TanDEM-X and TerraSAR-X-based Spring Flood Simulation on the Lena Delta, Siberia

Kurzfassung

Arctic watersheds heavily influence the oceans and the mean global temperature. They are the largest source of freshwater to the Arctic Ocean and affecting the sea ice cover and the ocean conveyor belt. An example of these impacts is the 7% rise of the average annual freshwater discharge from the six largest Eurasian rivers to the Arctic Ocean over the last century. The interaction between the Arctic water budget and climate change makes the study of arctic surface water important for climate- and environment-related research. The river Lena is one of the four main contributors to the Arctic Ocean freshwater, the others being Yenisei, Ob, and Mackenzie. It is the 10th longest natural river in the world, located in eastern Siberia with a basin the size of 2.4 × 106 km2 and an average annual water runoff of >500 km3. Extreme spring flood events take place annually in Lena, due to the melting of the accumulated snow from the previous winter. During this flood event, around 40% of its annual discharge is released to the ocean. The study area consisted of the central part of the Lena river delta where the main channel diverges into its major distributaries, namely the Trofimowskaya, Bykowskaya and Olenekskaya channels. This area consists of a braided river system with a vast floodplain area, surrounded by cliffs on its east and west side. Most of the floodplain area was flooded during the spring flood. This remote area cannot be easily accessed, making data acquisition challenging and costly. The aim of this thesis is to perform a feasibility study in implementing hydraulic modeling methods using the TerraSAR-X and TanDEM-X datasets during the annual spring flood events between 2013 and 2019. This approach integrated two different methods: remote sensing and river hydraulic methods. This integration is widely known as remote sensing of rivers and has been an emerging sub-discipline on river hydraulic studies. The remotely-sensed datasets provided the parameters for hydraulic measures on poorly gauged regions, whereas hydraulic modeling helped derive parameters which could not be acquired by remote sensing. HEC-RAS by the US Army Corps of Engineers was selected as the hydraulic modeling tool. The selected module for the simulation was the 1-dimensional unsteady flow (dynamic wave) analysis. The resulting simulated inundation boundaries were validated using multi-temporal TerraSAR-X images. The other hydraulic variables, namely flow velocity and flood depth, were also extracted from the simulation. Additionally, the outcomes from the feasibility study, the strengths and weaknesses of the approach, were also expected. By learning from the weaknesses of the approach, solutions to improve and extend this study in the future were devised. With very limited field data availability (no field data of the river bathymetry and surface roughness), the assessed accuracy of the model ranged between 65-95%. Considering these results were achieved with many approximations of parameters with high uncertainties, the approach was regarded applicable. This means, with better field data acquisition, the accuracy of the model can be pushed even higher, with better calibration and validation processes. The modeled flow velocity before the flood event was lower than 1.5 m/s. Then the velocity built up to 6 m/s on some parts of the channel during the peak of the flood. During the peak flood of 2014 (the most severe flood during of the simulated timeframe), the flood depth on the floodplain ranged between 0.001 – 2.5 m. In the future, the proposed method of this feasibility study can be extended for an improved hydraulic study of Lena delta, such as sediment transport, water quality, freshwater-seawater interaction, or ecological modeling. The accuracy can be improved with better field acquisition and methods with more precision, such as 2D and 3D modeling