A simple field method for stereo-photographic length measurement of free-swimming fish: Merits and constraints

A simple field method to correct for bias in stereo-photographic underwater measurements is presented and its usefulness to improve the accuracy of length estimates of free-swimming fish is tested. The calibration is based on the inclusion of stereo exposures of a slate of known dimensions at the beginning and end of each measuring session. and it allows correction for non-parallel optical axes, spherical aberration of the lenses, and for other sources of bias that may vary between sessions. A precision of +/-3% for replicate length measurements of a 30 cm slate is obtained at distances between 0.7 and 2.0 m. This compares well with reported values that are obtained with more sophisticated (laboratory) methods. However, application of this calibration does not significantly improve the accuracy of stereo measurements of the length of free-swimming fish (+/-2.5 cm or 7-11% of actual fish length), The limitations are caused by difficulties in recognizing identical extreme points at the body of fish that are photographed under field conditions. By averaging three or more replicate measurements the accuracy is improved to +/-1 cm (2.7-4.5%). We conclude that effort should be aimed at increasing sample size, rather than at improving equipment and correction procedures, when measuring free-swimming fish in their natural habitat