Expanded Method for Filtering Submerged Objects in Echosounder Data

Methods for filtering raw bathymetric field data are useful when developing multi-dimensional hydraulic models for river instream flow studies. Raw field data characterizing a submerged surface may include spikes or irregularities arising from instrument issues (e.g., signal loss, bubbles) or from actual objects (e.g., large woody debris) resting on or above the channel bottom. Identification and filtering of large woody debris (LWD) in single-beam echosounder data has been investigated by others and these concepts can be extended to other submerged objects. An improved application of a median filter approach is presented here that includes a method for varying the filter half-length to account for changes in boat speed. With improvements to the identification method, typical raw datasets can be used to more effectively locate submerged objects according to size class. When developing a hydraulic model, locating filtered objects aids in identifying areas of bathymetric complexity and calibrating roughness coefficients.