Comparing fixed- and variable-base-age site equations having single versus multiple asymptotes
Cieszewski CJ
FOREST SCIENCE
48 (1): 7-23 FEB 2002
Abstract:
Site equations compute values of a variable Y as a function of both variable t and a value of the variable
Y = Y0 measured at an arbitrary t = t0. For example, the plant size Y can be defined
as a function of both age t and a reference size Y0 measured at the base age t0.
The base age can be implicit (i.e., implied but hidden), as in fixed-base-age equations [e.g., Y = f(t, S),
where S is Y at t = 50], or explicit (i.e., readily exposed and changeable), as in dynamic equations
[e.g., Y = f(t, t0, Y0]. Using as the main criterion the ability of an equation to
generate concurrent polymorphism and multiple asymptotes, I compare a fixed-base-age height growth site
equation with several dynamic equations, derived through the traditional and the Generalized Algebraic
Difference Approaches. The comparison leads to conclusions about desirable model properties, the methodologies
of derivations, and expected outcomes of the different methodologies. The conclusions suggest that the ability
to simulate concurrent polymorphism and multiple asymptotes is an important property of site equations that
should be considered during modeling various growth trends. Furthermore, the conclusions suggest that both
algebraic difference approaches are more parsimonious and robust than the fixed-base-age approaches.
The Generalized Algebraic Difference Approach can increase model usefulness considerably through derivation
of more complex equations that can achieve more desirable properties.
Author Keywords:
model derivation, model conditioning, nonlinear models, biological models, site productivity,
base-age-invariant equations, dynamic equations
KeyWords Plus:
INDEX CURVES, HEIGHT GROWTH, DOUGLAS-FIR, YIELD, MODEL
Addresses:
Cieszewski CJ, Univ Georgia, Sch Forest Resources, Athens, GA 30602 USA
Univ Georgia, Sch Forest Resources, Athens, GA 30602 USA
Publisher:
SOC AMER FORESTERS, BETHESDA
IDS Number:
521NH
ISSN:
0015-749X
To Request a Reprint: