Recent developments in classifying bottom substrates and quantifying underwater vegetation make hydroacoustics an effective tool for monitoring and mapping habitat parameters in aquatic ecosystems. This study incorporated hydroacoustic sampling for bathymetry, substrate type, underwater vegetation, and fish distribution in Lake Washington, Seattle, USA. Data were collected using three independent echo sounder systems, to maximize data accuracy and vessel use, and geo-referenced using a Differential Global Positioning System, enabling the acoustic data to be used in a Geographic Information System. Results indicated that aquatic vegetation was limited to depths less than 8 m. This method proved to be an effective way to study various habitat influences on fish distribution, as well as map and monitor important physical and seasonal habitat parameters such as bathymetry, bottom character, and aquatic vegetation distribution.

Recent advances in seabed classification and submersed aquatic vegetation (SAV) detection using acoustic technology have made it possible to collect acoustic data on bathymetry, substrata type, plant coverage and height, and fish abundance and distribution. However, each type of acoustic application has optimal beam-width, frequency and pulse-width requirements for data collection. For example, when working in shallow water environments (0-100 m), bottom classification will require a low frequency, wide-beam echo sounder system; aquatic vegetation will require a high frequency, narrow-beam echo sounder system and fish data collection will require a 38-420 kHz, narrow-beam echo sounder. This may appear to expand the time and effort required to collect and process multiple types of acoustic data. However, by using multiple digital echo sounder systems, the entire data set can be collected in a single acoustic field survey, greatly enhancing the speed in which the data may be acquired.

Two recently developed software packages, BioSonics’ EcoSAV® and VBT Seabed Classifier, were used for data analysis. EcoSAV is designed to quantify SAV from hydroacoustic data. The software requires a digital echo sounder (BioSonics), differential GPS (DGPS) and computer for data acquisition. The echo sounder system collects and stores the digital echo signal from the echo sounder (either split- or single-beam) and DGPS data (latitude, longitude and time). The echosounder used to collect SAV distribution data was a DE 420 kHz, 6º split-beam echo sounder set to 0.1 ms pulse-width, 5 pings s-1, and a threshold of –130 dB. The processing algorithms in EcoSAV determine bottom depth, plant presence/absence, plant height, and aerial coverage.

A survey map was designed with a commercial navigation software package using BSB format maps produced by the United States National Oceanic and Atmospheric Association (NOAA). The navigation software was operated simultaneously with the hydroacoustic acquisition software in order to collect data along preestablished transects. The signal from an onboard Differential Global Positioning System (DGPS) was fed into both the navigation package and acoustic data collection software so that all data collected were concurrently geo-referenced.

Upon completion of the data acquisition, the data was processed to extract SAV height and cover. Accurate two-dimensional ground measurements (e.g., length, area) required that the DGPS data be mathematically transformed into a map projection system appropriate to the area of interest. All other data was subsequently imported into a GIS database for integration and mapping. EcoSAV calculations of plant height, water depth, and percent bottom coverage were transformed into percent plant biovolume: the percentage of water column with vegetation at a given report location. This transformation was performed by multiplying percent bottom cover by the ratio of plant height to water depth. Geospatial surface models were developed from final data on percent plant biovolume, percent bottom cover, sediment type, and bathymetry.

Aquatic vegetation was distributed widely and covered almost 60% of the survey area, but was confined to depths less than 8 m, presumably due to light-limitation. The highest density patches reached 100% cover, were relatively uniform in cover, and were most common in water less than 3 m deep. By biovolume, the greatest proportion of the area surveyed had plants that were between 1-19% of the water column. Three patches with high biovolume were observed: the northeast corner, the northwest corner, and along the west side of the area surveyed.

Overall, the study demonstrated that the various echo sounder systems could be combined to yield a wealth of data on aquatic habitat and fish distribution. Integrating the data in GIS was an effective way to evaluate fish distribution in regards to bathymetric features, substrate types, and patches of aquatic vegetation.

A surface of roughly 4sqkm has been covered with transects spaced 20m on average, resulting in a total track length of about 250km, taking some 800,000 samples (pings). A pre-release version of a specific software, based on SAVEWS (Submersed Aquatic Vegetation Early Warning System), developed by Bruce Sabol, USACE Waterways Experiment Station, Vicksburg, and currently under further development through BioSonics, Inc., Seattle, was deployed in order to process the split beam raw data (only single beam data is read by the program).

Previous to the survey, a number of fixed position observations over a frame (50x50cm) were done in order to dispose of data for calibration purposes. Afterwards the frame area has been fully sampled physically. Data on existence of plants as detected by the program was verified comparing the findings with the echograms. The percentage of plant detection over a cycle of 8 pings as a measure for the density or coverage of detected plants and plant mean height are also available from the program output. These parameters have been interpolated and subsequently presented on maps.

Additionally, ground truth data from 100 physical field samples is available to verify findings. The desire to distinguish the two species present in the area, Zostera marina and Zostera noltii, was not yet achieved based on the obtained data. Due to the specific configuration of the equipment during data acquisition, TS analysis with common tools are not easily done. An empirical approach to separate species based on a combination of height and depth has not yet concluded.

Finally, a series of problems, as for example the inclination of plants due to currents, are discussed (see attached document).