Prediction of meander migration for four rivers in Texas at roadway crossings

River meandering can endanger the hydraulic structures by lateral expansion and down-valley translation of meander bends. In this study, the RVR Meander model was used to predict the meander migration for four bridge crossings in the state of Texas, USA. The bridge crossings are SH 105 at the Brazos River, FM 787 at the Trinity River, US 90 at the Nueces River and SH 63 at the Sabine River. A widely used migration coefficient approach developed by Ikeda et al. (1981) was adopted, which relates migration rate linearly to vertically-averaged near bank excess velocity through the use of a coefficient of bank erosion. This study estimates erosion coefficient values for eroded banks from measured long term migration rates and computed near bank excess velocities. The values of erosion coefficients are calibrated using the RVR Meander model with the least prediction error. The predicted planform evolution using the model suggests that the migration coefficient approach used in the model can capture the growth of meander bends if they do not change the migration direction. In general, the model can predict migration well in lateral and downstream direction. This study also attempts to examine the extent to which erosion coefficient depends on resistance of bank material to fluvial erosion. Erosion Function Apparatus (EFA) erodibility values represent the resistance of bank material to fluvial erosion. At the upstream bends adjacent to the road crossings for which both erosion coefficients and EFA erodibility values were obtained, it was found that erodibility values follow the trend of erosion coefficients, unless significant hydraulic or geomorphic changes occurred in the watershed and in the channel due to the human activities. It opens up the possibility that erosion coefficients may be estimated directly from the bank material properties instead of using the historical maps. This suggests that for long term (several decades) meander migration riparian vegetation is not a major controlling factor. This hypothesis requires further testing with data sets from other road crossings over rivers.