HISTORICAL AND SEASONAL TRENDS IN FLOOD CONVEYANCE, TISZA RIVER, HUNGARY

We have compiled 45 to 170 years of hydraulic and hydrologic data for six measurement stations situated on the Tisza River in Hungary: Tivadar, Vásárosnamény, Záhony, Kiskoré, Szolnok and Szeged. These data were obtained from the Hungarian hydrological authority (VITUKI) in Budapest. This research has two objectives: (1) to identify the causal mechanisms driving a recent increase in flood stages, and (2) test for a seasonal change in the stage-discharge relationship. The first objective involved testing the hypothesis that increased flooding on the Tisza River is the result of conveyance loss due to (1) channel and floodplain aggradation and (2) changes in riparian vegetation growth. Increased flood stage for a constant discharge can be caused by three and only three geometrical changes: decrease in width, decrease in depth and/or decrease in flow velocity. The second objective of this project involved testing the hypothesis that there is an identifiable seasonal signal on the Tisza River, and secondarily that this signal is linked to vegetation. Decreased vegetation growth and the absence of leaves during fall can cause lower stages than those measured at the same discharge during spring. The higher stages result from an increased roughness of the channel banks and floodplain, due to vegetation and/or other factors (e.g., temperature-dependent bedform roughness), and the resulting decrease in velocity. Specific-gage analysis was used to assess the causes on increased flood stage through the historic record and seasonality in flood stages. This analysis separates upstream (climate and land-use) factors from instream (conveyance-capacity) factors. Results show a decrease in conveyance of above-bankfull stages at all stations due to increased roughness of the un-leveed floodplain as well as due to levee construction near several of the study sites. Seasonal analysis shows stages peaking in May for in-channel flows, while out-of-channel flows peak in March. These months do not coincide with temperature max/minim, suggesting that roughness is controlled by bank and floodplain vegetation rather than bedforms. Other vegetation roughness studies have seen peaks later in August and September but note that vegetative roughness is sensitive to specific river reach characteristics.