Simulating of Land Use Dynamics in Southeast Asia:
A Cellular Automaton Approach 2001 - 2005
Background
The spatial pattern of village territories in Nang Rong district is
complex. Behind the spatial pattern are complex spatial
processes. Human decision-making takes place on a variety of
spatial and temporal scales, and models are needed that integrate human
dimensions at these space-time scales.
Purpose
This project, using a temporally, substantively rich case study, is
developing cellular automaton (CA) procedures to predict a spatially
explicit model-based simulation of future land use and land cover
change scenarios for Nang Rong, Thailand, and the broader region
surrounding Nang Rong, including Cambodia, China, and Vietnam.
Scope of Work
The scenarios are based on empirically observed relationships in the following areas: a) history and spatial pattern of village settlement; b) road development, expansion of available vehicles, and changing geographic accessibility; c) electrification and the
accompanying rise in TV viewership and consumerism; d) outmigration and urbanization, including the role of foreign investment in the broader region; e) global economic factors, including world cassava prices and the 1997 economic crisis; f) climate and monsoon history; g) spatial and temporal land
titling and linkages to investment in various land uses; and h) basic demographic change, with an emphasis on changes in the number of households and their size/composition. Results of
the simulations will elucidate the spatial distribution and composition
of LULCC.
The project exploits a rich collection of interlinked data sets for
Nang Rong. A collection of previously analyzed Landsat images (TM
and MSS) dates to 1973. Other remotely sensed data include AVHRR,
SPOT, and SAR, as well as aerial photos dating to the 1950s.
Community and household level surveys are available for 1984, 1994, and
2000. Out-migrants have been followed, and in-migrants added to
the dataset. Digital coverages showing roads, rivers, elevation,
soil types and other spatial-thematic data are available within our
GIS. Daily precipitation and temperature data can be linked at
the village level for 1984, 1994, and 2000, and at the household level
for approximately 10,000 households in 2000.
Analytical Methods
The research draws heavily on recent work in remote sensing,
demography, sociology, complexity theory, and related social and
biophysical disciplines. Both CA and agent-based models are used,
and qualitative approaches back up and validate the models. A CA
model representing LULCC is being developed and validated using a time
series of remotely sensed satellite and aircraft images from Northeast
Thailand linked to spatially referenced biophysical and socioeconomic
coverages as input data combined with "rules" derived from empirical
analyses of those data. The CA model will be used in dynamic
simulations to explore LULCC as both cause and consequence. After
developing, calibrating, and validating the CA modeling scenarios for
Nang Rong through the use of a deep satellite time series, spatially
explicit LULCC patterns will be derived for the period 1950 - 2020.
Contribution to the Field
The spatial extent of the CA modeling and scenario evaluation will be
expanded regionally by comparing Northeast Thailand to other sites in
Northeast Cambodia, Southern China (Yunnan Province), and Northern
Vietnam. Using relationships, rules, and weights from these test
sites, relative to Nang Rong scenarios, investigators simulate LULCC
dynamics by perturbing CA-based Nang Rong simulations for conditions
that represent alternate development processes that have occurred
elsewhere within the Southeast Asia region. These are countries
with significant extant forest coverage, some of which has likely been
preserved due to difficult political and social histories in the past
50 years. While prediction is difficult, it seems that at least a
sub-set of these countries is poised for substantial social and
economic change, with resulting implications for LULCC and the carbon
cycle. See
publications for examples of work emanating from this
project.
Funding
This project is supported by a grant from the National Aeronautics and Space Administration.