Environmental Engineering Reference
In-Depth Information
To gain a better understanding of the spatial extent of disturbance (i.e. the
fraction of the plot that is disturbed) and the size of disturbed patches, two different
scenarios were created: First, the spatial extent of disturbance was varied at five
intervals ranging from 0 to 0.6 while the size of disturbed patches was kept
constant. In the second scenario, the total disturbed fraction of each plot per year
was kept constant and the size of disturbed patches was varied. The sizes of
disturbed patches followed a power law model in which frequency of disturbance
events is related to their spatial extent; i.e. smaller disturbances were set to occur
more often than larger ones.
Recruitment and background disturbances were updated yearly while growth
and competitive interactions were iterated once every 3 months.
Simulations were run for 500 years (complete cycle of the model) and percent-
age cover of each species was taken on an annual basis. The data derived from the
model simulations were fed into a Bray-Curtis matrix in order to determine the
sensitivity of the variables on species diversity, species composition and mortality
in response to each of the model parameters.
The results showed that in the absence of disturbance, the reef was occupied by
competitively dominant species and that those species featuring low aggression and
low growth rates were lost after short periods of time. Intermediate levels of
background disturbance favoured high coral diversity, which supports the classic
hypothesis of intermediate disturbance (Grime 1973). Accordingly, the amount of
bare substratum increased with higher levels of disturbance and was accompanied
by a decrease in biodiversity.
The relative importance of total colony mortality to partial colony mortality
changed with colony size class. Total mortality was more important for small
colonies while large colonies were most sensitive to partial mortality. For sensitive
species, competitiveness, i.e. aggressive potential had the greatest influence on
community composition. Growth was also an important factor whereas mortality
and recruitment had the least impact on the model.
Model evaluation showed that only five out of the ten simulated species were
comparable to actual field observations. Despite this relatively weak congruence
(mainly resulting from insufficient data), for coral population size structures,
the model was able to accurately represent growth and distribution for seven out
of the ten species. The results indicate that size structure of populations is a much
more precise indicator for testing the predictive abilities of the model than the
simple comparison of coral cover.
Another important aspect of the results was the apparent relationship between
the threshold of partial and total colony mortality and modal colony size on a log-
scale, which demonstrates that colony size and age may be decoupled earlier than
was previously thought (e.g. see also Bak and Meesters 1998). Also, the results
indicate that size of disturbed patches was as important in structuring coral
communities as the overall amount of disturbance (Fig.
17.3
). This poses an
important consideration when looking at recovery mechanisms of reefs since
the spatial extent of cleared substratum is rarely directly quantified in the field,
but is usually determined indirectly through differences in coral cover. Even though
