Search

We present here a strategy to obtain a realistic hydrological simulation over France with the ORCHIDEE land surface model. The model is forced by the Safran atmospheric reanalysis at 8-km resolution and hourly time steps from 1959 to 2020, and by a high-resolution DEM (around 1.3 km in France). Each Safran grid cell is decomposed into a graph of hydrological transfer units (HTUs) based on the higher resolution DEM to better describe lateral water movements. In particular, it is possible to accurately locate 3507 stations among the 4081 stations collected from the national hydrometric network HydroPortail (filtered to drain an upstream area larger than 64 km2). A simple trial-and-error calibration is conducted by modifying selected parameters of ORCHIDEE to reduce the biases of the simulated water budget compared to the evapotranspiration products (the GLEAM and FLUXCOM datasets) and the HydroPortail observations of river discharge. The simulation that is eventually preferred is extensively assessed with classic goodness-of-fit indicators complemented by trend analysis at 1785 stations (filtered to have records for at least 8 entire years) across France. For example, the median bias of evapotranspiration is −0.5 % against GLEAM (−4.3 % against FLUXCOM), the median bias of river discharge is 6.3 %, and the median KGE of square-rooted river discharge is 0.59. The spatial contrasts and temporal trends of river discharge across France are well represented with an accuracy of 76.4 % for the trend signal and an accuracy of 62.7 % for the trend significance. Despite inadequate performance in some specific regions (the Alps and the Seine sedimentary basin), this study offers a thorough historical overview of water resources and a robust configuration for climate change impact analysis at the nationwide scale of France. [ABSTRACT FROM AUTHOR]

Simple Summary: In national hospital databases, certain prognostic factors cannot be taken into account. Our objective was to estimate the performance of two models based on the Epithor clinical database and the French hospital database. The performance of the models was assessed with the Brier score, the area under the receiver operating characteristic (AUC ROC) curve, and the calibration of the model. For the Epithor and hospital databases, the training dataset (70% of the initial data) included 10,516 patients (with, respectively, 227 (2.16%) and 283 (2.7%) deaths) and the validation dataset (30% of the initial data) included 4507 patients (with, respectively, 93 (2%) and 119 (2.64%) deaths). The Brier score values were similar in the models of the two databases. For validation data, the AUC ROC curve was 0.73 [0.68–0.78] for Epithor and 0.8 [0.76–0.84] for the hospital database. The slope of the calibration plot was less than 1 for the two databases. This work showed that the performance of a model developed from a national hospital database is nearly as good as a performance obtained with Epithor, but it lacks crucial clinical variables. In national hospital databases, certain prognostic factors cannot be taken into account. The main objective was to estimate the performance of two models based on two databases: the Epithor clinical database and the French hospital database. For each of the two databases, we randomly sampled a training dataset with 70% of the data and a validation dataset with 30%. The performance of the models was assessed with the Brier score, the area under the receiver operating characteristic (AUC ROC) curve and the calibration of the model. For Epithor and the hospital database, the training dataset included 10,516 patients (with resp. 227 (2.16%) and 283 (2.7%) deaths) and the validation dataset included 4507 patients (with resp. 93 (2%) and 119 (2.64%) deaths). A total of 15 predictors were selected in the models (including FEV1, body mass index, ASA score and TNM stage for Epithor). The Brier score values were similar in the models of the two databases. For validation data, the AUC ROC curve was 0.73 [0.68–0.78] for Epithor and 0.8 [0.76–0.84] for the hospital database. The slope of the calibration plot was less than 1 for the two databases. This work showed that the performance of a model developed from a national hospital database is nearly as good as a performance obtained with Epithor, but it lacks crucial clinical variables such as FEV1, ASA score, or TNM stage. [ABSTRACT FROM AUTHOR]

From which part of the streamflow time series does a hydrological model learn most about the behaviour of a catchment? How does this part change with the objective function used for model calibration? We calibrated the hourly model GR4H (Génie Rural à 4 paramètres au pas de temps Horaire) on data subsets selected based on the flow duration curve, which narrowed the range of streamflow that influenced the identification of model parameters the most. Using 658 catchments in the United States and France, we found that calibrating GR4H on the highest 50% of streamflow values of a given period provided similar parameters to calibration on the whole period, with the top 10% dominating the identification of parameters. The logarithmic transformation of streamflow balanced the contribution from high and low flows, but only the reciprocal transformation emphasized low flows. The proposed methodology identifies the informative ranges of streamflow data and allows anticipating potential model failures in nonstationary hydroclimatic conditions. [ABSTRACT FROM AUTHOR]

Modelling and predicting the future of sandy shorelines is a key challenge in coastal research and is critical for sustainable coastal management. However, currently the most skillful shoreline models strongly rely on data to calibrate the free parameters, and are thus restricted to a few well monitored sites in the world. Here we address the challenges and opportunities offered by optical satellite imagery to provide useful information for equilibrium shoreline model calibration on cross-shore transport dominated sites. We focus on Truc Vert beach, southwest France, where previous work showed good equilibrium model skill to reproduce shoreline change from the time scales of hours (storms) to decades. Satellite derived waterlines are extracted over 11 years (2009–2020) and further transformed into satellite derived shorelines (SDS) with different water level corrections (e.g. tide and/or run up) and varying alongshore averaging lengths, and thus different uncertainties, in order to test model performance. Successively the timeseries duration and sampling frequency required for model calibration were also investigated. The model calibrated using the SDS data showed similar skill as the model calibrated using in-situ alongshore averaged shoreline positions, even for the uncorrected SDS dataset which Root Mean Square Error (RMSE) are approximately 30 m. Alongshore averaging was found to be the only necessary processing of the SDS data while any other site-specific corrections did not significantly improve model skill. Finally to further investigate the effect of sampling frequency and noise in the dataset we performed an analysis using a synthetic shoreline. Our results suggest that the effect of noise is negligible as long as the sampling frequency remains high (dt ≤ 30 days). Pending further validation, results show the strong potential of using uncorrected SDS dataset for shoreline model calibration at cross-shore transport dominated sandy coasts. • Model calibration using uncorrected satellite-derived shoreline data. • Simulated annealing extracts information from raw satellite derived shoreline data. • New perspective in modelling sandy shoreline change even when lacking field data. • Sampling frequency more critical than data quality in model calibration. [ABSTRACT FROM AUTHOR]

• Multi-criteria calibration of a lumped hydrological model based on discharge or stream-level and temporary stream data for 92 catchments in France. • The flow state of temporary streams was assumed to reflect groundwater storage. • Temporary stream observations did not improve overall model performance for most catchments but led to a better low-flow simulation for some catchments. • Temporary stream data constrained the uncertainty in the storage coefficient for the slowest reservoir, which affected the simulated groundwater level dynamics and total dynamic storage. Multi-criteria model calibration can lead to a better representation of hydrological processes and reduce parameter uncertainty compared to calibration on streamflow data alone. However, the additional data may be difficult to collect or aggregate into a representative catchment average value that can be used to calibrate a lumped model. Temporary streams are highly dynamic, and their flow state can be observed visually. However, data on the state of temporary streams are still uncommon and rarely used in hydrological catchment modelling. In this study, we used a unique dataset with discrete flow state observations for temporary streams in France and evaluated how informative these data are for calibrating a lumped, bucket-type hydrological model. We calibrated the HBV model for 92 catchments using discharge or stream-level data at different temporal resolutions (daily, one daily value per month, or one daily value per season) and used the observed flow states of temporary streams as a proxy of groundwater storage. Temporary stream data generally did not result in a better overall discharge simulation for the validation period. For catchments for which the model performance based on the calibration on only discharge or stream-level data was poor, it was more likely to lead to an improvement in model performance. The use of temporary stream data in combination with discharge data reduced the uncertainties in the low-flow simulations for up to half of the catchments. This improvement was caused by a better-constrained storage coefficient for the slowest groundwater reservoir and the elimination of parameter sets that led to substantial variations in groundwater storage. However, the improvements in low-flow simulations or parameter uncertainty due to the inclusion of temporary stream data in model calibration were not related to catchment characteristics. Thus, it remains unclear for which catchments temporary stream data can help to improve low-flow simulations and reduce parameter uncertainty. [ABSTRACT FROM AUTHOR]