AQUI-FR

Historical overview

The objectives of the first phase (since 2014) was to show the feasibility of such a tool via the construction of the AQUI-FR computer structure and the first evaluations, as well as to ensure the legal possibilities of exploiting hydrogeological applications in operational, and finally, to identify the elements of interest for the water managers. The construction of the AQUI-FR structure required several steps: (i) to gather the hydrogeological models on the same computer structure that can be mobilized in operational use, (ii) to integrate the different hydrogeological applications available, (iii) to converge towards a homogeneous treatment on a national scale, via at least the use of a common atmospheric forcing, (iv) to recalibrate the applications to make them more compatible with this new forcing.
This work was accompanied by an effort to process the input and output data. An important aspect of the input data is the integration of groundwater withdrawals over the simulated periods, which are difficult to acquire over long periods and for recent periods (2-year delay). The management of output variables requires specific work, taking into account the variables of interest for managers (variable, depth or geological layer, estimation period, estimation domain) and numerical constraints (disk volumes, calculation time). In parallel, two development efforts were carried out, in order to: (i) include a representation of the basement aquifers, with numerical experiments carried out on the Brittany aquifers, and (ii) to allow a correction of the initial states of the aquifers by integrating the available piezometric data.

Evidence of benefits from implementation

The results of the AQUI-FR project confirm the feasibility of bringing together independent hydrogeological models developed in different research institutes in the same coupling platform. All these models were initially developed and calibrated over shorter periods with heterogeneous geological and meteorological databases, but the evaluation of the long-term simulations that has been carried out shows a good comparison with the observations available for the same period. It confirms the relevance of using the AQUI-FR as a tool for long-term impact studies.

The other advantage of this platform is in its modularity. The AQUI-FR platform encourages the development of groundwater modeling where it is lacking and, more generally, it has the potential to be a valuable tool for many applications in water resources management and in water quality studies. research, for example in climate change studies and seasonal forecasts

Replication potential in SUDOE region

The trigger for implementing the practice was the publication of the Explore 2070 project results, which showed a disparate analysis, and thus the idea emerged to harmonize the modeling results to make forecasts. From a technical point of view, few obstacles were encountered. The obstacles encountered were in relation to intellectual property. Indeed, it was necessary to reach an agreement on the sharing of the source code of the Marthe model to implement it on the Météo France platform. On the conceptual side, the models were not perfectly calibrated: some model outputs overlap. From an administrative point of view, it was necessary to convince the regions, because the hydrogeological models were developed within the regional entities of BRGM. Finally, from a functional point of view, the maintenance of the platform requires a permanent researcher per partner. The project was also able to benefit from subsidies (about 100 k€ / year).

Future outlook

In terms of evolution, the project offers several perspectives in terms of: (i) improvement of physical processes: better integration of basement and karst aquifers, and (ii) development: evolution of numerical codes and techniques.

This project also offers prospects for development outside of France, in the French overseas departments and territories, but also prospects within the framework of PEPR (One Water program).

In the future, other regional models will be included to extend the coverage of AQUI-FR (bedrock aquifers located in Brittany). A new modeling method based on a rainfall-runoff model will be used to provide upstream river flows as boundary conditions for the MARTHE models that require it. Finally, since errors in the initial conditions can significantly alter the skill of the forecast, studies dedicated to data assimilation to improve the initial state conditions are also performed in parallel.

Key points of the innovative method

> Multi-model numerical platform
> Daily to seasonal forecasts
> Prospective modeling (up to 2100)