Standardization and calibration of AGRRA visual assessments

Atlantic and Gulf Rapid Reef Assessment

Standardization & Calibration of AGRRA visual assessments

Corals
        - Colony boundaries
        - Coral identification
        - Colony size
        - Estimating coral mortality
        - Coral condition
Algae

The limited ability to compare among methods is the strongest argument to adopt common sampling protocols (Sale 1997). One of the main objectives of the AGRRA approach is to provide a standardized methodology that will enable teams working in different areas to collect and compare data on a regional scale. With visual censuses, it is important to have consistent and comparable visual estimates and to minimize individual bias among different observers. Hence, it is essential to carefully standardize methods prior to data collection. This requires “consistency training”, particularly in learning to correctly identify coral and fish species, distinguish recent versus old coral mortality, estimate percentages of coral mortality and algal abundance, and correctly estimate fish numbers and sizes. The following highlights some of the differences we have found among observer and several exercises we have used to train observers on the AGRRA method, build consistency between observer estimates, and minimize differences in visual estimates. Other recommendations for maintaining data quality and consistency can be found in the literature (e.g., Bell et al. 1985, English et al. 1994).

Corals

For assessing coral condition, differences in observer estimates can arise from differences in:

· Delineating the boundaries of an individual coral colony,

· Coral nomenclature,

· Measuring colony size,

· Estimating the percent coral mortality and distinguishing between old and recent mortality, and

· Determining potential cause(s) of recent mortality.

Top of page

Colony boundaries

Before assessing individual corals for identification, size or condition, it is important to distinguish what constitutes a colony, i.e., what are the colonies boundaries? For most cases, colonies of mound corals (e.g., Diploria, Siderastrea) are fairly easy to distinguish, whereas recognizing distinct colonies of branching corals (e.g., Acropora palmata and A. cervicornis) or platy variety of corals (e.g., M. franksi) can be very difficult. When determining colony boundaries you also set the boundaries for maximum height and diameter. It is important that observers are consistent in distinguishing individual colonies prior to collecting data.

Here are some guidelines to use to establish colony boundaries:

1) A colony is defined as any autonomous, free-standing coral skeleton that is still identifiable to genus level (preferably to species level) based on the presence of living tissue or identifiable corallites.

2) Assess only colonies still attached to the substrate. For corals that have been knocked over, only survey it if it has either reattached to the substrate or is too large to move.

3) Include colonies that are 100% dead that you can identify to generic level based on colony morphology (e.g., Acropora palmata) or corallite characters (Diploria spp., M. cavernosa).

4) Identify the colony’s boundaries based on connecting or common skeleton, connecting living tissues, polyp size, and polyp color.

5) The living tissues of some corals may be distributed among several physiologically separate units. Species like the columnar Montastraea annularis (= M. annularis f. annularis) which grow as clusters of basally-interconnected lobes having live tissues only at their summits, should be treated as one coral, and observed only at the tops of the lobes.

6) Look for one vs. two colonies. Sometimes one colony of M. faveolata can encircle a M. annularis colony in the middle and superficially look like a single colony (like an apron on a person).

7) Corals like Acropora cervicornis which grow as thickets should also be considered as one “supercoral”, unless you are sure of the boundaries of the constituent colonies.

8) If two colonies are growing on top of one another and they both fall under the transect line, assess each colony individually. For example, this may occur when you have a mound coral growing underneath a branch of A. palmata.

9) If a colony has incurred partial mortality, be sure to include this when establishing the colony’s boundary.

10) Include all extended plates in plating varieties.

Coral identification

Once a colony’s boundaries are established, it is easier to make a correct identification of genus and species. To avoid differences or ambiguities of common names, while still recognizing that several different systems of nomenclature are in current use by reef researchers, a species list of the major reef building corals of the Caribbean is given in Appendix 2. This list represents the Caribbean species that are the primary focus of our analysis i.e., those that contribute most to reef building. To distinguish between colonies, it is useful to use colony shape (e.g., branching, mound); corallite size (e.g., compare Stephanocoenia intersepta, M. annularis, and Madracis decactis), and corallite shape (i.e., Dichocoenia stokesii and Favia fragum). Particular attention should be given to correctly identifying corals with similar appearances (e.g., Diploria strigosa, D. clivosa) as well as identifying different morphologies of the same coral species (e.g., small encrusting and large mound forms of S. intersepta). Observers should be proficient in identifying coral species prior to conducting AGRRA surveys.

Top of page

Colony size

After you identify the colony and its boundaries, measure to the nearest cm the coral’s maximum projected diameter (live + dead areas) in plan view and maximum height (live + dead areas). Use these guidelines:

1) Establish the coral’s axis of growth, especially for corals growing at an angle to the seafloor or on an inclined or tilted surface (see diagrram below right).

2) The diameter should be measured in plan view perpendicular to the axis of growth. Plan view is assessed from an angle that is parallel to the axis of growth.

3) The height should be measured parallel to the axis of growth.

4) Measure height from the base of the colony’s substratum, not the reef substratum or ocean floor.

5) For a colony growing on top of another colony, only measure the diameter and height of the individual colony being assessed.

For a colony that has been knocked over and too large to move (see below) or reattached, and does not yet show redirected growth, then measure size along original growth axis. But if a colony has started to grow in a new direction, be sure to measure diameter and height along the new axis of growth (not shown).

Example of size measurements for knocked over coral too large to move. Note dead area and live area. Mortality measurements would be take from planar view from the side (eye and arrow).

Example of size measurements for large coral (left) and small coral (right). Note plan view is tilted for small coral because it’s substrate is at an incline (eye and arrow).

Examples of corals with complex morphologies

·         Basally-connected columns of Montastraea annularis

·         Interlocking thickets of A. palmata, A. cervicornis, or Porites porites

·         Densely-spaced extensive colonies of Madracis mirabilis or M. formosa

·         Platy or shingle varieties of M. faveolata or M. franksi, especially if they form an apron around another coral like M. faveolata or M. cavernosa

·         Large mounds or heads of Agaricia tenuifolia, Millepora complanata

Above figure, in planar view, of basally-connected
(gray area) columns of Montastraea annularis with
only tops alive (hatched areas).

Top of page

Estimating coral mortality

The most common difference we found among observers is in estimating the percentage of coral mortality. To minimize observer bias, follow these guidelines:

1) Quantify mortality by visually estimating the amount from above in “plan” or “map” view, and ignoring all tissues on the sides or beneath upward facing surfaces (e.g., Acropora).

2) Plan view is assessed from an angle that is parallel to the axis of growth.

3) Mortality estimates for A. cervicornis, Madracis mirabilis, and all other branching corals that ordinarily are dead at the branch bases, should be restricted to the peripheral parts of the colony or thicket.

4) AGRRA makes a further distinction of mortality into partial mortality that includes both “recent” dead and “old” dead.

5) “Recently dead" is defined as any non-living parts of the coral in which the corallite structures are white and either still intact or covered over by a layer of fine mud. For recent mortality, there are several “stages” of recent that are all considered in the recent category:

            Very recent = white intact skeleton is still visible (dead w/in 1 month or less)
                Recent = corallite may be covered by thin turf algae or sediment (up to 6 months)
             Older recent = corallite structure slightly eroded or covered but still identifiable to genera (< 1-2 yrs)
                (except see 7 below for exception).

6) In contrast, "old dead" is defined as any non-living parts of the coral in which the corallite structures are either gone or covered over by organisms that are not easily removed (e.g., certain algae and invertebrates).

Corals that have been overgrown by the brown encrusting sponge, Cliona, but you can still identify corallite structure beneath the sponge, should be considered old dead (noticeable in Diploria spp., M. annularis, M. cavernosa, and Colpophyllia natans).

Top of page

Coral Condition

Determining potential cause(s) of recent mortality is often very difficult. To standardize assessments, at least until a more common consensus on disease nomenclature is reached among reef scientists, we suggest using color categories when identifying diseased colonies: Black (BB), Red (RB), Yellow (YB), White (WB= white band, WP=white plague, WS=white spots/patches), or Unknown (UK). To become more familiar with coral diseases, we recommend using disease cards (Bruckner 1998) and visiting the disease website http://ourworld.compuserve.com/homepages/mccarty_and_peters/coraldis.htm. If you are unsure of disease it is better to put unknown and describe it and take a photo if possible. Remember we are only interested in large scale bleaching events due to elevated sea surface temperatures, not localized patches of bleaching due to algal overgrowth. In some cases, corals may be pale or partly bleached (or remnant bleaching) due to the previous year’s bleaching event and are often still recovering. It is important to differentiate between corals with fish bites, recovering tissue from fish bites or bleaching and bleached tissue.

Exercise 1: This exercise is to help you become familiar with partial coral mortality. Appendix 5 shows a diagrrammatic representation of small patches of partial mortality on a “coral” and how tissue loss can expand over the entire colony. These figures also show the actual percentages corresponding to the amount of illustrated partial mortality. Review these diagrrams to become familiar with patterns of mortality and visually estimating the amount of mortality. Observers should also review pictures of recent and old mortality and be able to distinguish between the two.

Exercise 2: An additional exercise to practice estimating mortality is to use “flash cards” of corals with different degrees of partial mortality ranging from low to high. To do this, take photographs of a coral (in plan view) then determine the amount of partial mortality through digital analysis. Make a series of pictures with the amount of mortality written on the back of each photograph. Then look at each coral in the series of flash cards and determine the amount of partial mortality and identify the coral species. Compare your estimates with the digitally determined estimates. The resolution of the photographs usually only allows you to estimate total partial mortality and not distinguish between recent and old. Each observer should use the flash cards until his/her answer is close to the known value (+- 5%). Teams of observers should practice until differences in answers between observers is less than 10%.

Exercise 3: A way to standardize visual estimates between observers is to do in the field consistency training. Snorkeling exercises are especially effective because they allow observers to share their observations and questions. For this exercise, first lay out a 10 m transect, then (with another observer) swim the transect line and estimate the amount of coral mortality for each coral under the line, writing down your estimates on a datasheet. As each coral is estimated, compare your answers with your partner and discuss any differences in estimates. After completing the transect, all observers should discuss difference among their estimates. A similar exercise is also done while diving. For this, a series of set transect lines are laid out and each observer surveys the transects for coral condition. Afterwards, the estimates among observers are compared. Variation in data estimates should be discussed and additional exercises should be done until differences are minimal. Consistency training should be conducted fairly frequently, especially if an observer has not done visual estimates in awhile.

Top of page

Algae

For algal abundances, we are estimating absolute abundances of crustose coralline, fleshy macroalgae and calcareous macroalgae (we are not including estimates of turf algae). Normally we found differences in algal estimations to be fairly minimal after practicing the following exercises. The definition we are using for macroalgae is anything you pick up with your fingers. Crustose coralline algae may be present beneath turf algae or thin layers of sediment. Include any coralline algae in your estimate that you can clearly see without removing sediment, turf algae, or macroalgae.

Exercise 1: While snorkeling or diving, place a series of set algal quadrats along the 10 m transect line. For each algal quadrat, estimate the abundance of the crustose corallines, fleshy macroalgae and calcareous macroalgae. After each quadrat assessed, compare the answers among observers and discuss reasons for differences in estimates. Repeat the exercise with a different set of quadrats until there is a high consistency among observers.

Exercise 2: While in the field, survey several algal quadrats (in plan view) with varying degrees of algal abundance then visually estimate the absolute abundances. For each quadrat take a photograph in plan view and scan it into a computer. Use a digitizing program to analyze algal abundances and compare to your estimates. Continue with this type of exercise to increase your precision.

Top of page

Robert N. Ginsburg
Atlantic and Gulf Rapid Reef Assessment
MGG-RSMAS, University of Miami
4600 Rickenbacker Causeway
Miami, FL 33149
USA Telephone: (305) 421-4664
Email: info@agrra.org
Send data to: data@agrra.org
URL: http://www.agrra.org