|
Atlantic and Gulf Rapid
Reef Assessment
|
Index | Overview | Methodology | Coral
standards | Fish standards | Calibration | Data
processing | Data sheets
Coral reef fishes
download
in Word format
 Estimating 2m wide belt
transect
Counting fish
Estimating length
Rover diver
For fish sampling, biases are often due to difficulties arising from seeing
fishes, accurately counting or estimating length, and differentiating among
certain similar species (Sale
1997). An effective way to census such mobile organisms as fish are through a
method that most imitates an instantaneous survey of a pre-defined area (Sale
1997). The AGRRA fish method includes two different methods to characterize
reef fish populations – a belt transect and rover diver method.
|
|
|
|
|
A 1 m t-bar is used to help visualize the 2 m wide belt
transect.
|
|
Each 2 m segment of belt transect is surveyed separately.
|
Estimating 2 m wide belt transect
The first challenge is to visualize the boundaries of a 2-m
wide X 30 m long belt transect for fish censuses. The sample unit has been
specifically delineated to a small 2-m wide width because it is a small enough
width for the observer to be able to estimate the distance easily while
allowing for a large enough sample area. A 1 m t-bar (with 5 cm markings) is
used to assist observers in estimating the distance (below left). To estimate
the 30 m length, try to visualize the transect as a 2-m wide square tunnel (from
the bottom to the surface of the water). Start the first box at least 1-2 m
ahead of t-bar. When sampling the transect belt, the observer should give
uniform attention to each successive 2-m segment (below right). This requires
swimming at a more or less constant rate, and looking consistently about 2 m
ahead, except when actually recording data
Top
of page
Counting fish
Uniform
attention should be given to each successive 2-m segment of the transect when
surveying. It is important to swim in a consistent manner while actually
sampling the fish, although it is permissible to pause while recording data,
and then to start swimming again. A speed that counts each 30-m transect in 6-8
minutes should be attempted. High densities of counted species will slow this
rate in some cases. By remembering to keep effort equivalent on all segments of
the transect, you can focus on counting just those members which happen to be
within the boundaries of the transect and limit the tendency to count all
members of a large school of fish on either side of the transect. In other
words, keep your eyes focused straight ahead of the transect and do no get
distracted side to side as a large school swims by. Large groups of individuals
of a species that occur within a 2-m segment will be classified by attempting
to put them into one or more size categories as necessary.
Estimating length
Fish
observers should be trained to estimate fish lengths by using consistency
training methods both on land and underwater (e.g., GBRMPA 1979, Bell et al.
1985, English et al. 1994). The AGRRA method assigns fish lengths to the
following size categories (0-5 cm, 6-10, 11-20, 21-30, 31-40, >40cm). A 1 m
T-bar with 5 cm increments is used to assist in estimating sizes. We use a
variety of fish models cut out of closed-cell foam into the different size
classes (plywood or plastic can also be used). The shape of the fish models
used represent some of the species included in the AGRRA survey.
Top of page
Rover diver
Fish surveyors should be proficient in identifying Caribbean
reef fishes. For minimizing and recognizing observer bias and differences with
the REEF rover diving technique, we suggest reading the guidance provided by
Schmitt et al. (1998). The “Reef fish identification (Florida Caribbean, Bahamas)”
book and cd-rom by Paul Humann are very useful for improving and testing your
fish identification.
|
Exercise
1: While snorkeling or diving, each observer should have a T-bar
marked with 5 cm increments and a datasheet (see below) and should be 2-m
from the fish models. One person should have a series of different sized fish
models (of known lengths). One person holds up one fish model at a time and
each observer estimates the size of each fish and assigns it to one of the
following size categories (0-5 cm, 6-10, 11-20, 21-30, 31-40, >40cm). The
holder should write down on a datasheet the order in which the fish models
that are shown and the corresponding actual lengths. After at least 10 fish
models have been estimated, the observers should rotate to allow each person
a chance to practice estimating length. After the exercise, the divers
compare their answers to the actual answers and determine any differences
from the actual known amount and the difference between observers. The
exercise should be repeated until individual observer estimates are > 95%
accurate and observers are consistent with each other.
Example datasheet
|
Size (cm)
|
0-5 cm
|
6-10
|
11-20 cm
|
21-30 cm
|
31-40 cm
|
>40 list
size
|
|
Fish 1
|
|
X
|
|
|
|
|
|
Fish 2
|
|
|
X
|
|
|
|
|
Fish 3
|
|
|
X
|
|
|
|
|
|
Exercise
2: In this exercise, observers practice fish counting along a belt
transect and estimating fish size. Each observer should have a T-bar, a 30-m transect
tape, and a datasheet. A practice transect line is laid on the substratum and
fish models (made from styrofoam and attached to a weighed line of varying
lengths) are haphazardly placed along the transect (see figure to right) with
some models placed just out side the 2 m wide belt transect. One observer
starts at the beginning of the transect line and uses a T-bar to estimate the
width of the 2-m belt transect and the length of fish. Observers should mark
on their data sheet the size class for each fish (model) observed in the
sample area. Each observer should run the survey. After completing the
exercise, compare the answers of all the observers with the correct fish
lengths. Repeat exercise until observers are consistency between each other
and their answers are close to the correct answers.
|
|
|
Exercise
3: While diving, two observers (one hovered above the other) survey a 30
m transect at the same time (with one laying out the transect). The data is
then compared between the two observers. This approach is especially
effective if one of the observers is already well trained and experienced.
|
Top
of page
Index
| Overview | Methodology
| Coral standards | Fish
standards | Calibration | Data processing | Data
sheets
© 2000-2006 - AGRRA
|
