Sampling technique

Up until the 2006 field season a video camera was used to film a 50cm wide swathe for benthic transects. Since the beginning of the field season in 2007 the LTMP has used digital still cameras. Fifty photographs are taken at 1m intervals along each 50　m transect using a digital still camera held approximately 40 cm above the substrate. Percent cover of corals and other benthic categories are estimated using a point sample: approximately 200 systematically-dispersed points are sampled from each photo transect.　

Christie and Mapstone (unpublished) used different sampling strategies to investigate whether or not percent cover estimates of benthos differed significantly. They found that percent cover estimates did not differ significantly when the face-centered cube (5 points per screen) was compared to single random points. Therefore using random points or fixed points to estimate the percent cover still represents an efficient sampling strategy for those using basic equipment configurations. This study also established that 750 random points per transect represents the "best" possible sampling intensity for transects filmed at this focal distance. A comparison of mean and precision estimates for fewer than 750 points indicated that at 200-250 points there was no significant difference for any benthic group or lifeform (3). An example of power calculations for hard coral at a sampling intensity of 200 points per transect follows; In 2000 Agincourt Reef had moderate coral cover at 37%. The minimum detectable rate of change for this reef is 10%. The power to detect a trend occurring over the six years of survey is greater 5% (4).

Sample points for the benthic transects are typically identified to genus level for hard and soft coral. Algae and other organisms are placed into functional groups. A comparison of field identification and identical points on the video footage found that 'Benthic groups’ (mean accuracy of 90% ( 8)) and ‘families of hard coral’ (91% ( 7)) were identified most accurately and least variably from video images. Life-forms of hard corals proved to be the least accurate and most variable level of classification with a mean accuracy rating of 74% ( 16) (4). Observer error was small relative to estimates of cover. For example, if hard coral cover is 30%, the mean confidence interval due to both between- and within-observer variability was 27.2% to 32.8% (4).　

Fishes from a list of 191 species, representing 10 families, are counted on the five 50 m transects at three sites on each reef. All species are largely non-cryptic, easily　 identified underwater and include both commercial and non-commercial taxa. Because the surveys span the annual recruitment season, 0+ individuals are excluded from counts. These are distinguished from adults by their small size and often distinctive colouration. Large mobile fishes and damselfishes are counted separately on transects 5m and 1m wide, respectively. The general survey procedure at each site involves an experienced observer swimming along the central transect line counting large mobile fishes to 2.5m either side of the mid line. Absolute numbers of target fish species are recorded using a pencil and underwater paper (attached to a slate). An assistant swims approximately 10m behind the observer laying a tape measure over the substratum along the centre line of the transect. On completion of the five transects in which large mobile fishes have been counted and the length of predatory fish has been estimated, the observer returns along the same transects (which are now marked with a tape along the centre line) and undertakes counts of smaller damselfishes (Pomacentridae).

Observer swimming speeds average 10 metres per minute. In order to reduce observer bias, at the end of every transect the observer identifies an object estimated to lie at the outer edge of the transect. The perpendicular distance between this object and the transect centre line is then measured, thus providing the observer with a frequent reference to the desired transect boundaries. To further limit observer bias, throughout the sampling season regular pair wise comparisons are made between all observers used in this program. Once field sampling is complete all data is entered onto a customised database. This data is then checked against the original data sheets to ensure no transcription errors have occurred. Abundance data for each fish species is subsequently summed over the five transects at each site on each reef to provide reasonable sample sizes for analysis and interpretation.

Counts of juvenile hard corals

Juvenile hard corals that are less than 5cm diameter are counted along a belt transect (length of dive slate (34cm)) for the first 5m of the transect. These counts have been conducted since the beginning of the field season in 2007. Within the water quality program, small soft and hard corals have been recorded for three categories for inshore reef monitoring since 2004.