Na. Lagos et Jc. Castilla, BAYESIAN STATISTICAL-INFERENCE IN ECOLOGY - AN EXAMPLE OF ITS USE IN A CHILEAN MARINE PRESERVE, REV CHIL HN, 70(4), 1997, pp. 565-575
Bayesian inference allows setting hypotheses under uncertainty, for wh
ich requires explicit assignment of prior probabilities for a paramete
r, quantities or situation. This a priori probabilities, together with
observed data described by a likelihood function and using Bayes' the
orem, allows to obtain an a posteriori distribution which updates the
initial uncertainty given the new evidence. In this study, the potenti
al of bayesian inference in ecology is illustrated using available inf
ormation about the effect of spatial heterogeneity and human exclusion
upon two intertidal key-hole limpet populations of a single non repli
cated Marine Preserve and a harvested site at central Chile. Bayesian
inference has a direct interpretation about the parameter or uncertain
value, is operative in non replicated studies; and furthermore, its f
ormal structure provides a potential tool allowing the improvement of
ecological theory. This can then be applied to decisionmaking processe
s and the management of natural systems.