Mangukiya & Sharma (2025)
"Integrating Reservoir Dynamics Into Differentiable Process-Based Hydrological Model for Enhanced Streamflow Estimation"

Develops a hybrid differentiable process-based hydrological model (dPLHBVRes) that integrates reservoir dynamics within a neural-parameterized HBV framework. The model enables joint learning of streamflow and unobserved internal hydrological states, bridging the gap between process understanding and predictive performance. It demonstrates that combining process knowledge with data-driven flexibility improves model reliability and interpretability in human-influenced catchments.

Mangukiya & Sharma (2025)
"Deep Learning-Based Approach for Enhancing Streamflow Prediction in Watersheds With Aggregated and Intermittent Observations"

Develops LSTM-based framework to estimate daily streamflow from aggregated or intermittent observations, addressing the challenge of temporal gaps in the data. The study shows that reliable daily predictions can be achieved even with monthly or weekly observations, offering a practical solution for hydrological modeling in data-limited and sparsely sampled catchments.

Sahu G, Mangukiya NK, & Sharma A. (2025)
"Does MC-LSTM model improve the reliability of streamflow prediction in human-influenced watersheds?"

Evaluates a mass-conserving LSTM against a standard LSTM across hydrologically diverse, human-influenced Indian watersheds. MC-LSTM delivers higher accuracy, reduced bias, and improved high-flow representation, particularly in highly regulated basins. The study underscores the value of physics-guided, mass-conserving deep learning models for more robust and reliable hydrological forecasting.

Mangukiya et al. (2023)
"How to enhance hydrological predictions in hydrologically distinct watersheds of the Indian subcontinent?"

The first large-sample evaluation of LSTM-based hydrological model across Indian catchments. The study demonstrates that DL models can significantly improve streamflow predictions over conventional methods, even under data-scarce conditions. It highlights how data integration across diverse catchments enhances model generalization and identifies the limitations in semi-arid and non-perennial basins.

Mangukiya & Sharma (2024)
"Alternate pathway for regional flood frequency analysis in data-sparse region"

Introduces a deep learning-based alternate pathway for regional flood frequency analysis (RFFA), combining data-driven streamflow prediction with at-site flood quantile estimation. The approach outperforms traditional ML-based RFFA methods in data-sparse regions by reducing prediction errors and improving model reliability. Beyond enhanced accuracy, it offers the flexibility to generate continuous streamflow time series, enabling estimation of multiple flow attributes and return periods in ungauged basins.

Mangukiya et al. (2024)
"A novel multi-model ensemble framework for fluvial flood inundation mapping"

Develops a multi-model ensemble framework that integrates flood extent and depth models for efficient fluvial flood mapping. By incorporating key flood conditioning factors and evaluating predictive, extrapolative, and generalization capabilities, the study demonstrates that the approach accurately captures flood dynamics across varying streamflow conditions, including unseen events. The framework offers a computationally efficient and scalable alternative to conventional hydrodynamic models.

Mangukiya & Yadav (2022)
"Integrating 1D and 2D hydrodynamic models for semi-arid river basin flood simulation"

Implements a coupled 1D–2D hydrodynamic modeling framework to simulate the 2017 Banas River flood in western India. The study highlights how man-made structures, particularly canal cross-drainage works, exacerbate flooding by restricting flow capacity. Model validation using extensive field surveys demonstrates the framework’s reliability and provides critical insights for improving flood management and canal design in semi-arid, infrastructure-influenced basins.

Mangukiya et al. (2026)
"Integrating Analytical Hierarchy Process and Machine Learning for Enhancing Flood Hazard Mapping"

Introduces an integrated Analytical Hierarchy Process–Random Forest (AHP–RF) framework that combines expert-based and data-driven insights for high-resolution flood hazard mapping in the Upper Assam region. The hybrid approach significantly improves hazard prediction accuracy compared to standalone methods, providing a reliable and interpretable tool for identifying highly susceptible zones and supporting flood mitigation planning.

Mangukiya & Sharma (2022)
"Flood risk mapping for the lower Narmada basin in India: a machine learning and IoT-based framework"

Proposes a machine learning-based framework to identify flood risk zones in the Lower Narmada Basin using geomorphological, hydrological, and socio-economic indicators. The study highlights elevation, land-use, and proximity to the river network as key drivers of flood risk. It also proposes the integration of IoT-based sensors for developing real-time flood monitoring and early-warning systems, supporting informed decision-making and community preparedness.