HSHongren Shen

Hongren Shen

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Hongren Shen has a total of 84 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Hydrological model intercomparison in cold regions: challenges, insights, and future paths – a case from Nelson-MIP II

    by Mohamed Moghairib, Tricia Stadnyk, Alain Pietroniro, Juliane Mai, Bryan Tolson, Hongren Shen, James R. Craig, Mahkameh Taheri

    Hydrological model intercomparison studies have been conducted to test the usefulness and limitations of currently available hydrological models. Such studies have been conducted mainly to identify...

  • Open Access

    Network-based subbasin-scale mapping of streamflow alteration in Ontario, Canada

    by Hongren Shen, Bryan A. Tolson, James R. Craig, Robert A. Metcalfe, Jonathan Romero-Cuellar, James J. Luce

    Dams, reservoirs, waterpower facilities, and other water-control structures can modify streamflow and propagate their influence downstream through connected river networks, complicating hydrologic modelling and aquatic ecosystem assessment. Large-domain assessments often rely on binary agency classifications at hydrometric gauges (e.g., “natural” and “regulated”) or on storage- and discharge-based metrics that are difficult to apply consistently where operational data are incomplete. We introduce the Streamflow Alteration Index (SAI), a parsimonious, network-based screening metric that estimates the percentage of upstream drainage area whose routed flow passes through mapped alteration signal sources. Under a first-order drainage-area scaling approximation, SAI provides a proxy for the fraction of streamflow potentially exposed to upstream alteration sources, rather than a measure of the exact magnitude of hydrologic alteration. The framework comprises two paired indices, SAI_I (0 %–100 %) and SAI_II (0 %–100 %), independently representing confirmed Level I and potential Level II sources; values are propagated downstream by drainage-area-weighted mixing and reset at new sources of the same level. We compiled 2528 infrastructure records, of which 2476 were retained after geospatial alignment and screening. Following source classification and consolidation of multiple records within the same subbasin, 643 unique subbasins were identified as Level I and/or Level II alteration signal locations. The resulting SAI v1.0 database provides province-wide subbasin and river-reach layers for screening potential instream alteration at gauged and ungauged locations. We evaluated SAI-derived natural, conditional, and altered classes against 148 Reference Hydrometric Basin Network (RHBN) gauges, 125 additional Water Survey of Canada (WSC) gauges labelled natural, and 215 WSC-regulated gauges. SAI classified 86 % of both natural reference groups as natural, with no RHBN gauge classified as altered. Among WSC-regulated gauges, 26 % were classified as altered, 40 % as conditional, and 34 % as natural, highlighting differences between binary agency labels and the SAI screening framework and identifying sites requiring follow-up assessment. SAI v1.0 therefore provides a practical, transparent, and reproducible screening baseline for large-domain hydrologic and ecological applications without requiring detailed operating rules and offers a scalable framework for broader network-based assessments of potential streamflow alteration. The SAI v1.0 dataset is available from Zenodo (Shen et al., 2026; https://doi.org/10.5281/zenodo.19547932).

  • Open Access

    Calibration Strategies for Improved Hydrological Predictions in Lake-dominated River Systems

    by Menaka Revel, Bryan A. Tolson, James R. Craig, Hongren Shen, Etienne Gaborit, Ming Han

    Efficiently calibrating watersheds with hundreds of lakes and mixed observational data remains a key challenge. This study assesses global and local calibration strategies to better represent the lake–river continuum in Southeastern Ontario. Implemented within the Raven Hydrological Modelling Framework, the routing model is forced with runoff from Environment and Climate Change Canada's GEM-Hydro model and incorporates over 40 anthropogenically screened streamflow and lake level gauges. To address water level biases from unknown vertical datums, we introduced a modified evaluation metric. We optimized routing parameters including bias correction, baseflow, routing coefficients, and lake crest widths. Global calibration used a single multi-metric objective function incorporating discharge and lake levels, while local calibration optimized parameters for gauge-specific sub-regions. Local calibration improved the median Kling-Gupta Efficiency (KGE) of river discharge from 0.37 (baseline) to 0.73. Furthermore, the local approach outperformed global calibration, improving discharge predictions at 88.6% of gauges during calibration and 67.7% during validation, increasing median KGE from 0.62 to 0.78. Lake level simulations also benefited from individual crest width calibration, improving KGE deviation (excluding bias) from 0.72 to 0.88. Spatially distributed bias correction of runoff inputs enhanced discharge accuracy, correlating with underlying geology. These findings emphasize the value of localized calibration for capturing site-specific dynamics in semi-distributed lake–river routing models. We recommend calibrating lake crest widths locally to improve the accuracy of surface water dynamics in lake-dominated watersheds.

  • Open Access

    Network-Based Subbasin-Scale Mapping of Streamflow Alteration in Ontario, Canada

    by Hongren Shen, Bryan A. Tolson, James R. Craig, Robert A. Metcalfe, Jonathan Romero-Cuellar, Jamie Luce

    Dams, reservoirs, and waterpower facilities are central to regional water management, but they can induce streamflow alterations that propagate downstream through connected river networks and complicate hydrologic modeling and aquatic ecosystem assessment. Existing streamflow indicators often quantify alteration only at discrete locations (e.g., hydrometric gauges), thus limiting spatially continuous, subbasin-scale mapping of where upstream infrastructure may influence downstream flows. We propose the Streamflow Alteration Index (SAI), a network-based screening metric that propagates point-based alteration signal sources downstream through a routing network to map potential alteration influence. The framework includes paired indices: SAI_I (0–100 %) and SAI_II (0–100 %), representing two levels of influence derived from confirmed (Level I) and potential (Level II) alteration signal sources. We implemented SAI to develop the SAI v1.0 database for Ontario, Canada, using the Ontario Lake and River Routing Product Version 2 (OLRRP v2) network and an inventory of 643 alteration signal sources compiled from provincial datasets. The resulting product provides seamless, high-resolution, network-based, subbasin-scale mapping of streamflow alteration across Ontario and enables SAI estimates at both gauged and ungauged nodes within the routing network. We validated SAI by classifying subbasins containing hydrometric gauges as natural, conditional, or altered and compared these classes against two independent references: Reference Hydrometric Basin Network (RHBN) gauges (with minimal human impact) and Water Survey of Canada (WSC) gauge “Natural” and “Regulated” labels. SAI shows strong agreement for gauges labelled natural: 86 % of RHBN natural gauges and 86 % of WSC-natural gauges are classified as natural by SAI. For WSC-regulated gauges, SAI classifies 74 % as conditional or natural. Overall, SAI v1.0 provides a practical, objective screening product for large-domain hydrologic and aquatic applications without requiring detailed dam operating information, and it offers a scalable pathway toward national- to global-scale, network-based screening products for streamflow alteration.

  • Open Access

    Learning Basin Similarity Through Combined Deep Learning and Random Forest Approaches for Improved Parameter Transfer in Ungauged Basins

    by Rojin Meysami, Qiutong Yu, Bryan Tolson, Hongren Shen, Rezgar Arabzadeh

    Parameter regionalization for ungauged basins remains a critical challenge in hydrological modeling. While traditional approaches rely on physical catchment descriptors or spatial proximity, and recent machine learning applications have focused primarily on direct streamflow prediction, there remains significant potential to leverage machine learning for improved parameter transfer strategies. This study explores novel approaches that combine Long Short-Term Memory (LSTM) networks and Random Forest (RF) models to predict basin similarity and optimize parameter transfer for physically-based hydrologic models. Using case studies from British Columbia's Fraser River Basin and Ontario's Great Lakes region, we test multiple methodologies for integrating deep learning with traditional parameter transfer approaches. Our primary benchmark is established through an exhaustive parameter transfer experiment using the Raven hydrological model, where parameters from each potential donor basin were transferred to every possible receiver basin across 10 independent trials. This benchmark represents the best achievable KGE via parameter transfer methods. Our framework employs a regional LSTM model to capture complex streamflow dynamics and characterize basin similarity, then explores various RF-based approaches to predict optimal donor-receiver basin pairs for parameter transfer. These methods are evaluated against both the exhaustive transfer benchmark and emerging machine learning approaches. Results indicate that thoughtfully combining deep learning and random forest techniques can capture nuanced relationships between basin characteristics and hydrological response similarity, advancing the state-of-the-art in parameter regionalization for ungauged basins while maintaining physical interpretability.

  • Open Access

    Streamflow Alteration Index (SAI): Mapping Dam and Reservoir Impacts on Streamflow for Improved Hydrological Modeling

    by Hongren Shen, Bryan Tolson, James Craig, Robert Metcalfe, Jonathan Romero Cuellar

    Dams and reservoirs are integral to regional water management, providing critical services such as flood and drought control, water supply, hydropower generation, and recreation. However, their streamflow regulation often disrupts hydrological connectivity, sediment transport, and biodiversity, leading to significant ecological consequences. These alterations modify flow regimes across various time scales (hourly to annual), complicating the accuracy of hydrological models in affected regions. Thus, understanding how dam-induced streamflow alterations propagate through river networks is essential for informed water resource management.Current flow regulation indicators, such as those official flags from Water Survey Canada (WSC), are point-scale binary values that often under- or over-estimate regulation effects and lack spatial continuity. To address this limitation, we propose a spatially continuous metric, the Streamflow Alteration Index (SAI), which incorporates point-based alteration signals from dams, reservoirs, lakes, hydropower facilities, and hydrometric gauges into a subbasin-scale river and routing network. The SAI allows hydrologists to quantify cumulative upstream streamflow alterations at any point in a vector-based routing network. Using Ontario, Canada, as a case study, we applied the SAI to a network encompassing 245,576 subbasins, 82,928 lakes, and over 3,000 alteration sources identified from provincial and global datasets. This approach produced a seamless, high-resolution map of streamflow alteration signals across Ontario (total area: 1.07 million km2) at the subbasin scale, importantly covering both gauged points (including 1,320 flow and level gauges) and ungauged locations within the routing network. The SAI was validated against nearly 500 hydrometric gauges with WSC regulation flags.Results demonstrate that the SAI effectively identifies near-natural gauges with over 95% accuracy while revealing that more than 40% of gauges that are flagged as regulated by WSC could instead be reconsidered as model calibration targets, as many of them show little signs of significant regulation. By offering a less restrictive yet more reliable alternative, the SAI enables hydrologists to retain a larger pool of near-natural gauges for calibration, thereby enhancing streamflow predictions, particularly in data-sparse or ungauged regions. Furthermore, the SAI approach can be generalized to other routing networks in Canada and globally.

  • Top Cited Papers
    Open Access

    Enhancing long short-term memory (LSTM)-based streamflow prediction with a spatially distributed approach

    by Qiutong Yu, Bryan A. Tolson, Hongren Shen, Ming Han, Juliane Mai, Jimmy Lin

    Deep learning (DL) algorithms have previously demonstrated their effectiveness in streamflow prediction. However, in hydrological time series modelling, the performance of existing DL methods is often bound by limited spatial information, as these data-driven models are typically trained with lumped (spatially aggregated) input data. In this study, we propose a hybrid approach, namely the Spatially Recursive (SR) model, that integrates a lumped long short-term memory (LSTM) network seamlessly with a physics-based hydrological routing simulation for enhanced streamflow prediction. The lumped LSTM was trained on the basin-averaged meteorological and hydrological variables derived from 141 gauged basins located in the Great Lakes region of North America. The SR model involves applying the trained LSTM at the subbasin scale for local streamflow predictions which are then translated to the basin outlet by the hydrological routing model. We evaluated the efficacy of the SR model with respect to predicting streamflow at 224 gauged stations across the Great Lakes region and compared its performance to that of the standalone lumped LSTM model. The results indicate that the SR model achieved performance levels on par with the lumped LSTM in basins used for training the LSTM. Additionally, the SR model was able to predict streamflow more accurately on large basins (e.g., drainage area greater than 2000 km2), underscoring the substantial information loss associated with basin-wise feature aggregation. Furthermore, the SR model outperformed the lumped LSTM when applied to basins that were not part of the LSTM training (i.e., pseudo-ungauged basins). The implication of this study is that the lumped LSTM predictions, especially in large basins and ungauged basins, can be reliably improved by considering spatial heterogeneity at finer resolution via the SR model.


  • Open Access

    Enhancing LSTM-based streamflow prediction with a spatially distributed approach

    by Qiutong Yu, Bryan A. Tolson, Hongren Shen, Ming Han, Juliane Mai, Jimmy Lin

    Deep learning (DL) algorithms have previously demonstrated their effectiveness in streamflow prediction. However, in hydrological time series modelling, the performance of existing DL methods is often bound by limited spatial information, as these data-driven models are typically trained with lumped (spatially-aggregated) input data. In this study, we propose a hybrid approach, namely the Spatially Recursive (SR) model, that integrates a lumped long short-term memory (LSTM) network with a physics-based hydrological routing simulation for enhanced streamflow prediction. The lumped LSTM was trained on the basin-averaged meteorological and hydrological variables derived from 141 gauged basins located in the Great Lakes region of North America. The SR model involves applying the trained LSTM at the subbasin scale for local streamflow predictions which are then translated to the basin outlet by the hydrological routing model. We evaluated the efficacy of the SR model on predicting streamflow at 224 gauged stations across the Great Lakes region and compared its performance to that of the standalone lumped LSTM model. The results indicate that the SR model achieved performance levels on par with the lumped LSTM in basins used for training the LSTM. Additionally, the SR model was able to predict streamflow more accurately on large basins (e.g., drainage area greater than 1000 km2), underscoring the substantial information loss associated with basin-wise feature aggregation. Furthermore, the SR model outperformed the lumped LSTM when applied to basins that were not part of the LSTM training (i.e., pseudo-ungauged basins). The implication of this study is that the lumped LSTM predictions, especially in large basins and ungauged basins, can be reliably improved by considering spatial heterogeneity at finer resolution via the SR model.


  • Open Access

    Learning from hydrological models’ challenges: A case study from the Nelson basin model intercomparison project

    by Mohamed Ismaiel Ahmed, Tricia Stadnyk, Alain Pietroniro, Hervé Awoye, Ajay Bajracharya, Juliane Mai, Bryan A. Tolson, Hongren Shen, James R. Craig, Mark Gervais, et al.


  • Open Access

    On the elaboration of a robust calibration strategy for the large-scale GEM-Hydro model

    by Etienne Gaborit, Daniel Princz, Juliane Mai, Hongren Shen, Bryan Tolson, Vincent Fortin

    As part of the Great-Lakes Runoff Inter-comparison Project (GRIP-GL; Mai et al., 2022), which aims at comparing the performances of different hydrologic models over the Great-Lakes when calibrating them using the same meteorological inputs and geophysical databases, the GEM-Hydro hydrologic model used at Environment and Climate Change Canada (ECCC) to perform operational hydrologic forecasts was calibrated using different strategies. Following the calibration work related to GRIP-GL, progress has been achieved with regard to improving the calibration of the GEM-Hydro model.The work presented here focuses on improvements achieved with regard to calibrating the GEM-Hydro model, compared to the default version of the model and to the performances obtained during the GRIP-GL project. For various reasons explained, the GEM-Hydro calibration performed as part of GRIP-GL was suboptimal. The general calibration framework remains the same as in GRIP-GL, for example by using the MESH-SVS-Raven model to speed-up simulation times and transferring the calibrated parameters into GEM-Hydro afterwards, by relying on global calibrations for each of the 6 Great-Lakes subdomains, etc. However, several important changes have been made compared to the work performed in GRIP-GL, like a new approach to represent the effect of Tile Drains, changing the set of flow stations used for calibration, revising the objective function, etc.The proposed calibration methodology updates significantly improve GEM-Hydro streamflow performance across the Great-Lakes domain and in addition also improve or maintain similar performance levels as the default version of the model, with respect to auxiliary variables and surface fluxes: snow, soil moisture, evapotranspiration, 2m air temperature and dew point. Indeed, the model relies on 40m atmospheric forcings for wind speed, temperature and humidity, and simulates its own 2m atmospheric variables. To achieve this, it was necessary to constrain some parameter interval values during calibration, in order to prevent the calibration algorithm to choose physically-irrelevant parameter values that could allow to improve streamflow performances while degrading other hydrologic variables, due to equifinality.Reference:Mai, J., Shen, H., Tolson, B. A., Gaborit, E., Arsenault, R., Craig, J. R., Fortin, V., Fry, L. M., Gauch, M., Klotz, D., Kratzert, F., O'Brien, N., Princz, D. G., Rasiya Koya, S., Roy, T., Seglenieks, F., Shrestha, N. K., Temgoua, A. G. T., Vionnet, V., and Waddell, J. W. (2022). The Great Lakes Runoff Intercomparison Project Phase 4: The Great Lakes (GRIP-GL). Hydrol. Earth Syst. Sci., 26, 3537–3572. Highlight paper. Accepted Jun 10, 2022. https://doi.org/10.5194/hess-26-3537-2022