A General Guide for Deriving Abundance Estimates from Hydroacoustic Data














Analysis expectations

Although it is easy to get caught up in logistical considerations (how long? lunar phase? equipment issues?) before a survey, it is essential to know that the collected data will answer the primary survey questions.  Scheaffer et al. (1996) suggest visualizing the final report before conducting the survey.  This reflection on survey and project objectives should be done:

  • throughout survey design, and;
  • frequently while conducting the survey. 
  • By frequently reviewing survey objectives, priorities can be established should it appear that key project questions are not being answered.

    Visualizing the final report also puts analysis options into perspective while the survey is being designed.  For fisheries acoustic surveys, both classical and geostatistical approaches may be used.

  • Classical, or design-based, analyses follow random sampling theory. The use of randomly selected sampling units facilitates computation by allowing the samples to be treated as independent.  Choosing an appropriate sampling design as discussed below may enhance precision in the estimates. 

  • Geostatistical, or model-based, analyses can make use of the information that exists in how organisms are distributed.  An underlying spatial pattern is detected from survey observations and that pattern is modeled to deduce the density distribution of the organisms elsewhere in the survey. 
  • The choice of survey design may depend on whether the analysis will be design-based or model-based.  Both methods, however, benefit from the appropriate allocation of samples across all possible elementary sampling units (i.e., the distribution of samples across the area to be surveyed).