Gravel augmentation has become common practice to mitigate the effects of decline in upstream sediment supply in gravel-bed rivers. However, the functional aspects of river systems are often left out of rehabilitation monitoring programmes. Despite temperature being a fundamental parameter determining the general health of rivers, a limited number of studies have tested whether gravel augmentation can aid restoring thermal functions. Using airborne thermal infrared (TIR) imagery, this paper explores potential positive outcomes through the monitoring of gravel augmentation actions on 3 rivers in France using hydro-morphological indicators within a trajectory-based Before-After Control-Impact (BACI) framework. This design, combining long-term geomorphic evolution with TIR-based CI strategy, indicated that restoring forms was not sufficient to restore thermal functions in their spatial dimension. Despite forms showing various degrees of recovery, the thermal regime of rehabilitated reaches (i.e. temperature gradient, cold-water patches density, etc.) failed to recover. Nonetheless, hydro-morphological indices can be used to estimate long-term evolution of groundwater-surface water interactions. We emphasise the benefits of trajectory-based BACI assessment to identify current conditions, understand the past evolution (trajectory) of the system to define the framework within which rehabilitation can be objectively assessed, especially when assessing hydrological (here thermal) processes. From a hydrological perspective, the gap between restored forms and restored thermal processes stems from the alteration of connectivity pathways, which potentially require more time (or efforts) to be recovered than forms do. With an increasing number of rehabilitation schemes and increasing pressure of global changes on rivers, we suggest that monitoring of water temperature becomes a routine part of rehabilitation projects.

Groundwater pumping influences the rate of River-Aquifer (R-A) exchanges and alters the water budget of the aquifer. Therefore, fulfilling the total water demand of the area, with an optimal pumping rate of wells and optimal R-A exchanges rate, is important for the sustainable management of water resources and aquatic ecosystems. Meanwhile, comparison of the output of different simulation-optimization techniques, which is used for the solution of water resource management problems, is a very challenging task where different Pareto fronts are compared to identify the best results. In the present work, mathematical models were developed to simulate the R-A exchanges for the lower part of the River Ain, France. The developed models were coupled with optimization models in MATLAB environment and were executed to solve the multi-objective optimization problem based on the maximization of pumping rates of wells and maximization of groundwater input into the river Ain through R-A exchanges. The Pareto front developed by different simulation-optimization models was compared and analyzed. The Pareto fronts were juxtaposed based on the convergence, total diversity, and uniformity with the help of different performance metrics like hypervolume, generational distance, inverted generational distance, etc. The impact of different groundwater models based on domain size and boundary conditions was also examined. Results show the dominance of MOPSO over other optimization algorithms and concluded that the maximization of pumping rates significantly changes after considering the R-A exchanges-based objective function. It is observed that the model domain also alters the output of simulation-optimization, therefore the model domain and corresponding boundary conditions should be selected carefully for the field application of management models. ANN models were also developed to deal with the computationally expensive simulation model by reducing the processing time and were found efficient. Keywords: Simulation-Optimization, Multi-Objective optimization, Artificial Neural Network, River-Aquifer exchanges.

In this study, we unveiled the lumped effects at the reach spatial scale over three decades in one of the braided rivers in the Qinghai-Tibet Plateau of China, the Upper Lancang River (ULR). Using Landsat images obtained in 13 years between 1989 and 2018, we extracted flowing and non-flowing channels, active channel widths (unvegetated bars and flowing channels), and calculated lateral shifting rates of the main channel for the 13 periods. We also developed an empirical equation between vegetation area (Av) calculated from the high-resolution ortho-photo derived from an Unmanned Aerial Vehicle survey and Normalized Difference Vegetation Index for pixels of the Landsat image obtained at the same time. This relationship allowed us to estimate Av for other 12 selected years. We found that (1) braiding intensity increases with low discharges, indicating that the ULR is a very well-connected braided system with groundwater providing a large set of aquatic habitats, (2) this braided system is very well-supplied and actively shifting in relation to peak flow and flood duration, and (3) The ULR supports a progressive vegetation encroachment, which seems to be linked to temperature rising. Our study showed several similar morphological patterns to those in other braided rivers, such as the ones observed in the European Alps but much more active, well-supplied and highly connected. These similarities suggest that similar morphodynamic processes might take effect in the braided rivers with very high elevations and potentially high spots of biodiversity, indicating the ULR may be a reference for this region similarly to the Tagliamento in the Alps, but it seems that this system can be very sensitive to global change due to vegetation encroachment following temperature rising and decreases of low flows.