ABSTRACT
     Hurricanes and other catastrophic wind events impact forests
over most of the globe.  I examine spatial patterns of damage and
the factors that influence these patterns, and the resulting
dynamics of vegetation recovery in a subtropical wet forest
following hurricane disturbance.  I develop an approach to
modeling disturbance dynamics in which any point in geographical
space can be represented as a position in a multiple gradient
space of abiotic factors; and disturbance and recovery are
quantified as movements within this gradient space.
     
     Spatial pattern analysis indicates a significant clumping of
hurricane damage.  Canonical correlation analysis is used to
compare the relative influence of abiotic factors to biotic
factors in determining the severity of damage.  Spatial patterns
of hurricane damage are most strongly correlated to biotic
factors, i.e. differential distribution of tree species.  I
quantify recovery in terms of vegetation community dynamics,
rates of biomass accrual, and restructuring of the canopy. These
recovery dynamics are predicted based on gradients of abiotic
environmental factors and severity of hurricane disturbance.  A
two-dimensional gradient space of damage severity, quantified as
structural damage (percent basal area lost) and compositional
damage (percent mortality), is effective in predicting the major
path to recovery: regeneration (sprouting of surviving trees)
versus recruitment (establishment of pioneer species).  A two-
dimensional gradient space, of simulated solar radiation and soil
moisture, allows prediction of rates of biomass accrual.  Post-
disturbance biomass increases are maximized under high levels of
solar radiation and moderate level of soil moisture and damage. 
The vertical restructuring of vegetation is associated with
community dynamics of recovery, as predicted by damage severity
and by topographic position.

     I integrate these responses with a spatially-explicit
computer model (RECOVER) which simulates patterns of damage and
recovery from a hurricane of a given intensity and track.   
Simulated patterns of disturbance and recovery indicate that the
ecosystem is more sensitive to changes in the frequency of
hurricanes than to changes in intensity of storms.  This model
can be used to investigate hypotheses about the dynamics of
hurricane disturbance and recovery.

Return to the Table of Contents