Predicting spatiotemporal changes of channel morphology in the reach of Teesta River, Bangladesh using GIS and ARIMA modeling

The Teesta is one of the most dynamic river systems in Bangladesh, supporting the local livelihood patterns of millions of riparian people over a long time. However, there is an increasing concern regarding such a dynamic system, due to its channels being subjected to regular shifting and threatening engineering projects, which results in various socio-economic and environmental problems. Therefore, the objective of the study is to investigate the spatiotemporal changes of channel morphology and predict mid-line channel shifting in the reach of the Teesta River, Bangladesh during 1972-2031, using multi-temporal Landsat images data with GIS, spatial autocorrelation index, and an autoregressive integrated moving average (ARIMA) model. The results show that the temporal changes in channel shifting rates were mostly affected by high sedimentation in the river flow path, whereas the spatial changes along the channel reaches were primarily controlled by the differences of river banks, as demonstrated by the various bar area formations. The relationship between the bar area and the braided index exhibits a strong positive correlation (R-2 = 0.96, p < 0.01), while a moderate correlation occurs (R-2 = 0.50; p < 0.05) between the sinuosity index and time, suggesting a high rate of sedimentation and changing the flow of the river affected by the channel banks. The results of the ARIMA model indicate that the maximum mid-line channel shifting will occur in the rightward direction for the period between 2017 and 2024, and leftward mid-line channel shifting between 2024 and 2031. The spatial autocorrelation results suggest that the decreasing bar area increases the spatial pattern of the channel. Overall, the maximum lateral channel shifting occurs in the right bank side; thus, the channel reach is sensitive to lateral shifting and requires protection utilizing engineering structures, as a huge population and extensive agricultural lands are being displaced on its rightward floodplain.