Record number  479374 

Title  Rivers running deep : complex flow and morphology in the Mahakam River, Indonesia 
Author(s)  Vermeulen, B. 
Source  Wageningen University. Promotor(en): Remko Uijlenhoet, copromotor(en): Ton Hoitink.  Wageningen : Wageningen University  ISBN 9789462572065  150 
Department(s) 
Hydrology and Quantitative Water Management WIMEK 
Publication type  Dissertation, internally prepared 
Publication year  2014 
Keyword(s)  rivieren  morfologie  waterstroming  hydrologie  indonesië  rivers  morphology  water flow  hydrology  indonesia 
Categories  Hydrology 
Abstract  Rivers in tropical regions often challenge our geomorphological understanding of fluvial systems. Hairpin bends, natural scours, bifurcate meander bends, tie channels and embayments in the river bank are a few examples of features ubiquitous in tropical rivers. Existing observation techniques fall short to grasp the complex governing processes of flow and morphology. In this thesis new observational techniques are introduced and applied to study the Mahakam River, East Kalimantan, Indonesia. The observations reveal a new type of morphological regime, characterized by nonharmonic meanders, scour and strong variation of the crosssectional area. The anomalous geometry induces a complex threedimensional flow pattern causing longitudinal flow to be concentrated near the bed of the river. In Chapter 2 a wavelet based technique is introduced to characterize meander shape in a quantitative, objective manner. A scale space forest composed of a set of rooted trees represents the meandering planform. Based on the rooted trees, the locally dominant meander wavelengths are defined along the river. Submeander scale spectral density in the wavelet transform is used to determine a set of metrics quantifying bend skewness and fattening. Negative fattening parameterizes the socalled nonharmonic or hairpin bend character of meanders. The supermeander scale tree represents the embedding of meanders into largerscale fluctuations, spanning from doubleheaded meander scales until the scale of the valley thalweg. The new approach is used to quantify the anomalous planform geometry of the Mahakam River in a comparison with the Red River and the Purus River. The geometry of the Mahakam River is analyzed into more detail in Chapter 3, where the highly curved nonharmonic meanders are related to deep scours in the river bed. A total of 35 scours is identified which exceed three times the average river depth, and four scours exceed the river depth over four times. The maximum scour depth strongly correlates with channel curvature and systematically occurs half a river width upstream of the bend apex. Most scours occur in a freely meandering zone of the river. A systematic reconnaissance of the river banks reveals a switch of erosiondeposition patterns at high curvature. Advancing banks normally observed at the inner side of a bend are mostly found at the outer side of highcurvature reaches, while eroding banks switch from the outer side for mildly curved bends to the inner side for bends with high curvature. The overall lateral migration rate of the river is low. These results indicate a switch of morphological regime at high curvatures, which requires detailed flow measurements to unravel the underlying physical processes. Taking flow measurements in the deep scours of the Mahakam River presents a challenge to contemporary methods in hydrography. Acoustic Doppler Current Profilers (ADCPs) are capable of profiling flow velocity over large distances from a research vessel, but the existent data processing techniques assume homogeneity of the flow between the divergent acoustic beams. This assumption fails for complex three dimensional flows as found in the scours. In Chapter 4 a new ADCP data processing technique is developed that strongly reduces the extent over which the flow needs to be assumed homogeneous. The new method is applied to flow measurements collected in a river bend with a scour exceeding 40 m depth. Results based on the new approach reveal secondary flow patterns which remain invisible adopting the conventional method. Chapter 5 aims to better understand flow in sharp bends, by combining analyses of the flow measurements from a deep scour with Large Eddy Simulations of the flow. The threedimensional flow field is strongly dominated by horizontal circulations at both sides of the scour. The dramatic increase in crosssectional area (from 2200 m2 to 7000 m2 ) plays a crucial role in the generation of the two horizontal recirculation cells. An existing formulation to predict water surface gradients in bends is extended to include the effect of crosssectional area variations, next to the effect of curvature changes. Variation in the crosssectional area develops adverse water surface gradients explaining the flow recirculation. The depth increase toward the scour causes a strong downward flow (up to 12 cm s − 1 ) creating a nonhydrostatic pressure distribution, steering the core of the flow toward the bed. The latter aspect is poorly reproduced by the Large Eddy Simulations, which may relate to the representation of turbulent shear stresses. In Chapter 6 a novel technique is introduced to better monitor turbulence properties in complex river flows from ADCP measurements, exploiting what is discarded in observations of the mean flow. It extends the socalled variance method, using two ADCPs instead of one. The availability of eight acoustic beams, four from each ADCP, changes an otherwise unsolvable set of equations with six unknowns into an overdetermined system of eight equations with six unknowns. This allows to solve for the complete Reynolds stress tensor, yielding profiles of Reynolds stresses over almost the entire water column. Widely applied assumptions on turbulence anisotropy ratios are shown to be incorrect, which reveals a knowledge gap in open channel turbulence. Chapter 7 uses the technique developed in Chapter 6 to investigate the degree in which bed shear stress can be monitored continuously from an ADCP mounted horizontally at the river bank (HADCP). A calibrated boundary layer model is applied to estimate timeseries of crossriver bedshear stress profiles from HADCP velocity measurements. It is concluded the HADCP measurement can represent the regional bed shear stresses, as inferred from a logarithmic velocity profile, reasonably well. These regional bedshear stresses, in turn, poorly represent the local estimates obtained from coupled ADCP measurements, which are more directly related to processes of sediment transport and complex river morphology. Detailed observations of turbulence properties may be the key to improve our understanding of complex river flow and morphology.

Comments 