1 Introduction At the most microscale, individuals and households make decisions that affect land-use patterns, which, in turn, affect human behavior through positive and negative feedback mechanisms that have their own space ^ time dependencies. Land-use change, such as converting a natural savanna-forest into an unimproved pasture, has implications for land cover as the pasture is improved to accommodate additional animals or the land is degraded through use with bare soil and barren patches emerging over time (Walker, 2003). Although a global accounting of landownership or land use is not available, nevertheless, it seems reasonable to assert that a substantial proportion of the globe's land surface is owned or influenced by individuals or households (as opposed to governments, corporations, or other large organized social entities) through direct or indirect forces and consequences. Individuals and households make decisions on the use of millions of parcels of land on the basis of their interpretation of what might be best for them within the context of legal, economic, technological, biophysical, temporal, and other operating constraints. At the other extreme, increasingly, there is a global market for agricultural and manufactured products, and this market can and does affect local land-use decisions through local to global effects that are manifested Dynamic spatial simulation modeling of the population ^ environment matrix in the Ecuadorian Amazon Joseph P Messinaô Department of Geography and Center for Global Change and Earth Observations, Michigan State University, East Lansing, MI 48824-1115, USA; e-mail: jpm@msu.edu Stephen J Walsh Department of Geography and Carolina Population Center, University of North Carolina, Chapel Hill, NC 27599-3220, USA; e-mail: swalsh@email.unc.edu Received 10 September 2003; in revised form 22 January 2005 Environment and Planning B: Planning and Design 2005, volume 32, pages 835 ^ 856 Abstract. This research uses multithematic and spatially explicit data combined from a longitudinal socioeconomic and demographic survey conducted in 1990 and 1999, GIS coverages of resource endowments and geographic accessibility, and a classified Landsat Thematic Mapper (TM) satellite time series. The goal was to combine such data with expert knowledge, a set of analytic results, and dynamic modeling approaches to describe, explain, and explore the causes and consequences of land use and cover change (LUCC) in the northern Oriente region of the Ecuadorian Amazon. First, a cellular automaton (CA) model representing LUCC was developed using a time series of remotely sensed Landsat TM images for a 90 000-ha intensive study area within the region and calibrated using alternative images from the time series. The classified images were linked to spatially referenced biophysical and socioeconomic coverages used as input data, and then combined with `rules' derived from empirical analyses. Second, the CA model was used in dynamic simulations to explore LUCC as both causes and consequences of (a) road development, (b) agricultural extensification and land abandonment, (c) major shifts in world markets and crop prices, and (d) urban expansion of the central city within the region. Finally, complexity theory was explored within the spatial and temporal dynamics associated with population ^ environment interactions, particularly, deforestation, urbaniza- tion, and subsistence and commercial cultivation of agricultural crops on lands made accessible by petroleum-company-built roads and the corresponding in-migration of spontaneous colonists begin- ning in the late 1960s. This research contributes to the study of population ^ environment interactions in a frontier environment, and examines how dynamic and complex systems can be modeled using CA-based spatial simulations. DOI:10.1068/b31186 ô Corresponding author.