Tackling “the Ultimate Challenge” in Greater Depth
article written by James Thornton
11.05.20 | 10:05

I am delighted to join the MRI as a Scientific Project Officer and, as part of GEO Mountains, look forward to working in an interdisciplinary and collaborative fashion to improve the availability and accessibility of data pertaining to the earth’s mountainous regions. Below, by way of self-introduction, I take the opportunity to say a few words regarding my recent doctoral research.

Hydrological models are key parts of the simulation chain that is routinely employed to generate predictions of future water availability in mountainous areas. Three decades ago, developing reliable models of mountainous hydrological systems was described as “the ultimate challenge” (Klemes, 1990). Whilst many important advancements have been made in the intervening period, notably in terms of remote sensing technologies and computational power, the assessment continues to apply in certain respects. In this post, a few specific outstanding challenges are initially highlighted. Then, some of my own recent research that was undertaken with the intention of helping to improve our collective hydrological simulation capabilities in complex mountainous terrain is briefly outlined.

Firstly, as a result of the very orogenic processes that led to their formation, most mountainous regions exhibit considerable geological complexity, and it is increasingly recognised that this complexity can exert a strong influence upon hydrological processes. For example, calcareous sequences whose formations have highly contrasting hydraulic properties and which have been folded and faulted into complex geometrical arrangements are encountered across approximately 30% of the European Alps – with major implications for subsurface flow pathways and water balances.

a mountain view by James Thornton
The spectacular folds in the north face of the Petite Dent de Morcles, western Swiss Alps.

Conceptual hydrological models are still overwhelmingly relied upon to generate hydrological climate change impacts assessments. Whilst such models have numerous important benefits, they provide only extremely simplified representations of subsurface processes. Consequently, the role of groundwater in such systems may be overlooked, ultimately leaving the answers to key questions – such as the extent to which subsurface storage may “buffer” the streamflow response to climatic change – uncertain.

Secondly, other components of alpine environmental systems besides the glaciers and snowpack, such as the vegetation and permafrost, are also responding to climate change. Since such components also affect hydrology, and are linked together via a complex web of interactions and feedback mechanisms, any direct climate-related hydrological changes could be modulated by the response of the broader integrated system. Yet these “indirect effects” are also usually neglected in hydrological impact assessments.

Finally, snow is – of course – another defining feature of the hydrological regimes of temperate mountain catchments. Whilst this subject has received considerably more research attention than geological influences, for example, many hydrological models continue to rely on simple index-based melt estimation methods whose performance may be limited in extremely steep and rugged terrain.

In this context, and as part of the interdisciplinary “IntegrAlp” project, my research sought to evaluate the capabilities of a more sophisticated class of hydrological model – namely fully-integrated surface-subsurface models – in complex Alpine terrain; where they had never previously been applied. Being able to simultaneously simulate i) 3D (variably-saturated) groundwater flow (taking into account any subsurface structures defined), ii) 2D surface flows (which hold importance in terms of ecology, flood risk, and sediment transport), iii) bi-directional surface-subsurface exchange, and iv) evapotranspiration in a physically-based, transient, and spatially-distributed fashion, such models hold great potential to make more reliable and holistic hydrological projections. On the other hand, their data and computational requirements are high – the former being especially troublesome in mountainous terrain.

A river view by James Thornton
L’Avançon de Nant, Vallon de Nant, western Swiss Alps.

For instance, despite a wealth of traditional geological information being available in the Alps, very few (if any) detailed, regional scale 3D geological representations of sections thereof, with which integrated models might be informed, exist. An initial phase of work demonstrated that the generation of such datasets is now feasible datasets in even the most complex geological settings (Thornton et al., 2018).

In a second contribution, a novel approach to the calibration of distributed snow models was proposed and exemplified (Thornton et al., 2021). The particular code employed had an energy-balance core, but – bearing in mind the (often substantial) uncertainties related to meteorological input data and the representation of complex snow redistribution processes (e.g. gravitational redistribution) – several uncertain parameters also required estimation. Ultimately, observed snow patterns could be satisfactorily reproduced, predictive uncertainties quantified, and possible climate change impacts on snow dynamics assessed.

Following further fieldwork (e.g. groundwater level monitoring, geophysical surveys), a fully-integrated model of a 37 km2 region of the western Swiss Alps was developed and applied (Thornton et al., 2022). The final results indicate that whilst climate effects will likely dominate hydrological changes well into the latter part of the 21st century, additional evapotranspirative losses due to forest expansion alone are non-negligible at catchment scale. Undertaking such work more routinely could increase the confidence with which predictions of the future evolution of alpine hydrology can be made. That said, an enormous effort is required to develop and calibrate such models.

A fully-integrated model of a 37 km2 region of the western Swiss Alps
Numerical mesh of the fully-integrated surface-subsurface flow model of a 37 km2 region of the western Swiss Alps.

Overall, the experience of working with various techniques and across traditional disciplinary boundaries I gained through this research should, along with my previous education and employment, should stand me in good stead for me new role with the MRI.

I am likewise confident that by bringing together the observational capacities of a diverse range of global organisations to provide more and higher quality data in remote mountainous areas through initiatives such as GEO Mountains, the capacities of the latest computational tools can be leveraged more fully across both disciplines and geographies, ultimately leading to improved predictions and management decisions.

Klemes, V. (1990). The modelling of mountain hydrology: the ultimate challenge. IAHS Publ, 190, 29-43.

Thornton, J. M., Mariethoz, G., & Brunner, P. (2018). A 3D geological model of a structurally complex Alpine region as a basis for interdisciplinary research. Scientific Data, 5 (1), 1-20.

Thornton, J. M., Brauchli, T., Mariethoz, G., & Brunner, P. (2021). Efficient multi-objective calibration and uncertainty analysis of distributed snow simulations in rugged alpine terrain. Journal of Hydrology, 598 : 126241.

Thornton, J. M., Therrien, R., Mariethoz, G, Linde, N, &. Brunner, P.  (2022). Simulating fully integrated hydrological dynamics in complex Alpine headwaters: potential and challenges. Water Resources Research, 58 (4)

Thornton, J.M. Fully-integrated hydrological modelling in steep, snow-dominated, geologically complex Alpine terrain (2020). Doctoral Thesis, University of Neuchâtel. University of Neuchâtel.

Thornton, J.M. et al. Hydrological climate change impact assessment in the European Alps using a fully-integrated surface-subsurface flow model. In preparation.

Cover image by Pixabay user TeeFarm. All other photos courtesy of James Thornton.