Modélisation Neige et Forêt

4 à 6 mois entre janvier et juillet 2023

Seasonally snow-covered forests make up a substantial share of the land surface in mountain and boreal regions of the Northern Hemisphere, therefore accurate representation of snow under forest is important for any large-scale snow model application in these areas. The presence of forest cover affects all energy and mass exchange processes between the atmosphere and the underlying snowpack, for example through interception of snowfall in the canopy, shading of the snow surface from solar radiation and emission of thermal radiation by the trees, leading to snow cover evolution dynamics that can be very different from those seen in open terrain. Forest-snow interactions are strongly dependent on the structure of the forest canopy. Consequently, and especially under discontinuous canopies, their magnitude can vary across scales of just a few meters. Commonly used vegetation schemes of land surface models, however, often represent canopy structure in terms of bulk properties such as Leaf Area Index and thus may fail to capture the effect of canopy structural heterogeneity on forest snow. At the same time, the small-scale variability of forest snow processes makes it challenging to evaluate models intended for coarse-resolution applications against point-scale observations (Rutter et al. 2009). In snow hydrology, a very recent approach to overcome this issue has been the development of hyper-resolution (1-5m) forest snow models that explicitly resolve canopy structure and thus successfully capture small-scale forest snow distribution patterns (Broxton et al. 2015, Mazzotti et al. 2020), and the subsequent use of such hyper-resolution simulations in model upscaling experiments to arrive at coarse-resolution modelling strategies (Mazzotti et al. 2021, Broxton et al. 2022).

The forest-snow scheme MEB-Crocus (Boone et al. 2017) was recently implemented within the SURFEX Earth surface modelling platform developed at CNRM for large scale applications (from about 50 km in General Circulation Models to hectometric resolution for the most-detailed configurations of Numerical Weather Prediction or hydrological modelling systems). In particular, Météo-France plans to operate by 2026 a new snow modelling system at 250m resolution over all French mountains. The goal of this internship is to explore and assess the application of MEB-Crocus to represent snow under forest at this target resolution. So far, MEB-Crocus has only been evaluated against point-scale observations at few measurement sites, including Col de Porte in the French Alps (Vincent et al. 2018, Bouchet, 2022), 3 sites in the Canadian boreal forest (Napoly et al., 2020) and 2 sites in Finnish boreal forest (Nousu et al., in prep.). Within the framework of this internship, we will apply the model over a 35 km2 domain in the Eastern Swiss Alps and evaluate its performance at hectometric scale by comparing its results to simulations from the Swiss Operational Snow Hydrology Modelling System. The Swiss model, FSM2, includes the state-of-the-art forest snow process representations from Mazzotti et al. 2021, with grid-cell scale canopy structure descriptors calculated based on a nation-wide 1m resolution canopy height model derived from airborne LiDAR. For the test region, downscaled meteorological forcing is available at multiple resolutions (25-250m) for water years 2016-2022, and meter-scale simulations exist for selected sub-domains and have been assessed against snow distribution observations (e.g. Mazzotti et al., 2022). This setup will provide a benchmark for MEB-Crocus simulations, allowing us to assess model performance for gridded applications in a novel way which extends existing evaluation efforts. This work will further serve as a test case for the future application of MEB-Crocus in the French Alps.

Overall, the work of the internship will include :
1) Setting up SURFEX / MEB-Crocus simulations over the test domain, potentially at multiple spatial resolutions, leveraging the meteorological forcing available through the Swiss modelling system and using ECOCLIMAP datasets as canopy structure inputs,
2) Comparing results to the output of the Swiss model simulations and investigate model performance for different canopy structures, topographic configurations and meteorological conditions.
3) Potentially identifying model shortcomings and suggesting improvements.

Potential candidates have a strong interest in cryospheric sciences and numerical modelling, and are proficient with the Python language and Linux environment. They should further enjoy winter fieldwork, as occasional support of ongoing forest snow monitoring efforts at Col de Porte to familiarize with the relevant physical processes will be part of the internship.

References :
Boone, A., Samuelsson, P., Gollvik, S., Napoly, A., Jarlan, L., Brun, E., and Decharme, B. : The interactions between soil–biosphere–atmosphere land surface model with a multi-energy balance (ISBA-MEB) option in SURFEXv8 – Part 1 : Model description, Geosci. Model Dev., 10, 843-872, doi : 10.5194/gmd-10-843-2017, 2017.

Bouchet, A. : Observation et modélisation de la neige sous la forêt de l’observatoire nivo-météorologique du Col de Porte (1325 m, Chartreuse). Internship report, Université Grenoble-Alpes / Météo-France, 2022.

Broxton, P. D., Harpold, A. A., Biederman, J. A., Troch, P. A., Molotch, N. P., and Brooks, P. D. : Quantifying the effects of vegetation structure on snow accumulation and ablation in mixed-conifer forests. Ecohydrol., 8 : 1073– 1094. doi : 10.1002/eco.1565, 2015.

Broxton, P. D., Moeser, C. D., and Harpold, A. A. : Accounting for Fine-Scale Forest Structure is Necessary to Model Snowpack Mass and Energy Budgets in Montane Forests. Water Resources Research, 57, 12, doi : 10.1029/2021WR029716, 2021.

Mazzotti. G., Essery, R., Webster, C., Malle, J., and Jonas, T. : Process-level evaluation of a hyper-resolution forest snow model using distributed multi-sensor observations. Water Resources Research, 56, 9, doi : 10.1029/2020WR027572, 2020.

Mazzotti, G., Webster, C., Essery, R., and Jonas, T. : Increasing the Physical Representation of Forest‐Snow Processes in Coarse‐Resolution Models : Lessons Learned From Upscaling Hyper‐Resolution Simulations, Water Resources Research, 57, 5. doi : 10.1029/2020WR029064, 2021.

Mazzotti, G., Webster, C., Queno, L., Cluzet, B., and Jonas, T. Canopy structure, topography and weather are equally important drivers of small-scale snow cover dynamics in sub-alpine forests. Hydrol. Earth Syst. Sci. Discuss. doi : 10.5194/hess-2022-273, in review, 2022.

Napoly, A., Boone, A., and Welfringer, T. : ISBA-MEB (SURFEX v8.1) : model snow evaluation for local-scale forest sites, Geosci. Model Dev., 13, 6523–6545, doi : 10.5194/gmd-13-6523-2020, 2020.

Nousu, J. P., Lafaysse, M., Mazzotti, G., Ala-aho, P., Marttila, H., Launiainen, S., Cluzet, B., Aurela, M. and Kolari, P. : Snowpack, soil and forest energy balance and flux partitioning in boreal ecosystems, in prep.

Rutter, N. et al. : Evaluation of forest snow processes models (SnowMIP2), J. Geophys. Res., 114, doi : 10.1029/2008JD011063, 2009.

Vincent, L., Lejeune, Y., Lafaysse, M., Boone, A., Le Gac, E., Coulaud, C., Freche, and G., Sicart, J.E. Interception of snowfall by the trees is the main challenge for snowpack simulations under forests. In : Proceedings of ISSW, 705-710. http://arc.lib.montana.edu/snow-science/objects/ISSW2018_O08.4.pdf, 2018.