Equipment and Software

Our lab hosts a variety of state-of-the-art equipment and software to increase the capacity of student research and facilitate the generation of high-quality and high-impact research.


  • Aeryon Labs SkyRanger
  • Icom VHF Air Band Transceiver IC-A14
  • Macbook Pro and Dell Laptops
  • 2 – Dell Precision 5810 Workstations each with NVIDIA Quadro K4200 4GB CUDA enabled graphics cards
  • 2 – PowerEdge R720xd Servers
  • Ominus Dromida – for some UAV training fun


  • ArcGIS
  • eCognition
  • Pix4D
  • ENVI

In addition to these specialized pieces of equipment and software we also have high resolution monitors, and regularly use software like R statistics package, eclipse, fragstats, Adobe Creative Suite, Office products, NetLogo, RePast, NetworkX, among others.


Advances in Modelling Land Change and Estimating Ecosystem Function Across Ontario – PhD Funding Available via NSERC-DG

The effects of land-use and land-cover change on how ecosystems function has become a critical issue of local, provincial, national, and global concern. Over 30% of the Earth’s surface has been altered by human activities and while the extent of alteration by natural disturbances is uncertain, these disturbances may have greater ecological impacts than human activities in countries with large areas and relatively small populations (e.g., Canada). Because human and natural activities driving land-use and land-cover change are so pronounced, land-use and land-cover change is inexplicably linked to ecological processes and often guides the quantity and quality of ecosystem function.

In Canada, measurement and monitoring of land-use and land-cover change is achieved using remote sensing (e.g., Canada Land Inventory, Annual Crop Inventory). Remote sensing products are extremely useful in terms of quantifying historical land-use and land-cover change and deriving estimates of ecosystem function. However, these estimates, which are often used in national assessments, typically use top-down approaches (e.g., benefits transfer) that fail to account for changes in ecological function due to climate conditions or landscape heterogeneity (i.e., site characteristics).

To incorporate biogeochemical (e.g., nutrient cycling), biogeophysical (e.g., heat flux, evaporation), and other process effects on ecosystem function, scientists use ecosystem process models (e.g., Century, Biome-BGC, CBM-CFS3). These models account for heterogeneity in soil, climate, and other factors that drive vegetation growth and other ecological processes. However, they are typically applied at coarse scales (e.g., 0.5 degree resolution). At this resolution they offer useful estimates of national scale ecological conditions, but results do not coincide with land units at which humans make decisions or at which policy is made. Furthermore, these models are ill equipped to address changes in land cover, land management, and among other issues, landscape fragmentation. Contemporary ecological modeling efforts use prescribed land-cover data and are unable to account for dynamic changes in land use, land cover, or land management. New models are needed to integrate a dynamic representation of land-use and land-cover change and ecosystem process to improve our understanding and quantification of how land-use and land-cover change affects ecosystem function.

Despite calls to improve research and models that link land-use and land-cover change with ecological processes by the Global Land Project, Future Earth, and the United States National Science Foundation’s Coupled Natural/Human Systems program, a national land-use and land-cover change model for Canada has yet to exist. This research program uses a bottom-up modeling approach to create a land-use and land-cover change model for Ontario that integrates with ecosystem process models for the explicit purpose of estimating the effects of different quantities and patterns of land use and land cover derived from socio-economic and policy scenarios on ecosystem function. Results from this research will 1) garner an improved understanding of the interactions, feedbacks, and thresholds associated with ecosystem function due to different quantity and patterns of land use and land cover, 2) identify sensitive ecological areas vulnerable to specific patterns of land use and land cover, 3) provide methodological advancements in terms of model integration and conceptual issues in coupling land-use and land-cover change with ecological processes, and 4) provide the first step towards a national land-use and land-cover change model and ecosystem function assessment program.


Reference Landscapes – PhD Funding Available

Within a larger interdisciplinary research team, our lab is applying the reference condition approach at the landscape scale to quantify the properties, function, and the pattern of habitats in natural landscapes for comparison with reclamation efforts. Using remote sensing, geographical information systems (GIS), and ecological fieldwork we identify suitable reference areas to the oil sands region, but where human disturbance is low. Then we characterize the quantity and pattern of habitats and identify aggregate landscape properties that “fingerprint” naturally occurring landscapes in Alberta. A number of metrics are used to quantify edge characteristics, the degree of fragmentation, the shape of individual patches of habitat as well as across the landscape to fingerprint these landscapes. Global measurements of spatial autocorrelation and local indicators of spatial association are used to characterize the interrelationship among habitats within reference watersheds and sub-areas. By quantifying the pattern and properties of reference landscapes at multiple scales, and their variance among reference landscapes, we can define criteria and ranges of criteria values to guide closure-landscape planning towards a more natural appearance. Incorporating spatial assessments of naturally occurring ecohydrologic networks and ecosite distributions constitutes a major innovation in mine reclamation. The project is funded by Alberta Innovates – Energy and Environment Solutions and has the capacity to fund 1 additional PhD on my project team. If this project is of interest then please send an email outlining your qualifications, unofficial transcript, and example of your writing.

Find out more →


Estimating Market Potential and Location Allocation for Land-Use Modelling

As companies expand their organization across Canada they face tough decisions about where to locate new stores. These decisions take place at different scales and require choices to be made about what towns or cities would benefit from a new store and where within these towns or cities should that store be located. The Estimating Market Potential and Location Allocation for Land-Use Modelling project extends traditional site-location analysis and develops new approaches and customized tools to assess the best locations for new stores. More broadly, the project establishes a baseline of analysis routines that can be incorporated into models of land-use change to evaluate how drivers of site selection change over time.


Infrastructure for Modelling Human-Environment Interactions in Agricultural Systems

Agricultural production is impacted by both the ecological characteristics and processes of a farm and the land-management decisions that alter them. This project improves our ability to collect data on agricultural systems and unpack the complex decisions made by farmers and policy makers, who are striving to increase agricultural production, farm livelihoods, and reduce associated ecological impacts. The project combines unmanned aerial vehicle infrastructure with tablet technology and ecological field equipment to collect new data in an innovative way and uses those data in a new agent-based agricultural land management system (AALMS) simulation model. The AALMS model simulates farmer responses to changing socio-economic, policy, and climate scenarios. Alms – the act of charitable giving – inspired the model name and reminds the project team that integrated farmer-science-policy research and collaborative learning have the potential to positively impact the lives of others. Simulation results can inform farmers and policy makers of the potential economic and ecological consequences of their decisions. Collected data are given to farmers and simulation results are made available to the public so that they may better understand the complexities associated with farming and the trade-offs and synergies amongst ecological functions that occur in response to management decisions. This research is possible with infrastructure provided by the Canadian Foundation for Innovation. Students: If you have funding available and would like to work on this project please send an email outlining your funding status, why you are interested in the project, and your unofficial transcript. Farmers: If you are interested in working with our project team and infrastructure please send an email along with contact information and a member from our team will contact you.


SLUCE (Spatial Land Use Change and Ecological Effects) University of Michigan, School of Natural Resources and Environment, PI Dr. Daniel G. Brown

Selection of the degree of fidelity versus parsimony when answering and modelling a problem, especially one that lies within a complex system, often leads to an iterative process between data collection and modelling. Within the SLUCE project, much of my involvement incorporated similar decisions, i.e. expanding the development of a model or eliminating model components. The project initially developed a LUCC model named SOME (SLUCE’s Original Model for Exploration) that incorporated two factors into residential location decision-making: 1) aesthetic quality, and 2) nearness to service centers (i.e. urban amenities). I expanded the SOME conceptual model to include a third factor, whereby settling residential households formed measurements of neighbourhood similarity between their preferences and those already established residents. Using results from a household survey (Fernandez et al. 2005), Daniel Brown and I empirically justified these three factors and used the data to inform the preference weights of the population of residential household agents in the SOME model. We incorporated the survey data at several different levels of analysis that altered the number of agent categories and variation in agent preferences (Brown and Robinson 2006). Results from this work illustrated that 1) introducing variability in preferences increased the amount of agent dispersion, or sprawl, the model produced, 2) relationships between groups of similar agents indicated that agent preferences, and their distributions across various factors, affect spatial patterns of development and the utility achieved by agents, and 3) generalist agents achieved highest average utility levels. The group then developed a more complex conceptual model that incorporated farms, developers, subdivisions, and townships in addition to residential households. I coauthored a paper with the group on linking empirical data to these two conceptual models (Brown et al. In Press). Altogether I coauthored three publications under the SLUCE project and presented related research at four conferences. The NSF Biocomplexity in the Environment Program (BCS-0119804) project ultimately led me to a forth publication and dissertation chapter (Robinson and Brown Submitted)

Effects of land-use policy, forest fragmentation, and residential parcel size on land-cover and carbon storage in Southeastern Michigan

The overarching goal of this dissertation was to improve our understanding of the coupled natural-human land-use system in Southeastern Michigan. To accomplish this task Chapter Two presents an implementation of the DEED (Dynamic Exurban Ecological Development) model, which was used to evaluate the effects of land-use policies on forest cover. This research demonstrates one way to improve our understanding of how policy and land-use and land-cover change (LUCC) interact and can influence aggregate forest cover. The chapter provides novel contributions in the form of a framework for evaluating land-use policy effects on development and land cover, an approach to integrate an agent-based model with a geographic information system (GIS), and new examples of methods to empirically inform agent-based models.

To extend coupled natural-human systems research to include the ecological effects of LUCC and policy scenarios, Chapter Three presents an analysis of the effects of forest patch size and shape, and landscape pattern, on carbon storage estimated by BIOME-BGC. New insights from this research showed 1) the inclusion of within-forest-patch air-temperature heterogeneity can significantly influence carbon storage estimates, 2) that carbon storage estimates increase logarithmically with increasing forest fragmentation when only within-patch heterogeneity of air temperature is considered, and 3) the utility of integrating GIS and BIOME-BGC for site data collection and visualization of results. (Photos from the field work)

To better evaluate the effects of land-use development policies on land-cover change and ecosystem function, effectively combining the products of Chapters Two and Three, an analysis of land cover at the residential parcel level was necessary. Land-cover analyses at the parcel level have rarely been done. Chapter Four takes a step to remedy this problem by presenting new data that describe the quantity, fragmentation, and autocorrelation of land-cover within residential land-use parcels. Results from Chapter Four could extend the policy analysis in Chapter Two, conducted at the subdivision level, to the individual parcel such that we could evaluate policies that affect individual residents and their behavior. By capturing the distribution and patterns of land-cover types across different parcel sizes we can begin to understand the linkages between household land-cover behaviors, neighborhood interactions, and landscape patterns.

PHESI (Program in Human and Environmental Systems Interactions), University of Michigan, School of Natural Resources and Environment, 2007, PI Dr. Arun Agrawal

In this project we address common pool resource management questions related to land use and land cover change. The project attempts to answer how user behavior and resource-related outcomes change as users place greater emphasis on externally imposed institutions that rely on one way flow of information about resource condition relative to spontaneously generated institutions that permit a two-way flow of information about forest conditions. The project is ongoing and currently I am co-writing a paper that investigates how changes in the relative dependence of users on different institutions affect user behavior, harvesting levels, and forest-related outcomes. We investigate these institutional influences using an agent-based model that is informed by the results of a survey of 2200 users in the Indian Himalaya. The paper is planned for submission this year or early 2008 and the project is funded by an NSF Human and Social Dynamics (BCS-0527318) grant.


Back row: Derek Robinson, Rick Riolo, and Dan Brown. Front row: Arun Agrawal and Gautam Rao.

ESRI (Environmental Systems Research Institute)/Argonne National Laboratory/University of Michigan, Aug. 2004 – Aug. 2005

Traditionally, the analytical and computational tools used by geographers poorly represent time. It is often the tools from this community that are used in the spatial modelling of land use and land cover change. In an effort to improve the ability to represent temporal processes associated with spatial data models and link geographic data and spatial processes to agent-based models this project sought to design a tool, called Agent Analyst. My role in the project was to provide theoretical, technical, and design input as well as debug prototypes, implement sample models, and contribute to papers. From these efforts I presented at two conferences and coauthored a conference proceeding and refereed journal. The conference proceeding compared a number of different ABM platforms and their ability to represent geographic data and processes (Rand et al. 2005). The results illustrated the degree to which each toolkit was useful for novice modellers and that none of the available toolkits provided a useful integration of GIS and ABM. The journal publication was more theoretical and used examples to describe the types of linkages between ABM and GIS and their associated advantages and disadvantages (Brown et al. 2005). An ESRI Press book is due to come out on using Agent Analyst with ArcGIS. I have lead a chapter in this book and posted some screen shots of the model on the abm page of this website.

Agent Analyst
From left to right: Derek Robinson, Robert Najilis, Michael North, Kevin Johnston, and Bill Rand. Missing are Daniel Brown and Rick Riolo.

LUCITA (Land Use Change in The Amazon), University of Waterloo & Indiana University, 2001-2003, with Dr. Peter Deadman.

A fundamental step in LUCC research is the identification of drivers, or causal mechanisms, that influence landscape change. An area that has received widespread attention in LUCC research is the Amazon basin, due to the rapid rate of change and the region’s importance as a carbon sink and source of biodiversity. Emilio Moran and Eduardo Brondizio of the Department of Anthropology at Indiana University had been conducting field research along the trans-Amazon highway near Altimara, Brazil, since its inception in the early 1970’s. They hypothesized that much of what was driving land use changes (i.e. deforestation trends) on individual properties was based on household demographic characteristics. Our goal with the LUCITA project was to formalize research on colonist household farming practices and decision-making behaviours using an agent-based model to determine if the hypothesized conceptual model could produce overall patterns of land use change that agreed with those observed in remotely sensed images. The results of my Master’s thesis (Robinson 2003) and coauthored paper (Deadman et al. 2004) illustrated that the use of simple heuristic decision-making strategies representing colonist household farming decisions were capable of producing land cover change trends that compared well with remotely sensed data from 1970-2000. We also determined that households did not make land-use decisions based on their length of time on the frontier, but simply on the basis of available household resources (i.e. labour and capital availability as a secondary result from changing household demographics and structure), the performance of previous crops, and the landscape characteristics of the farm parcel (e.g. soil and burn quality). Pre-existing lot characteristics were also found to influence household land use choices. Overall, our contribution suggested that the conceptual model of household demographics and structure changes could explain a portion of LUCC trends but that additional research focussing on lot effects, and information dissemination amongst farmers was needed. The LUCITA project was a sub-component of a larger NSF biocomplexity in the environment program (SES00835) project titled “Biocomplexity in Linked Bioecological-Human Systems: Agent-Based Models of Land Use Decisions and Emergent Land Use Patterns in Forested Regions of the American Midwest and the Brazilian Amazon, whose PI was Elinor Ostrom.

Back to Top
Enter your Infotext or Widgets here...