geometry. Please report the actual canopy height, forcing height, LAI, displacement height, and roughness parameters used in every configuration, explaining which quantities are prescribed and which are calculated. A comparison with matched site geometry would make the interpretation of the u* improvement much stronger. Matching canopy height alone may be insufficient if other prescribed aerodynamic properties still differ. The temporal treatment of LAI also needs clarification, particularly for DE-HAI during September–October. If LAI remains constant, please explain the basis for that assumption and assess its influence where seasonal canopy changes may matter. The archived configuration should clearly identify how these site attributes are supplied and any run-specific source or table modifications.
4. Last but not least, the numerical robustness and model computational cost need quantitative assessment. In section 4.3, authors report difficulty recovering from dry conditions with a damping factor of 0.8, followed by adaptive reduction to 0.4 and an approximately 0.005 m³ m⁻³ separation between residual and wilting water contents (p. 20, lines 406–411). These choices need explicit activation criteria, sensitivity tests, and a distinction between numerical safeguards and changes to physically meaningful hydraulic parameters. More importantly, authors report increased cumulative water-balance error when adaptive damping is activated, without providing its magnitude (lines 412–414). Given the description of the solver as mass-conserving, please report cumulative and maximum conservation errors, convergence tolerances, and failure frequencies for the site simulations and demanding dry-to-wet tests. Please also explain what happens when convergence cannot be achieved at the minimum permitted time step.