Alps Are Losing Snow

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.

This study provides the first comprehensive long-term (1991-2021) analysis of snow water equivalent changes in the Po River District, Italy, one of Europe’s second most climate sensitive regions. Our findings show stark elevation-dependent changes in snow water storage and duration with profound and immediate implications for water security and climate adaptation.

Using a high-resolution (500m, daily) dataset from 1991-2021 (Dall'Amico et al., 2025), we observed two primary findings: First, we observed a profound loss of snow volume and decrease in duration below 2000 meters, with some low-elevation bands losing over 30% of their total snow-water storage. In contrast, high-elevation zones (>2500 m) are experiencing increased accumulation (needs, but a continued shortened snow season. However, the increase in snow water storage at high elevations requires careful interpretation due to methodological constraints and systematic overestimation of high elevation precipitation by ERA5. Second, we show that this shorter snow season is not just an artifact of earlier spring melt, but is primarily driven by a delayed onset of snow accumulation in early winter. 

These elevation-dependent changes and loss of the seasonal snowpack highlight a fundamental shift in the hydrological regime of the Po River Basin, with significant implications for the timing and volume of runoff and the future availability of water in the region.Therefore, the Po River Basin is moving from a stable to a more volatile system.

Note: To know more, click on the picture (study area) above to read our preprint.

References:

Dall’Amico, M., Tasin, S., Di Paolo, F. et al. 30-years (1991-2021) Snow Water Equivalent Dataset in the Po River District, Italy. Sci Data 12, 374 (2025). https://doi.org/10.1038/s41597-025-04633-5

Roots2025 - A presentation of the GEOSPACE system

This is the presentation I am giving at the Roots2025 event . It talks about the GEOSPACE infrastructured to study the soil-plant-atmosphere interactions.  Being a very compressed presentation I cannot go to all the details which are better grasped by reading the references below or browsing the various contributions that can be found in this blog under the keyword  GEOSPACE. 

GEOSPACE infrastructure is very modular and its peculiarity is that it is based on "components" that are joined with a scripting language just before being executed. The system managing such components is OMS3.  GEOSPACE integrated two big subprojects, WHETGEO the subsystem that deals with soil and infiltration and GEOET the system that contains various solutions for estimating evaporation and transpiration. To get the slides, clik on the figure above. To get the code look at the Gitub repository. Video lectures (here) or to a certain extent here. To get more information read the following references.

References

D’Amato, Concetta. n.d. “Exploring the Soil-Plant-Atmosphere Continuum: Advancements, Integrated Modeling and Ecohydrological Insights.” Ph.D., Università di Trento.

D’Amato, Concetta, and Riccardo Rigon. 2025. “Elementary Mathematics Helps to Shed Light on the Transpiration Budget under Water Stress.” Ecohydrology: Ecosystems, Land and Water Process Interactions, Ecohydrogeomorphology 18 (2). https://doi.org/10.1002/eco.70009.

D’Amato, Concetta, Niccolò Tubini, and Riccardo Rigon. 2025. “A Component-Based Modular Treatment of the Soil–Plant–Atmosphere Continuum: The GEOSPACE Framework (v.1.2.9).” Geoscientific Model Development 18 (20): 7321–55. https://doi.org/10.5194/gmd-18-7321-2025.

Tubini, Niccolò, and Riccardo Rigon. 2022. “Implementing the Water, HEat and Transport Model in GEOframe (WHETGEO-1D v.1.0): Algorithms, Informatics, Design Patterns, Open Science Features, and 1D Deployment.” Geoscientific Model Development 15 (1): 75–104. https://doi.org/10.5194/gmd-15-75-2022.

Bonus Reference (unpublished so far)

Tubini, N., and R. Rigon. n.d. “WHETGEO-2D: A Framework to Solve 2D Partial Differential Equation Domain within GEOframe System. The Richardson-Richards Equation.” http://abouthydrology.blogspot.com/2022/08/whetgeo-2d-open-source-tool-fo-solving.html.

Research in Arctic Permafrost. A new start

In the rapidly changing Arctic, understanding permafrost behavior is critical for infrastructure, ecosystems, and climate science. Marianna Tavonatti's master's thesis at the University of Trento has delivered some achievements that points towards the advance our understanding of Canadian Arctic permafrost dynamics and provide essential tools for climate adaptation.

Permafrost—permanently frozen ground—covers 24% of the Northern Hemisphere and is rapidly thawing due to climate change. This thesis focused on the Canadian Arctic near the Inuvik-Tuktoyaktuk Highway, using advanced computer modeling to understand how permafrost responds to changing temperatures over time scales from decades to seasons. 

Permafrost - https://www.maggiebaylor.com/permafrost

Marianna Thesis covers: 

1. GEOtop Model Implementation: First comprehensive application of the GEOtop model for Canadian Arctic permafrost, successfully validated against real ground temperature data.

2. Historical Analysis: 73-year simulation (1950-2023) revealing clear evidence of accelerating permafrost warming and active layer deepening.

3. Future Projections: Advanced climate scenarios showing significant future changes in permafrost stability with direct implications for infrastructure planning.

4. Advanced Theoretical Framework: Enhanced understanding of frozen soil physics, improving how we model phase changes in complex soil systems.

5. Methodological Innovation: Created an integrated GlobSim-GEOtop modeling chain applicable to Arctic regions worldwide.

6. Practical Applications: Provided quantitative data essential for Arctic infrastructure design and climate adaptation strategies.

7. Scientific Contributions: Advanced climate change understanding with implications for carbon cycling and global climate feedbacks.

However, the best things is to read the thesis that you can get by clicking on the figure. 

Marianna's work establishes a foundation for future advancement in Arctic permafrost science. The research identifies specific opportunities for enhanced spatial modeling, improved climate projections, and expanded ecosystem coupling—providing a clear roadmap for continued innovation in this critical field.

As the Arctic continues to experience rapid environmental change, research like Marianna's becomes increasingly vital for understanding system responses and supporting sustainable development in one of Earth's most climate-sensitive regions. 

About erosion and hillslope evolution: a lost master thesis rediscovered

 This 2003/2004 Master thesis by Martina Brotto deserves some more visibility. It presents an exploration into how slopes evolve over time through erosion processes and tackles one of the fundamental challenges in geomorphology: understanding and predicting how hillsides change their shape through the complex interplay of soil production and erosion.

What makes this research particularly interesting is its departure from traditional landscape evolution models. While most existing models assume that erosion is limited only by the transport capacity of flowing water or wind (what scientists call "transport-limited" processes), Brotto introduces a more realistic approach. Her model considers that erosion can also be limited by how quickly rock breaks down into soil and how much material is actually available to be moved ("detachment-limited" processes). This distinction might seem technical, but it's crucial for understanding real-world erosion, especially in areas where bedrock is close to the surface or where soil production rates are slow.

The heart of the research lies in developing two complementary mathematical models. The first focuses on diffusive erosion processes – the slow, continuous movement of soil particles down slopes through countless small disturbances like frost action, animal activity, and the impact of raindrops. Using sophisticated numerical techniques including the conjugate gradient method, Brotto shows how these processes gradually smooth out irregularities in the landscape, creating the gentle, rounded hills we often see in nature. Her simulations, some extending over 20,000 years of landscape evolution, reveal how factors like the initial slope angle and the rate of soil creep influence the final landform shape.

Beyond this introduction is the good thing to do is to read the thesis. You can find it by clicking on the Figure above (one beautiful painting by Vincent Van Gogh). 

Giulia Merler using GEOSPACE for simulations on a wineyard

Giulia Merler's Bachelor thesis investigates the effects of climate change on Trentino vineyards using GEOSPACE simulation software. While this research is preliminary and requires further validation, it reveals compelling insights into how global warming affects vine health.

For a high-resolution view of the poster, simply click on the image below. All supporting materials and data are accessible via the QR code provided.

The findings may surprise those unfamiliar with viticulture: elevated temperatures place significant stress on grapevines, posing a serious threat to wine production. While experts in the field may find this unsurprising, seeing the impact quantified through sophisticated modeling tools brings the reality into sharper focus and provides valuable data for future planning. This work pairs with the one by Marco Feltrin that can be found here. 

Anna De Nardi works on EPNs

Our three years degrees on Environmental Sciences requires for graduation a work exercise that can be really interesting.  Among the last graduations, I followed the work by Anna De Nardi, on the Extended Petri Net which I think was particularly valuable.  Anna, in fact, elaborated  all the MARRMoTs hydrological models using the Extended Petri Nets graphs. This show in practice that the statement that all the integral distributed hydrological models  (a.k.a. lumped distributed models, the hydrological dynamical systems, semi-distributed hydrological models) can be represented by means of the EPN is actually true. 

The MARRMoT (Modular Assessment of Rainfall-Runoff Models Toolkit) is a collection of hydrological models designed to simulate rainfall-runoff processes across different scales and conditions. Developed in MATLAB, MARRMoT provides a modular framework that allows researchers to compare, assess, and implement various conceptual hydrological models consistently

Key Features and Purpose:

• Model Variety: MARRMoT includes 47 different rainfall-runoff models, ranging from simple lumped models to more complex, semi-distributed structures.
• Comparative Framework: Its standardized environment enables researchers to test models on the same data sets and conditions, making it easier to assess model performance, suitability, and sensitivity.
• Customization: Users can adapt model configurations or modify parameters within the toolkit to better fit specific research objectives or study basins.
• - Open-source and Accessible: MARRMoT is open-source, encouraging community contributions, modifications, and application across various hydrological research areas.

This toolkit is particularly useful for researchers looking to improve or validate hydrological models and for those who want a structured way to compare the effects of model structure on hydrological simulation outcomes.

Extended Petri Nets (EPNs) are an enhanced version of traditional Petri Nets, a mathematical modeling language originally designed to describe distributed systems with concurrent, asynchronous processes. EPNs incorporate additional features to increase modeling flexibility and allow for more complex system representations. In scientific modeling, they are used for representing and analyzing systems in which discrete and continuous interactions are critical, like biological networks, chemical reactions, ecological models, or engineering processes.

For getting the poster at higher resolution, please click on the above image. For all the material, including all the MARRMoT models, please look at the here:

• All the  MARRMoT models represented with EPNs
• All the  EPNs-MARRMoT models catalogued by the number of reservoirs
• All the fluxes variables names found in the MARRMoT  collections

Other Hydrological models represented with EPNs can be found in Bancheri et al., 2019 below. 

References

Bancheri, Marialaura, Francesco Serafin, and Riccardo Rigon. 2019. “The Representation of Hydrological Dynamical Systems Using Extended Petri Nets (EPN).” Water Resources Research 55 (11): 8895–8921. https://doi.org/10.1029/2019WR025099.

Marialaura, Bancheri, Francesco Serafin, and Riccardo Rigon. 2019. “Supporting Material for: The Representation of Hydrological Dynamical Systems Using Extended Petri Nets (EPN).” Water Resources Research.

Knoben, W. J. M., J. Freer, K. J. A. Fowler, M. C. Peel, and R. A. Woods. 2019. “Modular Assessment of Rainfall-Runoff Models Toolbox (MARRMoT) v1.0: An Open Source, Extendable Framework Providing Implementations of 46 Conceptual Hydrologic Models as a Ontinuous Space-State Formulations.” GMDD, February, 1–26.

Knoben, W. J. M., J. Freer, K. J. A. Fowler, M. C. Peel, and R. A. Woods. 2019, “Modular Assessment of Rainfall–Runoff Models Toolbox (MARRMoT) v1.2: An Open-Source, Extendable Framework Providing Implementations of 46 Conceptual Hydrologic Models as Continuous State-Space Formulations.” 2019. Copernicus GmbH. https://doi.org/10.5194/gmd-12-2463-2019-supplement.

Rigon, Riccardo, and Marialaura Bancheri. 2021. “On the Relations between the Hydrological Dynamical Systems of Water Budget, Travel Time, Response Time and Tracer Concentrations.” Hydrological Processes 35 (1). https://doi.org/10.1002/hyp.14007.

Trotter, L., W. Knoben, K. Fowler, M. Saft, and M. Peel. 2022. “Modular Assessment of Rainfall–Runoff Models Toolbox (MARRMoT) v2.1: An Object-Oriented Implementation of 47 Established Hydrological Models for Improved Speed and Readability.” Geoscientific Model Development, August. https://doi.org/10.5194/gmd-15-6359-2022.

STRADIVARI Project VI: Informatics and Small Language Models Revolution

Back to index <----

Complex Earth System problems demand sophisticated computational architectures, yet current modeling frameworks face significant limitations in balancing scientific rigor with accessibility. While Rigon et al. (2022) demonstrated that Modeling By Components strategies offer promising approaches for tackling such challenges, true implementation remains constrained by existing software architectures and knowledge transfer barriers across disciplinary boundaries. The interdisciplinary nature of coupled Earth System modeling creates substantial knowledge transfer barriers: effective research requires integration of informatics, software engineering, numerical methods, soil science, plant physiology, and atmospheric physics, competencies rarely unified in standard curricula. Traditional documentation approaches, written manuals, video tutorials, and workshops, fail to provide the interactive, contextual assistance needed for complex modeling frameworks.

Gould 

OMS3 framework (David et al., 2013) represents one of the most advanced implementations of component-based environmental modeling, yet several critical gaps persist. ML integration within existing frameworks remains rudimentary. While Serafin et al. (2021) demonstrated basic ML capabilities in OMS3, current implementations lack integration with modern ML libraries necessary for hybrid physics-ML approaches. This limitation prevents effective handling of massive datasets typical in large-scale studies and restricts development of computationally efficient surrogate models where full physics becomes intractable. The NET3 parallelization subsystem (Serafin et al., 2021) can represent complex systems as directed acyclic graphs, but performance limitations and inflexibility restrict its application to the dynamical systems pervasive in coupled ESS modeling.

Domain-specific SML trained on hydrological literature and extensive GEOframe documentation address this fundamental bottleneck by providing accessible interfaces to sophisticated modeling capabilities. Keeping in mind that technologies in this sector are rapidly evolving and breakthrough could change the technological approach, STRADIVARI will implement an innovative knowledge management system utilizing current compact language models (3-8B parameters) such as Phi-3.5-mini or Qwen2.5, fine-tuned on domain-specific content through parameter-efficient methods like LoRA (Hu et al., 2021). The system will integrate multiple knowledge sources: STRADIVARI GitHub repositories, approximately 1000 GEOframe tutorial videos, and the complete AboutHydrology blog archive (900+ posts spanning 15 years). Using retrieval-augmented generation architecture with vector embeddings and modern frameworks like LangChain, the system will provide interactive documentation and contextualized assistance. This democratization infrastructure is essential for community adoption: without lowering technical barriers, sophisticated coupling frameworks remain accessible only to specialists, limiting scientific validation opportunities.

STRADIVARI breakthrough: Modernizes the proven OMS3 framework to OMS4, implementing enhanced Service-Oriented Architecture with machine learning integration and domain-specific small language model (SML) creating an intelligent modeling platform trained on hydrological literature and extensive GEOframe documentation. This infrastructure will provide intelligent assistance for model configuration, parameter selection, and results interpretation. This democratizes access to complex Earth System modeling while maintaining computational rigor, enabling researchers worldwide to contribute to dynamic Earth System understanding through interfaces providing contextual assistance and automated workflow guidance. The OMS4 improvements address critical computational architecture limitations in three key areas. First, enhanced parallelization capabilities unify disparate computational paradigms within a coherent framework: NET3 improvements enable efficient handling of large-scale systems of ordinary differential equations governing biota population dynamics and vegetation processes, while seamlessly integrating with grid-based partial differential equation solvers for soil-atmosphere transport. Second, the Service-Oriented Architecture redesign facilitates dynamic coupling between previously isolated computational domains—population dynamics models can now exchange state variables with spatially distributed hydrological processes in real-time, enabling feedback mechanisms between biological activity and physical transport that were computationally prohibitive in OMS3. Third, the integration of domain-specific Small Language Models represents a fundamental shift toward intelligent modeling infrastructure: rather than requiring users to navigate complex parameter spaces and component interactions manually, the SML provides contextual guidance for model configuration, interprets results, and suggests optimization strategies based on the extensive hydrological literature and GEOframe documentation corpus.

Back to index <----

References - Informatics and Small Language Models

• Belcak, Peter, et al. 2025. "Small Language Models Are the Future of Agentic AI." arXiv [Cs.AI]. arXiv.
• Chen, Min, et al. 2020. "Position Paper: Open Web-Distributed Integrated Geographic Modelling and Simulation to Enable Broader Participation and Applications." Earth-Science Reviews 207(103223): 103223.
• David, O., et al. 2013. "A Software Engineering Perspective on Environmental Modeling Framework Design: The Object Modeling System." Environmental Modelling & Software: With Environment Data News 39(c): 201-13.
• Hu, Edward J., et al. 2021. "LoRA: Low-Rank Adaptation of Large Language Models." arXiv [Cs.CL]. arXiv.
• Moore, R. V., and A. G. Hughes. 2017. "Integrated Environmental Modelling: Achieving the Vision." Geological Society, London, Special Publications 408(1): 17-34.
• Rigon, R., et al. 2022. "HESS Opinions: Participatory Digital Earth Twin Hydrology Systems (DARTHs) for Everyone: A Blueprint for Hydrologists." Hydrology and Earth System Sciences, January, 1-38.
• Serafin, Francesco, et al. 2021. "Bridging Technology Transfer Boundaries: Integrated Cloud Services Deliver Results of Nonlinear Process Models as Surrogate Model Ensembles." Environmental Modelling and Software[R] 146(105231): 105231.

Back to index <----