Our first academically published paper demonstrates how an early version of our model outperformed traditional conceptual models, even though those models were calibrated to each basin individually, while our model was predicting “ungauged” in basins it had never seen before.
The paper demonstrated how physically informed, neural network based models learn fundamental hydrologic relationships which allow them to perform strongly in previously unseen basins and established that these models outperformed traditional physical hydrologic models across a broad range of hydrologic conditions.
Kratzert, F., Klotz, D., Herrnegger, M., Sampson, A. K., Hochreiter, S., & Nearing, G. S. (2019). Toward Improved Predictions in Ungauged Basins: Exploiting the Power of Machine Learning. Water Resources Research, 55(12), 11344–11354. https://doi.org/10.1029/2019WR026065