Bonanza Creek LTER
BNZ LTER Web Site Map

The Maps and Locals Project (MALS)

Background and Coordination: MALS is a collaborative effort of LTER sites that seek to research changing social-ecological systems using a mixed methods comparative approach. Coordinating the project are Gary Kofinas of the University of Alaska Fairbanks, Robert (Gil) Pontius of Clark University Geography Department, and Nathan Sayre of University of California Berkeley. Gary serves to coordinate and integrate project activities. Gil oversees the spatial analysis, and Nathan, with Gary, leads the work on the contributions of local knowledge. Eleven LTER sites have been funded by the National Science Foundation to participate in MALS. Other LTER sites and groups are participating at various levels. (See listing of participants below).

Table 1 - Participants

Table 2 - Site Descriptions

Table 3 - Maps

Table 4 - Major Drivers & Implications

Rationale for MALS: Understanding the dynamics of social-ecological interactions and organizing system-wide investigations through comparative analysis of LTER sites are key objectives of the ISSE Decadal Plan. These objectives address a grand challenge for both applied and theoretical ecology of discerning past human effects on ecological systems and distinguishing anthropogenic from non-anthropogenic drivers of change. Although most LTER sites have assembled historical data on land cover, climate, vegetation, and other ecological attributes, less is known about historical resource management practices and other human influences. In many cases social data are available for human communities of LTER regions, but are not well compiled. Local ecological knowledge (LEK) from long-time residents and users of landscapes provides an additional source of insight which has the potential to add to LTER studies, but methods for documenting and integrating such data into ecological research remain poorly developed and a subject of debate. Our objective in this collaborative research project is to develop methods and capacity for research into social-ecological systems both at individual sites and at the network scale.

Objectives and Research Plan: We are involved in a cross-site collaborative effort with several complementary areas of activity. Participating LTER sites will 1) use spatial representation of land cover and land use to identify patterns of landscape change in regions in and around LTER sites; 2) integrate LEK and other existing social data into theories and models of ecological change and their implications for human livelihoods. LTER sites participating in this program of research will emphasize these activities to varying degrees. Cross-site comparisons will aid in developing methods and questions, in testing hypotheses over larger scales, and in setting the stage for future cross-site comparative studies.

MALS was launched with a workshop at the 2009 ASM in Estes Park, Colorado. The project involves a series of virtual meetings at which participating LTER researchers evaluate and define problems of human land use and change, identify the group's methods and strategies, and arrive at a set of common approaches. A key focus will be identifying social and ecological processes that are hypothesized to delimit the region in question, and specifying the spatial and temporal scales of those processes. We will also assess availability and utility of documentary sources (e.g., historic aerial photographs, weather station data, etc.). A second face-to-face worksho will take place in spring, 2010.

Mapping: Concurrently, each site will assemble a time series (n=2) set of maps that represent known biophysical, infrastructural, and land-use changes in its region. At each site, maps will be used both as corroborating data and as research tools for use in collecting local ecological knowledge. At the network scale, we will gather the maps from several sites for development and application of methods for spatial/GIS analysis.

Local Knowledge Documentation and Social Data: For the documentation of local knowledge, we identify individuals and classes of informants for each site, with an emphasis on people who have had continuous or regular familiarity with specific places over long time periods (10-50 years, or potentially more through ancestors). Such familiarity may involve direct management of a property or repeated regular visits for specific purposes. Considerable information about local knowledge has already been documented in many regions through past and current research projects, and to the extent possible, participating LTER sites will draw on those data to meet the needs of this comparative study. Additional new data will be collected where opportunities and resources are present. The compiled data will serve both to generate hypotheses and as a means of corroborating and illuminating existing data.

Coordination and approach: We seek a Linux-like, open-source organizational approach to our collaboration and have intentionally left many of our specific activities and research questions defined in general terms for specification in individual site supplemental proposals and for further refinement over the course of the projects. We also commit to coordinating our work with the "Lawns LTER project" so our efforts are complementary. Coordination of the initiative is by Gary Kofinas (UAF/BNZ and ARC), Nathan Sayre (UC-Berkeley /JRN), and Gil Pontius (Clark U).

Details about spatial analysis: We propose to compile a database of land category maps from each participating site from at least two points in time. The set of maps from each site will show two or more land categories which will overlay on a single raster grid to facilitate statistical analysis. If we have maps from only two points in time, then we can characterize the changes over one time interval; whereas if we have maps from more than two points in time, then we can analyze whether the process of land transformation has been stationary across more than one time interval. This effort will advance the LTER agenda in two respects: 1) creation of the database, and 2) development of methodology.

Concerning the first goal of database creation, most sites already have the raw materials ready to participate; however these materials are either incomplete and/or in need of better documentation. For example, the BES site has a potential gold mine of JPG files that show urban expansion for fourteen non-consecutive years extending back to 1792. When we started to compile these, we realized that it was not easy to find metadata that stated the precise meaning of "urban" or explained why a substantial piece of Baltimore City lost urban land cover between 1925 and 1938. It is not clear whether this is a true loss of urban cover or a data problem since metadata are difficult to access. Furthermore, it is not clear how we should select the bounding coordinates of the study extent for the raster grid, since the existing maps are in vector format and many of the variables have somewhat different spatial extents. For other sites that we have examined, it is not clear how we should select the land categories. For example, we found that the CAP site has some maps that show 21 categories while other maps show 4 categories. We need to develop a methodology to reduce the number of categories so we can focus on a manageable number of land transitions, since 21 categories can produce a dizzying array of more than 400 possible transitions, while the use of 4 categories might mask some important transitions. Usually some category aggregation is useful to help focus on the most important categories. Pontius and Malizia (2004) have distilled the mathematical principles that dictate how category aggregation influences measurements of land change over time. Furthermore, we have found some substantial erroneous artifacts in some of the LTER sites' maps, such as seams in the elevation maps, which are not immediately evident until we use the elevation map to create a slope map. Typically, metadata concerning the accuracy of maps does not exist, in which case we can use the methods of Pontius and Lippitt (2006) to examine how the suspected errors in the maps influence the estimates of land transformation.

Concerning the second goal of methodology development, we have a hypothesis that some of our recently proposed methods would be particularly well suited for this type of cross site comparison. Pontius et al. (2004) proposed a novel method to test for systematic transitions among land cover categories when maps are available from two points in time. Their methods revealed that new expansion of built land in Massachusetts targets open space and avoids forest, in spite of the fact that most of the gross gain in built displaces forest. The reason is that the initial landscape is mostly forest, so the gain of the built environment is likely to displace forest even when developers prefer to avoid deforestation. Alo and Pontius (2008) advanced this method in order to compare two different sites to test whether the processes of land transformation inside a protected area are different than the processes outside a protected area. The next advancement in this methodological approach will be to examine the transitions among maps from three points in time to test whether the transitions during the former time interval are different than the transitions during the latter time interval. Chen and Pontius (in review) have proposed a complementary method to test whether the process of land transformation is stationary along a gradient, e.g. slope or distance to forms of human infrastructure such as highways. The investigators already have a substantial amount of experience in performing cross site comparisons using raster maps of land cover over time, in the context of NSF's Human-Environment Regional Observatory (HERO) program (Pontius and Malizia 2004) and the international Land Use and Land Cover Change (LUCC) program (Pontius et al. 2008).

Slide Presentation: Identifying Stationarity of Transitions Among Land Categories Over Time


Clark University students created these posters as part of a course called "GIS & Land Change Science" in fall 2009 taught by Gil Pontius. Each student adopted one site from our Maps and Locals (MALS) research project. Most posters use a statistical methodology that analyzes the intensities of the land transitions among three points in time. We are now beginning to show the results to the locals who sent us the maps. Preliminary feedback indicates that we need to show the simpler results of the size of the transitions, before the audience can understand the intensities of the land change. We desire feedback from the participants at the sites to guide us in the next phase of analysis. The students will revise the posters and present them in Washington DC at the Association of American Geographers in April 2010.

MALS Project Posters

Title Author(s) LTER Site Poster Link
Sensitivity of land cover analysis to category aggregation

(April, 2010)
Thuy T. Nguyen, J. Hepinstall-Cymerman, T. Gragson, J. Chamblee and R. G. Pontius Jr CWT
The Markov matrix versus the Duration, Category and Transition (DCT) matrix: a case study of land change in the Florida Coastal Ecosystems

(April, 2010)
Wuxuan Xiang, Robert Gilmore Pontius Jr., Jeffrey A. Onsted FCE
Method to Show the Level of Stationarity of a Land Change Process along a Gradient: Forest Regrowth at Hubbard Brook

(April, 2010)
Fei Meng HBR
Land Change Process along Ecological Gradients: Forest Regrowth at Hubbard Brook, NH

(April, 2010)
Fei Meng HBR
Can Observed Differences Between Maps of Two Times be Explained by Error?

(April, 2010)
Jenner Alpern HFR
Land Change intensity analysis on Andrews, Oregon

(April, 2010)
Kangpin Si, Myrica McCune, Robert Gilmore Pontius, Jr, Hannah Gosnell HJA
Land Change Intensity Analysis for H.J.Andrews, Oregon

(April, 2010)
Kangpin Si, Myrica McCune, Robert Gilmore Pontius, Jr, Hannah Gosnell HJA
Jornada LTER Site: Three Level Analysis of Land Category Change.

(April, 2010)
Trevor Heburn JRN
Jornada Basin Long Term Ecological Research Site: Conversion of Grassland to Shrubland

(April, 2010)
Trevor Heburn JRN
Konza Long Term Environmental Research Site: A Characterization of Land Change from 1990-2009

(April, 2010)
Nicholas Wilson, Amintas Brandão Jr. , and Robert Gilmore Pontius Jr. KNZ
Assessment of Land Change Around Forest Conservation Areas in Luquillo, Puerto Rico

(April, 2010)
Eva Zhang LQU
Analysis of Trends in Land Category Transitions for More Than Three Points in Time

(April, 2010)
Yujia Zhang PIE
Do Builders Consistently Target Forest from 1971 to 1999?

(April, 2010)
Yujia Zhang PIE
Reconciling inconsistencies of maps that were created with different methods for 1999 and 2005

(April, 2010)
Yao Yao PIE


Alo, Clement and Robert Gilmore Pontius Jr. 2008. Identifying systematic land cover transitions using remote sensing and GIS: The fate of forests inside and outside protected areas of Southwestern Ghana. Environment and Planning B 35(2): 280-295.

Chen, Hao and Robert Gilmore Pontius Jr. in review. Diagnostic tools to evaluate a land change prediction along a gradient. Computers, Environment and Urban Systems.

Pontius Jr, Robert Gilmore and Nicholas R Malizia. 2004. Effect of category aggregation on map comparison. Lecture Notes in Computer Science 3234: 251-268. In M J Egenhofer, C Freksa, and H J Miller (eds): GIScience2004.

Pontius Jr, Robert Gilmore, Emily Shusas and Menzie McEachern. 2004. Detecting important categorical land changes while accounting for persistence. Agriculture, Ecosystems & Environment 101(2-3): 251-268.

Pontius Jr, Robert Gilmore and Christopher D Lippitt. 2006. Can error explain map differences over time? Cartography and Geographic Information Science 33(2): 159-171.

Pontius Jr, Robert Gilmore, Wideke Boersma, Jean-Christophe Castella, Keith Clarke, Ton de Nijs, Charles Dietzel, Zengqiang Duan, Eric Fotsing, Noah Goldstein, Kasper Kok, Eric Koomen, Christopher D. Lippitt, William McConnell, Alias Mohd Sood, Bryan Pijanowski, Snehal Pithadia, Sean Sweeney, Tran Ngoc Trung, A. Tom Veldkamp, and Peter H. Verburg. 2008. Comparing the input, output, and validation maps for several models of land change. Annals of Regional Science 42(1): 11-47.

weather Data Policies | Webmaster Click for Fairbanks, Alaska Forecast

The Bonanza Creek LTER, including this website, is supported by the National Science Foundation through awards DEB-1026415, DEB-0620579, DEB-0423442, DEB-0080609, DEB-9810217, DEB-9211769, DEB-8702629 and by the USDA Forest Service, Pacific Northwest Research Station through agreement number RJVA-PNW-01-JV-11261952-231. Any opinions, findings, conclusions, or recommendations expressed in the material are those of the author(s) and do not necessarily reflect the views of the supporting agencies or the program as a whole.

© Bonanza Creek LTER, 2011.
Last modified 19-Apr-14
UAF is an affirmative action/equal opportunity employer and educational institution.
This site is a member of the LTER Network.