Flash floods can be dangerous and happen quickly, affecting many areas. Scientists have created a new way to predict these floods globally using public data. They developed a method called TS-DUH that looks at how water flows through different landscapes and how rainfall and soil moisture affect flooding. By testing it with real flood events, they found that TS-DUH combined with GFMS-Runoff gives better predictions than other methods, with a high success rate in detecting floods and accurate simulations over a long period. This new approach shows promise for improving global flash flood forecasts.