When the panel speaks: reading the STRADIVARI evaluation

 A follow-up to "A double failure is a giant failure"

The ERC Step-1 evaluation for STRADIVARI has arrived. I promised in November that when it came I would read it carefully and share what I learned. The verdict: score B, ranking range 75–84%, where the top 37% advanced. So the project sits in the lower-middle of PE10 ADG 2025 — not borderline, not retained.

Let me work through what the four reviewers and the panel actually said, because some of it is fair, some of it is unfair, and one piece of it deserves a longer comment about how reviewing is changing in the age of AI suspicion.

The convergent verdict: "incremental"

The panel summary, and three of the four reviewers, converge on a single word: incremental. Reviewer 1 puts it most plainly — the system "falls more or less in the same category" as existing land-surface schemes and is "essentially built by assembling existing components." The panel adopted this reading almost verbatim: "Coupled atmosphere–land models of the suggested type already exist, and the panel was not convinced how the project would go beyond these existing approaches."

I should take this seriously, and I do. But I also want to say honestly what I think happened.

STRADIVARI was framed around what I called a "luthier" approach: building instruments that enable virtuoso scientific work, rather than promising the breakthrough finding myself. The proposal says it explicitly in Part B2: "The project provides the instruments; the community provides the scientific virtuosity."

This is what I believe. It is also, I now see, almost designed to be scored "incremental" by an ERC ADG panel. ADG rewards a PI who claims they will deliver the breakthrough. By framing my project as infrastructure that enables others to make breakthroughs, I handed the panel the exact line they used. They read my honest description of the project's character back to me as a weakness.

I do not regret the framing. I regret choosing the wrong instrument for it. The "luthier" vision is real, the GEOframe/WHETGEO/GEOSPACE work is real, and the integration questions are real. But ADG is not where that proposal belongs. That is a strategic lesson, not a scientific one.

What was actually in the proposal that nobody saw

The proposal contained several specific theoretical claims that, taken individually, are not incremental. The resilience-vs-optimization framing of plant hydraulics (D'Amato & Rigon 2025), the dynamic-SWRC-under-biota proposal, the deconstruction of the resistance framework so that ABL turbulence is resolved rather than parameterized and conductance is left as a purely physiological quantity — none of these are "assembling existing components." They are theoretical positions. But they appear in the proposal as sub-bullets inside a framework-integration narrative, and Reviewer 1, scanning for novelty, never reached them. Reviewer 3, who engaged scientifically, did not flag them either — which means they were not loud enough.

This is on me. I knew this, and I made the choice to lead with the integrating framework. It was the wrong choice for this audience. Everybody overlooked the citations of Diego Miralles et al. (2025): "Current ESS Models fail to capture critical feedbacks between soil evolution, plant hydraulics, and atmospheric processes required for understanding coupled hydrological and ecosystem functioning because Earth's system compartments are often treated as silos or heavily parameterized in crucial aspects of their dynamics."

The four-pillar problem

R2, R3, and R4 each separately worry about how the four pillars (Soil/Biota, Plant Hydraulics, Carbon, ABL) actually integrate. R2 says the proposal "does not convincingly demonstrate how the different parts will be combined." R4 says the methodology is "overly conceptual, with limited discussion of specific validation metrics, error propagation, or computational performance."

This is a real critique. The three-tier validation and minimum-viable-deliverables structure in B2 is there, but it arrives late and it does not pin the whole project to a single experiment that, if it works, validates the integration. ADG proposals win when there is a falsifiable claim at the centre. STRADIVARI has many — too many to land as one.

The Giono problem, and credit to existing literature

R3 makes a small but accurate point: I opened with Giono's L'homme qui plantait des arbres and never paid the framing off. The tree-planter never returns to the proposal. R3 also flags limited engagement with the existing forest-hydrology literature, which is fair: I cited what I needed for the argument, not what I owed to the field. Both points are correctable in any future writing.

And then: the Beven attribution

This is the part that has stayed with me longest. R3 writes:

"I also have to note one surprising aspect: on page 3 (bottom), the formulation 'parameters as garbage collectors' is used with reference to a paper by Keith Beven. Interestingly, this term can not be found in that publication (or any other publication of Beven), but appears as a (made-up) quote when checking with ChatGPT."

Let me be direct. The "garbage collectors" formulation was a shortcut of mine — a paraphrase that condensed something I take to be implicit in Beven (2006) and the equifinality literature, attributed informally with a (sensu Beven, 2006). It was not a quotation, and I did not present it as one. It was a writing shortcut, and a sloppy one, because attaching anyone's name to a paraphrase in an ADG proposal asks for a higher bar of fidelity than that.

I take the correction. Beven did not write those words, I should not have invoked him for them, and a future version of this argument will either find the exact passage or carry the claim in my own voice without his name attached.

But I want to say something about the second half of R3's remark, because it is no longer about my proposal — it is about how reviewing is changing.

R3 verified the phrase by querying ChatGPT and found it appears there as a "made-up quote." That is offered, in the review, as evidence of fabrication. It is presented in a tone that suggests the proposal itself may have been generated, or at least drafted unchecked, by an AI. This is not stated, but the implication is in the air, and the panel has now seen it.

Here is what I want to register. The search for hallucinations is beginning to exceed the people doing it. A reviewer who finds a phrase that does not appear in the cited source has, in 2025, two possible explanations: the author paraphrased badly, or an AI made it up. The second hypothesis is now reached so quickly that it crowds out the first. In my case the first was correct. I made the shortcut myself — the dumb old-fashioned way, sitting at a keyboard, trying to compress two paragraphs of Beven into seven words. That this looks identical, from the outside, to an AI hallucination is not a reason to stop writing carefully. It is a reason that every attribution must now survive the test of being searchable, verifiable, and grounded — because the cost of an unverifiable phrase is no longer "you paraphrased loosely" but "you may not have read the source." That is a real shift in what scientific writing has to do, and we should name it.

A second point worth making: a proposal that also foregrounds AI integration (the SML / OMS4 component) creates extra surface for this suspicion. The proposal is, in some sense, a candidate hypothesis for its own AI-suspicion test. I had not seen that read coming, and I should have.

What the panel said about me, the PI

All four reviewers were generous about the track record. R4 was particularly clear: "highly accomplished and internationally recognised," with the GEOtop/GEOframe/OMS lineage giving "strong credibility." R3 mentioned the AboutHydrology mailing list (almost 7000 subscribers, for those keeping count). R1 and R2 both said I had the necessary qualifications.

I mention this not to console myself but because it locates the problem precisely. The PI was not the problem. The project, as written, was. That is the more useful diagnosis, even if it is the harder one.

What I take from this

A few things, in plain form:

The "luthier" framing is honest, but ADG is the wrong instrument for it. Either Synergy, or a single-PI proposal built around one falsifiable theoretical claim with the infrastructure as means, would fit better.

The deepest scientific moves in STRADIVARI need to be the centre of any future proposal, not the load-bearing sub-bullets of an integration narrative.

The Beven shortcut is fixed for the future. Every named attribution gets verified or removed. Not because reviewers might think it was an AI, but because they are now correct to check, and the burden of verification has shifted onto authors in a way that did not exist five years ago.

The questions STRADIVARI raised — about dynamic feedbacks the current LSM/ESM generation parameterizes away, about whether minimum-energy-dissipation principles extend to the coupled vegetation-soil-atmosphere system, about whether plant strategies are optimal or resilient — are unresolved. The panel did not say otherwise. They said I had not convinced them I would resolve them inside 60 months with this team and this instrument. That is a different claim, and it does not retire the questions.

The proposals stay on OSF for anyone who wants to read what worked, what did not, and what the panel said about it. As the November post put it, a double failure is a giant failure. But it is the kind of failure that produces a written record, and a written record is more useful than a private wound.

Onward.

Musical Coda

The Statistical physics of unsaturated soil water: kinetic theory and non commutative pore water dynamics

I am giving this talk at the EGU General Assembly 2026 in Vienna last week, in the Hydrological Sciences division. The argument, in a single sentence: Richards' equation is not wrong, but it is the equilibrium limit of a deeper kinetic theory — in the same sense that the Navier–Stokes equations are the hydrodynamic limit of the Boltzmann equation for a gas. Mario Putti twenty years ago once asked me, "if not Richards, what else?"; this is my attempt at an answer that arrives after year dedicated to properly solve Richards equation, before with GEOtop and later with WHETGEO. 

The core object is a filling distribution g(r, x, t) : ℝ⁺ → [0, 1] that gives the volume fraction of pores of radius r that are water-filled at position x and time t. Theta is recovered as θ[g] = φ ∫₀^∞ g(r) f(r) dr. Hysteresis becomes the non-commutativity [W, D] ≠ 0 of the wetting and drying operators — geometry, not memory. Richards' equation is recovered as the small Damköhler limit Da → 0, with K(ψ) emerging as a derived transport coefficient built from the connectivity kernel C(r, r') rather than being postulated.

Materials

• Slides (PDF)  the deck I'll use in the presentation.
• Storyboard (DOCX)  the slide-by-slide reading guide, in five columns: spoken text, visual content, speaker notes, mounting comments. Useful if you want to present the same material yourself, or if you just want to follow along with what I actually said.
• Extended version of the slides — give me a few days — an annotated version with the full speaker text, more references, and the bits I had to cut for time.

Notebooks

These are the Jupyter notebooks I used to generate some of the figures in the slides, plus a few that produce supporting evidence in the supplementary material of the upcoming PRE papers. All run on top of OpenPNM 3.x and a small custom Y–L percolation code.

• Hysteresis_SWRC.ipynb — drainage and wetting branches in the (ψ, S_e) plane on a 3D pore network, with internal scanning curves. The figure on slide 9 of the talk comes from here. The notebook also documents an algorithmic artifact near the air-entry value (the missing air-trapping term during imbibition) — which is honest enough that I left it in.
• OpenPNM_Da_overshoot.ipynb — non-equilibrium overshoot in (θ, ⟨r⟩) and the universality crossover when the pore-size distribution becomes bimodal, governed by the Bhattacharyya overlap of the two modes.
• Percolation_K_threshold.ipynb — the percolation scaling K ∝ (θ − θ_c)^t with t ≈ 2, with finite-size scaling on three lattice sizes.
• subsection_pnm_mapping.tex — a short LaTeX subsection on how a two-tier pore-network maps onto the kinetic theory through a bimodal f(r) and a block-structured C(r, r'). Background reading for the OpenPNM notebooks.

Please find them zipped at this link.

Two upcoming papers

The full theoretical development is in two manuscripts, going to arXiv soon and submitted thereafter to Physical Review E --- give me a couple of weeks after EGU26:

1. The Statistical Physics of Unsaturated Soil Water: kinetic theory and non-commutative pore-water dynamics — the long paper. Builds the kinetic equation from the network thermodynamics, identifies the Onsager–Rayleigh gradient-flow structure, and proves that hysteresis is a geometric property of the configuration bundle (not a memory effect).
2. Richards' equation as a hydrodynamic limit: Chapman–Enskog derivation from the kinetic equation for unsaturated soil water — the short companion. Walks through the Chapman–Enskog expansion that recovers Richards' equation in the Da → 0 limit, with K(ψ) derived from the connectivity kernel.

Where this connects

The framework absorbs and extends a number of existing approaches that have been circling the same physics from different angles:

• Mixed-form Richards as the Da → 0 limit, with K(ψ) derived rather than postulated.
• Hassanizadeh–Gray as a thermodynamically consistent extension — pore-class-resolved here.
• Phase-field methods (Cahn–Hilliard) as gradient flow on a free energy — with explicit pore-network connectivity through C(r, r').
• Lucas–Washburn and its fractal variants as the single-capillary kinetic building block of C(r, r').
• Percolation-based hillslope frameworks with Damköhler and Péclet, where macropore activation is the Da > 1 transition.
• Compressible statistical soil mechanics (Einav–Liu 2023) — same occupancy dynamics governs the (ψ, σ') coupling.
• Freezing soil thermodynamics (Rempel et al. 2023, and our own work with Wani and D'Amato) — same kinetic framework with capillary pressure replaced by freezing-point depression.

This kinetic theory is not a parallel universe to Richards. It absorbs the existing physics, and it opens new measurements — directly observing g(r) is the obvious next experimental challenge

Stomata close to maximize transpiration ?

This is the talk I am going  on EGU 2026 and I co-authored with Concetta D'Amato.  It talks about the complexity behind the plant reactions to various environmental factors and the interactions that control stomata openings. 

The slides of the talk can be found here.  The various figures were created within Jupyter Notebooks that are here with the helps of Claude. Please also consider to watch my other presentation on a new statistical theory on the dynamics of soil water in vadose zone, also presented at EGU 2026. This latter presentation is here. 

EGU WIEN 2026

Please you can find what we present and do at the EGU 2026 General Assembly in Wien. 

The high resolution pdf is here. 

Below one by one the contributions:

• Marianna Tavonatti & Al. (On Tuesday) Pico Presentation

GEOSPACE Validation Paper: Application and Testing in the "Spike II" Lysimeter Experiment

We have just submitted a new paper — A flexible open-source modular framework for ecohydrological modeling: Application and validation of GEOSPACE-1D — by Concetta D'Amato, Paolo Benettin, Andrea Rinaldo, and Riccardo Rigon. It is the natural companion and follow-up to the GEOSPACE framework paper published in Geoscientific Model Development earlier this year, which described the design principles and modular architecture of GEOSPACE. That paper introduced the framework; this one puts it to work.

The validation is built around the "Spike II" experiment, a carefully instrumented lysimeter study carried out in 2018 on the EPFL campus in Lausanne. Four soil columns — a willow tree (L2), two grass-covered lysimeters (L1 and L4), and a bare-soil system (L3) — were monitored over two months, providing high-quality weight-based evapotranspiration estimates alongside measurements of soil water content, pressure, and drainage. These data allow a thorough assessment of the model across contrasting vegetation types and soil configurations.

The core of the paper addresses three questions: Can GEOSPACE reproduce observed ecohydrological dynamics across such diverse conditions? What are the practical advantages of its modular structure? And does it enable novel analyses — the kind that open new scientific doors rather than merely close validation loops?

On performance: GEOSPACE reproduces soil water pressure dynamics, depth-resolved water content, bottom drainage, and evapotranspiration fluxes across all four lysimeters with R² values of 0.87, 0.81, 0.83, and 0.73 for L2, L1, L4, and L3 respectively. Mean residual biases are negligible throughout. The slightly lower performance over bare soil reflects a known structural limitation of the Penman–Monteith formulation for soil evaporation under conditions where thermal inertia matters — an honest diagnosis rather than a defect to be papered over.

On modularity: the willow lysimeter was simulated with three alternative evapotranspiration formulations — GEOET-Prospero-PM, GEOET-Priestley-Taylor, and GEOET-Penman-Monteith FAO — keeping the soil component (WHETGEO) and the partitioning solver (BrokerGEO) identical across all three runs. All formulations close the cumulative water balance (~600 mm over the experiment), but the Prospero-PM formulation captures sub-daily peak dynamics with roughly half the residual spread of the other two. The calibrated Priestley-Taylor α = 4.16 and FAO Kc = 3.9 — both well above standard values — are informative precisely because they expose structural limitations of simplified formulations when applied to a high-transpiration system dominated by stomatal control.

On novel capabilities: the model computes root water uptake (RWU) for every control volume at every time step, yielding a full depth–time distribution of uptake intensity. The willow shifts its water sourcing dynamically in response to moisture depletion and atmospheric demand, with the mean uptake depth varying over time in a way that closely mirrors the measured root density profile. This kind of depth-resolved diagnostic is directly relevant to isotope-based ecohydrology, where xylem water provides only a bulk integrated signal — GEOSPACE's spatial resolution of the RWU can help interpret what that bulk signal actually means.

The paper grew out of Concetta D'Amato's PhD work at the Center Agriculture Food Environment (C3A) at the University of Trento, supported by the WATZON COST Action and the PRIN 2017 WATZON project. Readers interested in the longer history of GEOSPACE and its components can find much of the background documented here on AboutHydrology: see the posts on GEOSPACE and WHETGEO, and in particular the earlier post on Concetta's PhD thesis and the exploration of the SPAC.

The source code is on GitHub at https://github.com/geoframecomponents/GEOSPACE-1D, with a frozen version on Zenodo. All simulation data are openly available. GEOSPACE continues to grow.

Waiting for the official preprint, you can download it here. Here instead, find the supplemental material. 

Three Decades of Snow Water Equivalent Dynamics in the Po River Basin, Italy: Trends and Implications

Seasonal snowpack is a key component of the mountain cryosphere, acting as a vital natural reservoir that regulates runoff downstream in snowfed basins.

 In mid- and low-elevation mountain regions such as the European Alps, snow processes, such as accumulation and ablation, are highly sensitive to climate change, having direct implications for hydrological forecasting and water availability. In this study, we present the analysis of a 30-year (1991–2021) long dataset of snow water equivalent (SWE) in the Po River District, Italy, which includes parts of the Alps and Apennines. The data is available at a 500 × 500 m2 spatial resolution and at a daily temporal scale (Dall’Amico et al., 2025). This data was generated using the “J-Snow” modeling framework, which integrates the physically based GEOtop model with in situ snow height observations and earth observation snow cover products such as MODIS. Our results show that the long-term (30 year) basin-wide mean annual SWE volume equals 3.34 Gm3. The elevation-wise statistical analysis of key snow volume and duration metrics shows that the most pronounced snow water equivalent losses occur below 2000 m a.s.l. Below this threshold, both snow volume metrics and duration metrics show a significant decrease, indicating decrease in snow water storage and earlier melt. Above this elevation, the snow volume metrics show increasing trend while as the duration metrics continue to show a shortened (decreasing trend) snow season except at the highest elevations (> 2500 m). The findings of this study highlight the changes to the mountain seasonal snow storage and the timing of snow disappearance across the Italian Alps. This combined effect highlights a fundamental shift in the hydrological regime of the Po River Basin, with significant implications for water availability and management under ongoing climate change. The data used in this paper are those freely available in Dall'Amico et al., 2025. 

References

Dall’Amico, Matteo, Stefano Tasin, Federico Di Paolo, Marco Brian, Paolo Leoni, Francesco Tornatore, Giuseppe Formetta, John Mohd Wani, Riccardo Rigon, and Gaia Roati. 2025. “30-Years (1991-2021) Snow Water Equivalent Dataset in the Po River District, Italy.” Scientific Data 12 (1): 374. https://doi.org/10.1038/s41597-025-04633-5.

Wani, John Mohd, Kelly E. Gleason, Matteo Dall’Amico, Federico Di Paolo, Stefano Tasin, Gaia Roati, Marco Brian, Francesco Tornatore, and Riccardo Rigon. 2025. “Three Decades of Snow Water Equivalent Dynamics in the Po River Basin, Italy: Trends and Implications.” EGUsphere. https://doi.org/10.5194/egusphere-2025-5520.

From Snow Depth to Streamflow: Reducing Snowfall Uncertainty in Alpine Headwaters with Sentinel-1 based snow depth retrievals

In mountainous regions, the sparse distribution of precipitation gauges at high elevations is a major source of uncertainty in snowfall estimation. This matters beyond the local scale: uncertainties originating in headwater areas propagate through hydrological modelling, affecting the estimation of all water balance components downstream. Yet establishing dense gauge networks in complex mountain terrain remains logistically and economically challenging — which makes it worthwhile to ask whether remote sensing can fill the gap.

This study assimilates Sentinel-1 C-band snow depth observations into the snow module of the GEOframe hydrological model, coupled with a snow-density scheme, to jointly update snow depth, snow water equivalent (SWE), and snowfall estimates. The method is applied to two key Alpine catchments: the Aosta River catchment and the headwaters of the Piemonte catchment in the upper Po River basin. Both are critical contributors of snowmelt-driven discharge to the Po Valley — sustaining its agricultural water supply — and both suffer from limited high-elevation gauge coverage.

Results show that assimilating satellite-derived snow depth systematically increases snowfall estimates across elevation gradients relative to the model's partitioned snowfall, and substantially improves simulated river discharge during the snowmelt season. Notably, similar improvements persist in years without active data assimilation, suggesting that the approach has a lasting positive influence on model state and performance.

This work¹ has been submitted to The Cryosphere and is currently under review.

References

Azimi, S., Girotto, M., Rigon, R., Roati, G., Barbetta, S., and Massari, C.: From Snow Depth to Streamflow: Reducing Snowfall Uncertainty in Alpine Headwaters with Sentinel-1 Based Snow Depth Retrievals, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2026-793, 2026.

Girotto, Manuela, Giuseppe Formetta, Shima Azimi, Claire Bachand, Marianne Cowherd, Gabrielle De Lannoy, Hans Lievens, et al. 2024. “Identifying Snowfall Elevation Patterns by Assimilating Satellite-Based Snow Depth Retrievals.” The Science of the Total Environment 906 (167312): 167312. https://doi.org/10.1016/j.scitotenv.2023.167312.