"Take risks. Ask big questions. Don't be afraid to make mistakes; if you don't make mistakes, you're not reaching far enough." — David Packard
Key Research Contributions by the Group
Recession Flow Analysis
Hydrological processes responsible for transforming rainfall into streamflow and evapo-transpiration are infinitely complex and largely un-observable. Therefore, we generally utilize limited information to understand catchment hydrology through novel analysis. Recession flow analysis in particular provides us opportunities to understand the relationship between catchment storage (un-observable) and discharge. We follow the famous Brutsaert-Niber analysis method that expresses time-rate of change of discharge (dQ/dt) as a function discharge itself. However, we contradict the suggestion Brutsaert and Nieber that the (dQ/dt,Q) data points from all the recession events of a basin need to be analyzed together, which means the relationship between dQ/dt and Q is unique for a given basin. Instead, we have shown that dQ/dt-Q relationship changes considerably across events, and thus recession curves need to be analyzed individually.
The idea of individual recession analysis is already popular now. We have made several contributions based on individual analysis. We have shown that the dQ/dt-Q power law recession coefficient is a function of initial storage. Along those lines, we have established that past discharge information is useful for predicting future recession coefficients. We have also found that the key recession flow characteristics are shaped by catchment physio-climatic features.
For details about our contributions on recession flow analysis, see the following articles: https://www.sciencedirect.com/science/article/pii/S0309170814000116, https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/wrcr.20379, https://www.sciencedirect.com/science/article/pii/S002216941500445X, and https://onlinelibrary.wiley.com/doi/full/10.1002/hyp.11441.
Recession Flow Modelling
Recession flow modelling is an age old subject with early attempt dating back as early as 1877 by Boussinessq. The boussinesq model is based on the hypothesis that recession flow in river channels is a result of drainage from homogeneous phreatic aquifers that are laid on impermeable beds (the Dupuit assumption). However, subsurface domains of natural catchments are highly heterogeneous in most parts of the world, particularly in such regions. Using well water table data it has been shown that the usefulness of Dupuit-Boussinesq model in mountainous regions is very questionable (for details, see https://www.sciencedirect.com/science/article/pii/S0309170814000086). One of alternative recession flow models is geomorphological recession flow model (GRFM), which hypothesizes that recession flow characteristics of a basin are linked to its drainage network structure. The model basically suggests that recession flow in a basin is controlled by the dynamics (expansion and contraction) of its active drainage network (ADN), i.e., the part of the network that is witnessing water flow. Although the assumptions behind GRFM is not very robust, there is increasing evidence supporting the notion that recession flow, at least partly, is governed by ADN dynamics (e.g., see the recent paper: https://www.sciencedirect.com/science/article/pii/S0022169419300435 ). To overcome the limitations of the original GRFM, a modified model has been proposed that allows inclusion of a wide range of mechanisms (see, https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/wrcr.20379). Finally, suitable modifications have been incorporated to use GRFM in continuous streamflow simulation (see https://www.sciencedirect.com/science/article/pii/S0309170816305322).
Calibration-free Continuous Hydrological Modelling
Long‐term partitioning of rainfall is generally achieved using the zero‐parameter Budyko model which defines a dryness index. However, this approach is not suitable for dynamic partitioning particularly at diminishing timescales, and therefore, a universally applicable zero‐parameter model remains elusive. Recently, an instantaneous dryness index is proposed which enables dynamic hydrologic modeling using the original Budyko model (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2016GL070173 ). By introducing a “decay function” that characterizes the effects of antecedent rainfall and solar energy on the dryness state of a basin at a time, Dynamic Budyko (DB) model propose the concept of instantaneous dryness index to perform continuous simulation in totally ungauged basins. Although originally the model was developed and tested in the US, a recent study has shown that it performs reasonably in India without any calibration exercise (https://iwaponline.com/hr/article-abstract/doi/10.2166/nh.2019.081/65688/Evaluation-of-an-instantaneous-dryness-indexbased).