In the Siberian Altai Mountains, where the sources of the River Ob are located, ice-dammed lake outburst floods, so-called jökulhlaups, occurred in Pleistocene times. Valley glaciers extended within the upper Chuja River catchment and dammed the river upstream of the village of Aktash, which generated ice-dammed lakes in the intramountainous Kuray and Chuja Basins. Indicated by shorelines and ice-rafted boulder deposits, the maximum lake level reached an altitude of 2100 m above sea level, which reveals a maximum lake volume of 607 km 3 . The failure of the ice dam caused outburst floods, which left traces by giant bars, fluvial gravel dunes, and boulder deposits. Run-up sediments deposited in front of local obstructions along the valley slopes indicate a maximum depth of flow of 400 m above the valley bottom. Like the giant bars, they consist of characteristic relatively fine suspension gravels. Occasionally, secondary lakes are formed in tributary valleys that were blocked by giant bar deposits. The alternation of lacustrine deposits and suspension gravels in the tributary Injushka valley near Inja village give evidence for at least three large outburst floods. Age estimations by a variety of dating methods, such as luminescence methods, exposure dating, and accelerator mass spectrometry radiocarbon applied on different flood features and lake sediments, indicate that the outburst floods occurred between 40 ka and 13 ka. This estimation should be taken as preliminary, as problems occurred with the different dating techniques, resulting in partly contradictory results. This study focuses on the paleohydraulic reconstruction of the flood rather than describing flood-related features in detail. Seven different approaches are applied to estimate the discharge of the floods. Several new methods are developed within the study, partly based on previous approaches carefully considering the hydraulic background. Data for paleohydraulic estimations are obtained during repeated field trips by observations, surveys, and sedimentological investigations in Chuja and Katun valleys. Peak discharge is estimated by the elevation of the surfaces of the giant bars indicating the depth of flow and additionally by the run-up sediments, which reveal information about the velocity head of the flood flow along the valleys. A value of 10 × 10 6 m 3 /s is estimated for peak discharge, which is considerably less than previous estimations derived. Dimensions of gravel dunes and obstacle marks obtain information of flow conditions during the decreasing flood. The application of regression formulae between drained lake volume and peak discharge, an established method to estimate discharges of jökulhlaups, and estimation of flow velocity considering flow competence indicated by transported boulders yield problems as the magnitude of the outburst floods and the size of the boulders are beyond the level of experience. Based on the estimated peak discharge and the drained volume of water for the lake, the duration of a flood is estimated to be on the order of days, clearly less than one week. Considering the time needed to refill the lake after an outburst event leads to the rough estimation that recurrence intervals of floods are on the order of centuries. Open questions are intensively discussed throughout the study. For example, the relation of flood features to individual floods of the repeated outbursts is not possible unequivocally at the current state of knowledge. Further studies are required to obtain information on the extension and failure process of the ice dam, which would allow detailed modeling of the dynamics of the jökulhlaups. Also, methodological problems of dating techniques must be solved to develop a chronology of the floods in the Altai Mountains.