Video or dive? Methods for integrated monitoring and mapping of marine habitats in the Hvaler-Koster area

Sammendrag

In order to prepare for integrated monitoring and mapping in the marine national parks Ytre Hvaler and Koster at the Norwegian-Swedish border this pilot study has evaluated alternative drop video techniques and diving observations in terms of taxonomic resolution, precision and cost. The report further gives recommendations regarding sampling design and the expected effort required for a certain precision is given using a specific drop video technique. Lastly the report highlights how monitoring and predictive mapping may be integrated by providing examples of species distribution modelling using various sampling designs. Two drop video techniques were initially compared, Free and Guided. The Guided technique, whereby ten randomized stops are analysed quantitatively and the entire film qualitatively, was shown to be less dependent of the person analysing the film, thereby providing a more precise estimate compared to Free interpretation. Taxonomic resolution was substantially better when diving compared to video, yielding about twice as many vegetation taxa and seven times as many animals. The difference between video and dive increased with diversity, and although video appeared to capture relative diversity, diving is required for certain identification of individual species. However, in terms of cover, video and diving largely identified the same dominating habitat forming groups. Similarly, the more comprehensive comparison from the related “Visual methods project” could in general not find any difference between Free video and dive regarding repeatability and uncertainty in terms of cover estimates. No comparison of uncertainty and repeatability between the Guided video technique and diving could be performed and the suitability of the Guided technique for other areas needs to be further evaluated. A stratified sampling design clearly reduced the effort needed for attaining a normally accepted precision and power in terms of monitoring cover and diversity. On hard substrates between 0 and 15 m depth, precise estimates may be obtained with as low as 100 stations for a majority of habitat forming taxa. Combined with estimated costs, this was roughly equivalent to a minimum of 75-95 person hours under optimal conditions. However, there is a risk of underestimation over a 6 year assessment period (in relation to the WFD) since not all issues of total uncertainty has been covered. Future studies need to address the uncertainty related to temporal dynamics. Integrating predictive mapping and monitoring is feasible, both regarding occurrence and abundance distribution. Although there are some conflicts regarding the benefits of stratification versus random sampling, there are suitable modelling techniques for these kinds of data available. It is important to note however, that if the aim is to obtain full coverage maps of both the qualitative distribution and quantitative abundance the effort would have to be increased compared to the minimum requirements outlined above.