Carbonate karst aquifers, which cover 15% of Earth's surface and provide freshwater to nearly 25% of the global population, are facing depletion stress due to increased water abstraction, land use change, and climate change. The assessment of water resources in karst aquifers remains a challenge compared to classical porous aquifers due to their dual-to-triple porosity behavior. 

Karst aquifers are characterized by non-linear flows, with diffusive recharge and slow flow through a fissured matrix, and concentrated recharge and quick flow via conduit networks. They also encompass an original uppermost weathered layer called epikarst, which is sensitive to evapotranspiration and controls the recharge from the vadose zone towards the matrix and conduits. Moreover, for large aquifers with impermeable surfaces, known as binary karst, two recharge components coexist: direct infiltration of precipitation over karst outcrops and concentrated infiltration of runoff water from neighboring non-karst terrains. Nowadays and especially since 2022, a particular attention is devoted to the estimation of the annual low water dynamics in these aquifers. 

This constitutes a new challenge for the scientific community since this critical period is largely dependent of the winter and spring recharge process and of the different withdrawals on the watershed. Now, the accuracy of the estimation of the annual minimum discharge is not sufficient for the water resources agencies and a strong societal pressure lead to a need for rapid advances. The project aims to answer to two main methodological issues: 

[1] How the incertitude inherent to the different meteorological/climate forcings of the hydrogeological systems propagates in a hydrological conceptual model and generate incertitude in the low flow estimations statistics? This issue is raised based on KarstMod platform that already demonstrates its ability in karst discharge modeling. In terms of input data, we will compare gauge-based precipitation datasets in estimating the water balance and streamflow of the watersheds against precipitation from reanalysis products (SAFRAN, COMEPHORE, and ERA5-Land) and the satellite-based precipitation (GPM-LR and SM2RAIN-ASCAT), both at original resolution and downscaled to 1 km. Concerning evapotranspiration estimation, the remote sensing driven model SimpKcET, which estimates plant transpiration and soil evaporation based on soil’s water retention capacity and the vegetation fraction cover, is considered. The daily ET from SimpKcET will be compared to the ET from SAFRAN, GLEAM, ERA5-Land, MOD16 as well as eddy-covariance station. A propagation of ET incertitudes will be implemented in order to be able to include the final ET product from the next TRISHNA mission.

[2] How a dedicated optimisation procedure can improve the estimation of the low flow discharge statistics? We will explore here the possibility of a dynamic optimization process that will allows on one hand to consider a non-stationarity of the model parameters during the hydrological year and on the other hand a time varying objective function over the hydrological year. The introduction of a non-stationarity of the model parameters (function of time or function of the time dependant variables of the models such conceptual water levels) constitutes an original way to take into account a time dependant dynamics of the system. Moreover, a single objective function on a large interval often reduces its impact on high and medium discharge intervals. The introduction of a time-dependent objective function based on different power-law of simulated and observed discharge differences will allow a better constraint of the conceptual models and then a better assessment of the low flow simulations.

The project aims to propose a new paradigms in karst hydrogeological modeling by the introduction of new incertitude propagation methods and an original optimisation procedures that could be used more globally in hydrological modeling. The methodology will be applied to three catchments located over the Adour Garonne Water Agency territory: the Baget, the Ouysse and the Touvre. The not anthropized Baget catchment will constitute the test catchment since long time series are available. The Ouysse watershed is fed by both autogenic and allogenic recharge from external impermeable land. Its karst spring is a primary source of domestic water supply. The Touvre karst spring is the second largest water resurgence in France, recharged by an autogenic area on a karst impluvium and by water losses from sinking streams that drain the external Tardoire, Bandiat, and Bonnieure non-karst basins. The watershed is heavily exploited for crop irrigation and domestic water supply for over 110,000 inhabitants in Angoulême City. 

The application of the methodolody to these three catchments will allow an easy dissemination of the results to academic an public water resources communities.

=================

For more Information about the topics and the co-financial partner (found by the lab!); contact Directeur de thèse - david.labat@utlouse.fr ; david.labat@get.omp.eu

Then, prepare a resume, a recent transcript and a reference letter from your M2 supervisor/ engineering school director and you will be ready to apply online  before March 13th, 2026 Midnight Paris time!