Inter
MODELING PATCH DYNAMICS FROM INTEGRATION OF CELLULAR AUTOMATA $
SIMULATION AND HISTORICAL AIR-PHOTOGRAPHS ANALYSIS Dual
com]
chan
M. Shoshany and E. Kelman char:
fragr
Department of Geo-Information Engineering, Faculty of Civil & Environmental Engineering, TECHNION , Israel patei
Institute of Technology, Haifa 32000, Israel px
Commission IV, WG IV/6 1956
KEY WORDS: Geo-Information, Landscape Change, Patch-Dynamics, Patch Pattern, Fragmentation
ABSTRACT:
Dynamics in nature are multi-faceted and difficult to follow. Computerized simulations offer new ways for investigating dynamic systems J
by incorporating their principle aspects . These simulations might be instrumental in discovering fundamental properties of dynamic ¢
systems such as duality between fragmentation and shape of changing patch patterns. Do fragmentation processes which occur over a
range of scales, from planets to microecosystems, share a common geometrical evolution? Using remote sensing data and computerized
simulation, we have formulated and assessed here the duality between geometrical properties of patch patterns. Fragmentation and shape,
were found to be mutual and complementary in the way that they change, facilitating reconstruction of patch patterns from partial data over
a wide range of natural phenomena.
i: Ln: : : ; 1976
1. INTRODUCTION connectivity within fragmentation analysis can be achieved by ie
adding transition zones or buffers to the delineated patches. In
Fragmentation processes occur over a wide range of scales ecological terms, increasing fragmentation represents
attracting attention in diverse fields of research: planets (Boss, intensification of disturbance, disconnection and decrease in
2000; Mayer et. Al, 2002) forests (Cochrane, 2003), habitat; while patch expansion and aggregation indicates
microecosystems (Gonzalez et.al, 1998) and recovery (Allan et al, 2003; Cochrane, 2003; Hanski &
microscopicstructures (Ilani et al, 2001). Quantitative Ovaskainen, 2000) .
measurements of the spatial attributes of fragmentation are Here the parameterization of fragmentation is based on the well
mostly incomplete, allowing very limited discrete description of known Shannon and Weiner13 Information Index, :
their course of action. While some of these processes are A
phenomenologically specific, others share common geometrical — SW = - > Si In Si (1) e
characteristics (i.e., self similarity (Mandelbrot, 1982)), which
may facilitate closing the information gaps. Mutuality and ; ; ; d
; u «s d ia where Si represents the relative area of a patch i out of the total
complementarity between fragmentation and shape in patch
patterns was observed in an earlier work (Shoshany, 2002).. area of patches, and the summation is for all patches.
Here we generalized, formulated and tested a fragmentation — Shape is parameterized mainly according to relationships
shape duality hypothesis, using both real data and simulation. between perimeter length and area (Forman, 1995). Such Ta
Its practical implications lie in reconstruction or prediction of parameterization facilitates differentiation in the elongation and
spatial behaviour from partial data. The simulation emulates convolution of patches, which in turn determine the flow of 1990
processes combining random patch generation with a non- energy and matter in and out of patches. Increasing the relative [m
random element (Hanski & Ovaskainen, 2000) represented by boundary length may increase the variability of habitat
the adhesiveness of new patches to the existing ones. conditions, which may in turn lead to an increase in species
richness and vice versa. Within the context of a pattern of
patches rather than an individual patch, the shape is measured
using WAP1 and WAP2 weighted area-perimeter indices
2. PATTERN PARAMETERIZATION (Santiago & Javier, 2000) in the form:
Landscape ecology terminology was adopted here to allow WAPn= XSi*Ai/ P", Q) 5
intuitive interpretation within the context of issues of wide "n
^nvir © ^ o : : : AI Ee
on Sn, ire tom yrinnllion 10 Seas where Ai and Pi are the area and perimeter of a patch indexed i; ze.
Fraementation represents (Form: the "breakdown of WAPI and WAP2 are closely related, with the first representing
gmentation represents (Forman, 1995) the eakdc :
; ; ac? ; 4 i ; the area per unit boundary length, and the second a
objects into pieces”. It is parameterized mainly by : ; :
characteristics of the relative size distribution of distinctive dimensionless index .
(unconnected) patches. Fragmentation is related to the level of
connectivity between areas, which needs to be addressed in >
respect to modes of dispersion and migration of different FA
appe
species (Andren, 1994; Young & Jarvis, 2001) . Incorporating
560
