Hydrologic similarity based on width function and hypsometry: An unsupervised learning approach

The prediction of hydrologic conditions in watersheds has manifold applications, ranging from flood disaster preparedness to water supply and environmental flow management. In with scarce or no data, it is difficult make accurate predictions. Past work used similarity single-valued properties the terrain (for example, drainage area, mean slope) as basis relate gauged ungauged ones. resulting predictions show modest accuracy have a weak physical basis. this study, we develop physics-informed machine learning approach extract features that represent dynamics — width function hypsometric curve. These two geomorphometric measures are computed using functional forms fitted estimates derived digital elevation data. Furthermore, dynamically-similar groups identified based on results unsupervised clustering divergence measures. Our paves way towards flexible scalable can be assess improve prediction, one informed by physics surface generation transport watersheds. A case study involving 72 sub-watersheds Narmada River Basin (India) illustrate new methodology.