![[logo] US EPA](http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Validate Decision Processes With Additional Data Sets
- Classify coral reefs
- Develop testable hypotheses
- Define data collection protocols
- Biological sampling
- Screen attributes to define metrics
- Determine appropriate sampling effort
- Define data protocols
- Validate decision processes
- Define biocriteria
- Implement monitoring programs
- Diagnose causes of degradation
- Evaluate management effectiveness
- Communicate results
Biological metrics must be validated with an independent data set before being adopted into biocriteria. Associations between metrics and human disturbance should be confirmed for more than one location.
Ensure that classification is at the right level (not too few or too many).
The number of categories needed to classify the various habitat types should be kept to a minimum where possible. Reef areas may look different, for example, if they include different species or are shaped differently, but if expectations for coral reef metrics are similar, then they should be combined to simplify survey sampling.
For large regional areas, the effort required to individually sample small unique habitat types may not be merited. Time may be better spent sampling the dominant habitat types that represent the greatest percentage of the resource.
Selection of metrics should match regional human influences.
Global measures of coral decline are compelling indicators, but for local monitoring and assessment, coral metrics need to indicate the types of human disturbance that local managers can regulate. Loss of atmospheric ozone may be implicated in coral loss, but managers in the US Virgin Islands can do little to influence this stressor. In contrast, land use practices associated with agricultural erosion or nutrient enrichment from wastewater can be regulated by local authorities. Coral indicators may respond similarly to both global and local disturbance or some indicators may be uniquely associated with one or the other.
Select metrics that respond predictably to known human influences for an independent data set.
For the US Virgin Islands, St. Croix, St. Thomas, and St. John provide three replicate islands for testing whether observed differences in coral metric values are due to differences in human disturbance or better explained by underlying natural processes or gradients.