|
|
Instruction, Early Research, and Photos
|
|
|
Each semester, Prof. Pereira teaches four of the following classes at SJSU: Remote Sensing I (lecture and lab) Remote Sensing II (lecture and lab) Geography of the Physical Environment Topics in Physical Geography Biogeography Seminar: Geographic Information Technology Seminar: Advanced Geographic Techniques Seminar: Research Design Senior Seminar Geographical Field Studies Global Geography East and Southeast Asia Introduction to Mapping and GIS (lecture and lab) Mapping the World (Introduction to cartography and GIS)
Course information website
Course descriptions, syllabi, study guides, and selected course material are available through my SJSU faculty website.
Syllabus, Geography of the physical environment
Syllabus, Remote Sensing I
Syllabus, Remote Sensing II
Syllabus, Field Studies
Syllabus, Senior Seminar
Syllabus, Geographic Information Technology
Syllabus, Research Design
|
|
Flammability simulation in a fragmented region of the Amazon
|
Investigations of fire dynamics in seasonally dry regions of the Amazon (PhD Dissertation, University of Minnesota, 2002) Image above is taken from Chapter 5 (Applying the simulation model to remotely sensed landscapes). (a) the wetness component of a Kauth-Thomas transformation of Landsat TM data. (b) a binary flammability image with potentially flammable regions in white. (c) The flammability image derived by processing image b with an edge effect filter described in previous chapters. (d) a fire spread simulation, generating by igniting the central cell of image c.
Table of Contents
Chapter 1
Chapter 2
Chapter 3
Chapter 4
Chapter 5
Chapter 6
Chapter 7
Chapter 8
|
|
Near Hailar, Inner Mongolia
|
|
|
On the grasslands of Inner Mongolia
|
|
|
Fragrant Hills, north of Beijing
|
|
|
somewhere in Sichuan Province
|
|
|
A spot for contemplation near home
|
|
|
|
Theoretical Geography A typology of relations between any two geographical scales is established by qualitatively comparing their respective grains and extents. This typology is applied to spatial, temporal, and spatiotemporal scales. It describes seven relations between any two scales in either space or time. These basic relations yield a set of 169 qualitatively different spatiotemporal scale relations, a subset of which is portrayed diagrammatically. If it is possible to transform processes or patterns from one scale in the relation to the other, up to four scaling methods may need to be simultaneously applied, depending on the relation: grain generalization, grain decomposition, extent extrapolation, or extent selection.
A Typology of Temporal and Spatial Scale Relations
Geographical Analysis 34, 2002.
|
|
|
Remote Sensing, Geocomputation, and Geosimulation Links
http://www.python.org/
Python for beginners
http://www.esri.com/
ArcGIS and other tools from ESRI
http://3dnature.com/
Terrain visualization from 3D Nature
http://eros.usgs.gov/
Earth Resources Observation and Science (EROS), a source for a variety of aerial, satellite, map, DEM, and land cover data.
http://www.mrlc.gov/
Multi-Resolution Land Characteristics Consortium (MRLC): the source for downloading National Land Cover Database 2001 (NLCD 2001) and 1992 data.
http://glovis.usgs.gov/
USGS Global Visualization Viewer: a source for Landsat, MODIS, and other data
http://www.erdas.com/
IMAGINE and other software from Leica Geosystems
http://www.ittvis.com/
ITT Visual Information Systems (ENVI and IDL)
http://www.dfanning.com/
David Fanning's guide to the Interactive Data Language (IDL).
http://www.clarklabs.org/
Clark Labs, home of IDRISI GIS
http://ccl.northwestern.edu/netlogo/
NetLogo, a cross-platform multi-agent programmable modeling environment.
|
|
|
Graphical user interface of dynamic data-driven multi-agent system. Time series sensor readings are at left. A DEM forms the landscape at center, with sensor values superimposed. Agent-based model settings are at right.
|
Agent-based environmental monitoring and modeling Networks of sensors and simulation models of the physical environment have been implemented separately, often using agent-based methodologies. Some work has been done in providing real-time integration of sensors and models. Given the distributed nature of geospatial processes, the use of software agents for representation, communication, and analysis seems to be a reasonable basis for such integration. An agent-based system that senses the actual presence of mobile organisms in a spatial environment and simultaneously represents simulated organisms individually in a model of that environment is described here.
Dynamic Data Driven Multi-Agent Simulation
Proceedings, The 2007 IEEE International Conference on Intelligent Agent Technology, Fremont, CA
Agent-Based Integration of Sensor Networks, Remote Sensing, and Simulation Models
Proceedings, 5th International Symposium on Digital Earth (ISDE5), Berkeley, CA, June, 2007.
Dynamic Date driven Multi-Agent Simulation
Long version for journal publication
|
|
3D view from the dynamic data-driven multi-agent system
|
Each animal is an individual and may have its own unique charateristics and behaviors.
Digital Earth Presentation
Integrating Microsensing Technologies with Agent-Based Simulations for the Study of Individualized Diversity in Ecological Communities
|
|
Landsat TM band 6, 6/16/06, San Francisco Bay Area
|
Thermal remote sensing Thermal remote sensing has the capability of characterizing local and regional urban heat island effects and other microclimatic conditions as they change with changing global conditions. I am establishing a spatiotemporal series through recent years in California using, for example, the TM data portrayed here. Red colors indicate the warmest temperatures at approximately 10 AM local time. In this work, temperatures are calibrated to ground observations, and changes in related variables, including potential evapotranspirtation rates, are estimated over time.
|
|
Using a linear transformation and novel visualization of high resolution IKONOS data to estimate flammability in the Amazon rainforest
|
Modeling tropical forest microclimates using high resolution multispectral data Even after long periods without precipitation, undisturbed tropical rain forests do not usually become flammable. Their canopy structures moderate microclimatic conditions at ground level. However, seasonally dry rain forests that have been subjected to anthropogenic disturbances do experience large, often unintentional fires that have been a major contributor to their further deforestation. When the canopy is broken, exposure to sunlight increases temperature and reduces humidity and the wetness of materials at ground level. Slash provides fuel, and human activity often provides sources of ignition. This study characterizes flammability in Amazonian tropical forests by post-processing Tasseled Cap transforms of high resolution, pan-sharpened multispectral IKONOS data. It is demonstrated that the first three Tasseled Cap components of IKONOS data can be generally described at Brightness, Greenness, and Wetness, as they have been with Landsat TM data. Examination of landscape conditions indicates that a high Brightness value in conjunction with a very low Wetness value generally characterizes flammable conditions; a useful index is the simple quotient of Brightness and Wetness (tc1/tc3). Nonflammable conditions within a vegetated landscape can often be characterized by high values of both Greenness and Wetness, as indexed by their product (tc2xtc3). Composite images in which values of the tc1/tc3 index are portrayed in red and values of the tc2xtc3 index are portrayed in green and blue are used to provide visualizations of both disturbance and possible flammability more clearly than do natural color or color IR imagery.
Modeling Flammability in Disturbed Tropical Forests using an IKONOS Tasseled Cap Transform
Proceedings, The 2006 ASPRS Annual Meeting, Reno NV
|
|
|
Modeling tropical forest microclimates using high resolution multispectral data Detecting disturbances and features at fine scales: (a) natural color; (b) color IR; (c) Tasseled Cap index composite, where bright red indicates increased flammability. Note anthropogenic linear feature - a largely covered drag trail from selectively logged regions. Horizontal spatial extent is 2.1 km.
|
|
A standard 500 m buffer of the shoreline of San Francisco Bay (Landsat TM data). Note the scalloped appearance of the buffer around land features. This is a common artifact of GIS buffering operations.
|
|
|
A nearly equivalent buffer, created using the method in the paper described below. Results are more physically meaningful, and cartographic artifacts, including buffers around small features like bridges and vessels, are automatically removed.
|
An alternative field-based method of buffering geographic features Proximity buffers are used in many spatial applications in research and management. Nevertheless, they are limited in their representational validity. Since only the nearest points on the edge of an entity are used in calculating the buffer boundary, the various meanings of the influence of an entity on its surrounding environment are not well-estimated. An alternative implementation of a more generalized class of buffers is described here. This method considers the contribution that the internal spatial geometry and attribute values of entities have on the buffered environment. It also considers the cumulative influence of multiple entities. The method is easily implemented by using a class of integrative spatial filters. For many applications, it is likely to yield results that are more meaningful than those obtained through proximity buffers.
Alternative Buffer Formation
In: GIScience 2004, M. Egenhofer, C. Freksa, and H. Miller (Editors), Springer Lecture Notes in Computer Science, 2004.
|
|
|
Geographic Visualization Techniques Flythroughs created several years ago using Landsat TM and USGS DEM in ERDAS Imagine. While the movies below are not novel by today's standards, they exemplify the simplest kinds of dynamic visualizations that can be done using desktop geospatial software. Such software still offers much greater flexibility both in terms of visualization techniques and data being visualized than do online servers like Google Earth. However, the situation is changing rapidly.
|
|
|
|
Point Reyes to San Francisco
|
Click on the image to begin. Movie will run in a separate window.
|
|
|
|
San Francisco to Mount Diablo
|
Click on the image to begin. Movie will run in a separate window.
|
|
|
Click on the image to begin. Movie will run in a separate window.
|
|
|
Click on the image to begin. Movie will run in a separate window.
|
|
|
Jewels and Binoculars
www.millenniumassessment.org
Reports from the international effort to inventory global ecosystems, their contribution to human development, and the effects of the ongoing degradation of the world's environments
www.seva.org
Seva Foundation, dedicated to finding skillful means to relieve suffering caused by poverty around the world.
www.kiva.org
Kiva - microloans that change lives.
www.cgiar.org
Consultative Group on International Agricultural Research, a strategic
alliance of members, partners and international agricultural centers that mobilizes science to benefit the poor.
www.savedarfur.org
to alleviate the suffering of the people of Darfur
www.CircleOfBlue.org
Information and resources about the global water crisis
www.realclimate.org
Climate science from climate scientists
www.rmi.org/
"Rocky Mountain Institute(RMI) is a nonprofit organization that fosters the efficient and restorative use of resources so that companies, governments and organizations are more efficient, make more money, and do less harm to the environment. RMI is engaged in cutting-edge research on oil independence, renewable energy technologies, distributed energy, resource planning, green buildings , and radically efficient transportation.
|
|
|
|
Diversification, coherence, and resilience in geographies of process and change Currently conducing studies in complexity and evolution in geographical systems, in the context of agent-based generative modeling. Spatially distributed variability in the values of key attributes and a resulting diversity in the functional behavior of a broad range of geographical entities can be associated with the emergence of robust, persistent, and coherent patterns and structures. This principle can be applied to a broad range of physical, social, and cultural phenomena. A number of concepts are explored in this work: modularity, weak connectivity, exploratory behavior, adaptation, niche construction, and stigmergic structuration. If a geographical system is viewed in terms of these and related characteristics, the emergence of cooperation, coherence, and resilience can be understood in evolutionary terms. The long document is nearing completion. Below are chapter headings and selected excerpts. Chapters: Introduction Perspectives on diversity Facilitated variation Modularity, weak connectivity, and the recursive nature of agency Variation, noise, and exploratory behavior Adaptation, niche construction, coexistence, and cooperation Spatiotemporal coherence Resilience Is land use change evolutionary? Generative models of evolutionary geographies Discussion References
Is land use change evolutionary? (excerpt)
References
|
|
|
from Introduction Discussions of variation, noise, diversity, self-organization, coherence, and resilience from the physical, biological, economic, organizational, and computational sciences are joined here to inform a generative approach to geospatial dynamics and decision making, in order to examine the nature and significance of functional diversity in geographic domains. To illuminate these points, very simple, easily duplicated spatially distributed agent-based models will be used to illustrate the striking effects of variation on the formation of coherent patterns in both abstract and representational geographic landscapes. It is argued in this study that a diversity of strategies among geospatial entities, and a diversity of interactions between them, not only improve resilience across multiple scales, but often actually increase the efficiency, productivity, and long-term sustainability of geographic, socio-economic, and socio-ecological systems - more so than any single statistically optimal strategy, and even when these diverse entities and processes are far from optimal in their immediate, individual effects.
Introduction (excerpt)
|
|
Producer-Consumer ecosystem, no consumer diversity.
|
from Generative models of evolutionary geographies The section on generative agent-based models of evolutionary geographies begins with several simple, easily validated examples. Pictured here is a producer-consumer ecosystem (imagine, for example, a grassland with herbivores), where consumers (pictured) are mobile, reproductive, and adaptive. In the landscape above, a lack of functional diversity yields random behavior. Diversity in their individualized behaviors produces some startling effects. Most obvious are the formation of self-organized patterns, through stigmergic modification of the producer environment, shown below.
|
|
Producer-Consumer ecosystem with consumer diversity.
|
All consumers here have a variety of velocities. Patterns formed are robust and stable. The target pattern is centered on the original location of the ancestral population. This simple model ecosystem exemplifies niche construction and coherence resulting from stigmergic modification of the producer environment.
|
|
|
A scenario for behavioral divergence that establishes spatiotemporal coherence A single consumer agent reproduces at time step i plus 1. The velocity of the descendent is slightly greater than that of the parent. Descendent travels to right, riding crest of wave in resource availability. Parent travels to the left, utilizing regeneration in refractory zone behind crest. Agents moving against the wave are responsible for the generation of new waves when they converge on ancestral source region. The continuous generation of target patterns occurs under a diversity of agent velocities.
|
|
Cooperative behavior in agent diversity
|
Pictured here is a producer environment as modified by diverse consumer populations. Shading indicates biomass, which regenerates. While these wavelike patterns pulsate outward, they continue to be generated as a cohort of the consumer population travels back into the center of each pattern. Multiple patterns always reinforce each other; there is no destructive interference.
|
|
|
The patterns discussed in this work are essentially dynamic. Model simulations and their analysis are therefore spatiotemporal, best viewed as film clips. Click on the image above to view these simulations. The movie will open and run in a separate window. This streaming video was first presented at Geocomputation 2005, Ann Arbor, MI.
|
|
|
Click on the image above to view these simulations. The movie will open and run in a separate window. This streaming video was first presented at Geocomputation 2005, Ann Arbor, MI.
|
|
Exploring behavioral diversity in traffic.
|
Three representative runs of a traffic flow model, with different degrees of variance in vehicle acceleration rate. The road is represented along the horizontal axis, and is continuously joined at the ends.
|
|
|
The origins of this work in 2005 The modeling interface for this study consists of switches, sliders, and indicators for model structure and operation; graphs of important global statistics; and a dynamic representation of the model landscape. Distributed models of autonomous consumer agents on resource landscapes are used to demonstrate the effects of functional diversity. Comparisons between populations of agents with various degrees of diversity in the values of key functional variables indicate that moderate diversity results in the formation of spatial patterns, higher yield, and greater resilience. These results are discussed in the context of previous research, as are their implications for understanding and improving the resilience of geographical systems.
Investigating the effects of functional diversity in spatially distributed geographic doimains
Proceedings, Geocomputation 2005, Ann Arbor, MI
|
|
|
Attractor states differ under different conditions of consumer velocity diversity (deviation). Each phase portraits encompasses 1000 time steps after dynamic equilibrium is achieved in producer/consumer simulations, with consumer velocity diversity conditions indicated as standard deviation around 1 cell per time step. These portraits indicate that higher yield is achieved at similar population densities with greater velocity diversity, and that the instantaneous values of population and yield correlate differently with different diversities. Results are robust and repeatable over many runs.
|
|
|
