Underwater video as a monitoring tool to detect change in seagrass cover

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.jenvman.2005.08.021 Byline: Justin I. McDonald, Grey T. Coupland, Gary A. Kendrick Abstract: To date seagrass monitoring has involved the removal of seagrass from its environment. In fragile or highly disturbed systems, monitoring using destructive techniques may interfere with the environment or add to the burden of disturbance. Video photography is a form of non-destructive monitoring that does not require the removal of seagrass or interference with the environment and has the potential to be a valuable tool in monitoring seagrass systems. This study investigated the efficacy of video photography as a tool for detecting change in seagrass cover, using the temperate Australian species Amphibolis antarctica (Labill.) Sonder ex Aschers. Using visual and random point estimates of seagrass cover from video footage, it was possible to determine the minimum sample size (number of random video frames) needed to detect change in seagrass cover, the minimum detectable change in cover and the probability of the monitoring design committing a Type II error. Video footage was examined at three scales: transects (m apart), sites (km apart) and regions (tens of km apart). Using visual and random point estimation techniques, a minimum sample size of ten quadrats per transect was required to detect change in uniform and variable seagrass cover. With ten quadrats it was possible to identify a minimum detectable change in cover of 15% for uniform and 30% for variable seagrass cover. Power analysis was used to determine the probability of committing a Type II error from the data. Region level data had low power, corresponding to a high risk of committing a Type II error. Site and transect level data had high power corresponding to a low risk of committing a Type II error. Based on this study's data, managers using video to monitor for change in seagrass cover are advised to use data from the smaller scale, for example, site and transect level data. By using data from the smaller scale, managers will have a low risk of incorrectly concluding there has not been a disturbance when one has actually occurred. Author Affiliation: Faculty of Natural and Agricultural Sciences, School of Plant Biology M090, University of Western Australia, 35 Stirling Highway, Crawley 6009, Western Australia Article History: Received 16 April 2003; Revised 23 August 2005; Accepted 26 August 2005