Estimating longitudinal dispersion in rivers using Acoustic Doppler Current Profilers

Chaopeng Shen, Jie Niu, Eric J. Anderson, Mantha S. PhanikumarAdvances 2010 · DOI

TL;DR

This paper demonstrates that Acoustic Doppler Current Profilers (ADCPs) can provide accurate estimates of longitudinal dispersion coefficients in rivers, offering a more efficient alternative to traditional tracer studies. By analyzing ADCP velocity data and comparing it with tracer study results across various rivers, the study validates the ADCP method and highlights its potential for routine environmental monitoring.

What Problem Does This Paper Address?

Accurately quantifying how solutes spread longitudinally in rivers and streams is crucial for predicting contaminant transport, assessing water quality, and understanding ecological processes. The traditional method for estimating the longitudinal dispersion coefficient (D) involves expensive and time-consuming tracer studies. While empirical formulas exist, they often yield estimates with significant variability, sometimes within an order of magnitude of the true value. This paper addresses the need for a more practical and reliable method to estimate D, specifically by leveraging the capabilities of Acoustic Doppler Current Profilers (ADCPs) and the shear-flow dispersion theory.

Methods

The core of this research lies in applying the shear-flow dispersion theory to ADCP velocity data. This theory, expressed by Equation (2) in the paper, relates the dispersion coefficient to the river's velocity profile and geometry. The authors utilized ADCPs to collect detailed, high-resolution velocity and bathymetry data across river transects.

Key methodological aspects include:

  • ADCP Data Processing: Raw ADCP data underwent rigorous processing, including:
  • Orientation: Velocities were oriented in the streamwise direction or projected onto a normal to the transect track.
  • Smoothing: Vertical velocity profiles were smoothed using power law or logarithmic fits to reduce noise.
  • Bad Ensemble/Bin Removal: Data points with low confidence were identified and either discarded or interpolated.
  • Dispersion Coefficient Calculation: The processed ADCP data was used to numerically integrate Equation (2) to estimate D.
  • Comparison with Tracer Studies: ADCP-derived D values were compared against D values obtained from traditional tracer studies. Two tracer-based approaches were used:
  • Transient Storage (TS) Model: Explicitly accounts for dead zones.
  • Advection-Dispersion Equation (ADE) Model: Does not explicitly separate dead zones, thus its dispersion estimates tend to be higher, incorporating some dead zone effects.
  • Reach-Averaging: For heterogeneous river reaches, a distance-weighted average of ADCP-derived D values from multiple transects was used to represent reach-averaged conditions, as suggested by Rutherford [2].

Datasets

The study utilized a comprehensive dataset comprising:

  • ADCP Data: Over 505 ADCP transects collected from seven rivers in Ohio, Indiana, and Michigan. These transects were collected under various flow conditions between 2000 and 2006.
  • Tracer Study Data: Approximately ten tracer studies were conducted on some of these rivers. Simultaneous tracer and ADCP data were available for three rivers (Ohio River, Grand River, and Burns Waterway). Tracer data for the Red Cedar River was collected during different flow conditions.
  • Literature Data: The paper also incorporated ADCP and tracer data from previous studies by Carr and Rehmann [5] and Fischer et al. [1] for broader comparisons.

Key Insights and Ideas

  • ADCPs as a Viable Alternative: The study strongly supports the ADCP method as a practical and accurate alternative to traditional tracer studies for estimating longitudinal dispersion coefficients, especially for larger rivers where tracer studies are particularly challenging.
  • Influence of Dead Zones: The comparison between TS and ADE model estimates revealed that the difference between them increases with flow. This suggests that dead zones (e.g., riparian vegetation, woody debris, hyporheic zones) play a more significant role in dispersion at higher flow conditions. The ADCP method appears to capture these dead zone effects implicitly, similar to the ADE.
  • Importance of Multiple Transects: For heterogeneous river reaches, using a median or distance-weighted average of dispersion coefficients from multiple ADCP transects provides a more reliable and representative estimate compared to a single transect. This approach helps to average out local variations and potential errors.
  • Processing Choices Matter: The method used for orienting velocity data (streamwise vs. normal to transect) had a more significant impact on the dispersion estimates than the method used for smoothing vertical profiles. This is because the shear-flow dispersion equation primarily uses the depth-averaged velocity, which is less sensitive to vertical profile smoothing.
  • Data Quality is Crucial: The presence of a high percentage of "bad ensembles" in ADCP data can significantly impact the accuracy of dispersion coefficient estimates, leading to unrealistically high values. Careful data quality control and filtering are essential.

Key Results

  • ADCP-derived dispersion coefficients were found to be in good agreement with tracer-based estimates, particularly when using median values from multiple transects within a reach.
  • For the Grand River, log-law smoothing with velocity projection in the streamwise direction yielded the best agreement with tracer results among the tested processing methods.
  • The study found that ADCP estimates of dispersion coefficients were comparable to, and often lower than, those derived from empirical relations, but generally closer to tracer estimates.
  • The ADCP method's ability to capture the influence of dead zones was supported by its good agreement with ADE estimates, which implicitly include such effects.

Limitations and Future Work

  • Limitations of ADCP Measurement: The paper acknowledges that ADCPs cannot make measurements close to the banks or in very shallow stream reaches, which can lead to an underestimation of dispersion in such areas.
  • Suitability for Complex Geometries: The current ADCP method is not ideal for stream reaches dominated by meander bends, recirculating regions, or strong secondary flows, as these can violate the assumptions of the shear-flow dispersion theory.
  • Future Directions: The authors suggest that newer ADCP models with smaller sensor heads, shorter blanking distances, and multiple acoustic frequencies hold promise for improving dispersion estimation, especially in complex environments. They also emphasize the need for additional datasets, including simultaneous tracer and ADCP data, to further validate and refine the approach, particularly for large rivers.

Summary updated

BibTeX
@inproceedings{shen2010estimating,
  author = {Chaopeng Shen and Jie Niu and Eric J. Anderson and Mantha S. Phanikumar},
  title = {{Estimating longitudinal dispersion in rivers using Acoustic Doppler Current Profilers}},
  booktitle = {Advances},
  year = {2010},
  doi = {10.1016/j.advwatres.2010.02.008},
}

Loading related content...

Figure