 The paper proposes a novel data-driven approach using long short-term memory, LSTM network to model rainfall runoff and catchments with snow influence. The LSTM's ability to learn long-term dependencies is essential for modeling storage effects. The approach is tested on 241 catchments of the camel's data set and compared to the Sacramento Soil Moisture Accounting Model, SACSMA, coupled with the Snow-17 snow routine. The LSTM shows potential as a regional hydrological model that predicts discharge for multiple catchments and it is also shown possible to transfer process understanding learned at the regional scale to individual catchments, increasing model performance. The results show that the LSTM outperforms SACSMA plus Snow-17, highlighting its potential for hydrological modeling applications. This article was authored by F. Krautsert, F. Krautsert, D. Klotz, and others.