Biological Sampling

Biological sampling should represent the entire gradient from minimal human influence to severely altered.

In the early years of developing a biocriteria program, one efficient way to test data collection and data analysis protocols involves sampling across a gradient of human influence. Sampling across the spectrum from reference to degraded condition provides a framework for testing the sensitivity of candidate biological metrics, identifying the level of change that indicates degradation, and determining the level of sampling effort needed to detect that level of change.

Gradients need not be strictly numeric, but can be narrative or descriptive. In the US Virgin Islands, an industrial point source along the shore provides an opportunity to test and evaluate candidate coral metrics.

Figure 1

Figure 1: Schematic representation of a sampling design for testing coral metrics across a gradient of human disturbance.

Pictured is an area of disturbance associated with an industrial point source on land. Reef stations are arrayed within a single coral reef habitat type that parallels the shore. Sampling locations extend from the most disturbed area and beyond the zone of influence. Replicate samples and samples from a different habitat type are also pictured.

Data should be collected from the full set of habitat types.

Where possible, different habitat types should be tested for their sensitivity to human influence. In the US Virgin Islands, major habitat types within each coastal management zone (e.g., patch reefs, linear reef, etc.) will be sampled. In the example above, at least two habitat types will be sampled across the human disturbance gradient. If coral metric response is similar in the two habitat types, they can be combined and sampled as a single reef area. If coral metrics are more reliably associated with the disturbance gradient in one habitat than the other, the less responsive habitat may be excluded from future monitoring. If coral metrics show a strong association with the disturbance gradient but different ranges of values, expectations for coral metrics can be adjusted according to habitat type.

Enough data should be collected to evaluate sources of variability.

Repeat sampling within a reef area provides the data we need to estimate the relative contribution of different sources of variance (e.g., microhabitat differences or measurement error associated with different crews) to the overall variability of the coral reef metrics.

Components of variance analysis uses output from ANOVA testing to calculate variance associated with different factors. A study in the Florida Keys collected replicate data in coral reefs from 1) different halves (sections) of transects, 2) different dive crews, 3) different years, and 4) different stations within the sampling region.

Figure 2

Figure 2: Relative contribution to the overall variance of coral metrics due to differences associated with sections of the transect, duplicate samples of the same reef station, repeat visits in different years, and station differences. Shown are total reef surface area, % living corals, living reef surface area, and % living surface

For the 4 coral metrics picture above, microhabitat differences associated with different halves of the sampling transect were small and suggested that a smaller sampling area could be used. Both total reef surface area (TSA) and living surface area (LSA) had variance profiles indicating good potential as biological indicators, that is, only a small percentage of their total variance was associated with duplicate sampling by different dive crews. This result indicates that the sampling protocol and metrics are reliable and repeatable. In contrast, a large percentage of the total variance was associated with different reef stations. This result indicates a good potential ability for these metrics to detect differences in reef locations.

In contrast, percent living coral (%LC) and percent living surface area (%LSA) may need a greater sampling effort to detect a similar amount of change because of the high percentage of variance associated with duplicate sampling by different dive crews. Thus, either more duplicate samples or a larger reef area may need to be sampled for these metrics.