Hope C. Humphries (NWT) and Patrick S. Bourgeron (NWT)
The Maps and Locals (MALS; http://www.lter.uaf.edu/bnz_MALS.cfm) project started in 2009 as a collaborative effort funded as social science supplements for 11 participating LTER sites (AND, ARC, BNZ, CCE, CWT, GCE, JRN, KBS, KNZ, LUQ, and NWT) to investigate changes in socio-ecological systems using a mixed methods comparative approach. Other LTER sites and groups have been involved at various levels. MALS coordinators are Gary Kofinas, BNZ & ARC; Robert (Gil) Pontius, PIE; and Nathan Sayre, JRN. The specific objectives of MALS are to: (1) use spatial representations of land cover and land use to identify patterns of landscape change in regions in and around LTER sites; and, (2) integrate local ecological knowledge (LEK) and other existing social data into theories and models of social ecological change and their implications for human livelihoods. LTER sites participating in this program of research emphasize these activities to varying degrees. Cross-site comparisons are being conducted to develop methods and questions, test hypotheses over larger scales, and set the stage for cross-site comparative studies. Three workshops were held to coordinate MALS activities. A three-day training/planning workshop will be held in 2012 that provides LTER investigators, LTER graduate students, and others with a theoretical orientation, practical skills, and the research tools to document local ecological knowledge and integrate that knowledge with spatial analysis and other forms of scientific data to understand socio-ecological resilience.
In the initial phase of MALS, research was primarily focused on cross-site land-use land-cover change analyses. Each site has assembled or is assembling a time series (n=2) set of maps that represent known biophysical, infrastructural, and land-use changes in its region. At each site, maps are used both as corroborating data and as research tools for use to collect local ecological knowledge. At the network scale, maps are collected from several sites for development and application of methods for spatial/GIS analysis. A database of land category maps is being compiled from each participating site from at least two points in time. Maps from each site show two or more land categories that are overlaid on a single raster grid to facilitate statistical analysis. If maps are available from only two points in time, changes are characterized over one time interval; if maps are generated from more than two points in time, analyses can detect whether the process of land transformation has been stationary across more than one time interval. Metrics are also being developed for cross-site analysis; for example, to measure the level of stationarity across sites in a manner that sites can be ranked from less stationary to more stationary. The maps will be designed for use in conjunction with the LTER’s web-based map browser, LTER MapS (http://www.lternet.edu/map/). In interpreting MALS results, Pontius and Millones (2011) concluded that it is more useful and simpler to compare maps in terms of two summary parameters, quantity difference and allocation difference, than to follow the previous paradigm that compares the agreement between maps to the agreement that could be expected due to randomness (Kappa indices). In this respect, MALS is inducing a transformative shift in the conceptualization and mathematics of map comparisons. The MALS effort advances the LTER agenda in two respects: 1) creation of the database, and 2) development of methodology.
Local knowledge documentation and social data
For the documentation of local knowledge, individuals and classes of informants are identified 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 draw on these data to meet the needs of this comparative study. Additional new data are being collected where opportunities and resources are present. The compiled data serve both to generate hypotheses and as a means of corroborating and illuminating existing data.
MALS work at NWT LTER
The MALS project at NWT integrates spatial data and social data into ongoing efforts that test components of the ISSE framework (http://www.lternet.edu/planning/). For this work, we have identified an area that ranges from mid-elevation (2500 m) to the alpine zone (up to 4000 m), which encompasses NWT, including the alpine and subalpine zones. In addition, a fairly large portion of Colorado Front Range forests falls within this elevation range, where substantial social and economic changes are taking place. Our initial MALS efforts focused on the area surrounding Niwot Ridge and the adjacent Green Lakes Valley. An additional site, the nearby small town of Nederland and its surroundings, is representative of many mountain towns in the Colorado Front Range that had an early history in resource extraction but have recently experienced residential development in the wildland-urban interface.
NWT human impacts and local knowledge
Although the subalpine, treeline, and alpine ecosystems of Niwot Ridge and the Green Lakes Valley appear pristine, they have experienced impacts from a variety of post-settlement human activities. A chronicle of such activities at NWT was compiled based on historical documents and interviews with long-time residents in the area (Komarkova et al. 1988). To identify key drivers of socio-economic changes in both areas, we will continue the process of acquiring census data, land transfer data, and land survey notes.
Settlement of the area began in 1861 with mining in the Green Lakes Valley, including the area surrounding Lake Albion, where the Albion townsite housed as many as 200 people in the late 1800s. The town was one of the country’s highest and most remote settlements at the time, but was abandoned by 1910; some of its buildings are still standing (Figure 1). Dam construction occurred in the Green Lakes Valley and the valley to the north of the ridge beginning in the late 1800s, with impacts that included road building, tree cutting, borrow pit excavation, and subsequent dam enlargement (Figure 2). Forests in the area, primarily at lower elevations, were cut to supply timber for mining activities. The early history of the town of Nederland and the surrounding area was associated with resource extraction, primarily mining, including several cycles of booms and busts. The economy now depends heavily on recreation and tourism. In recent years, housing density has greatly increased in the wildland-urban interface in Nederland and throughout the Colorado Front Range.
Figure 1. Albion townsite just below Lake Albion in Green Lakes Valley. Lower left photo taken pre-1890 (source: Boulder Carnegie Library, 213-1-3). Upper right photo shows remaining structures in 2008.
Figure 2. Composite aerial photos of Niwot Ridge, showing changes over time due to road and dam construction in upper right (Lefthand Reservoir) and lower center (Silver Lake).
Climate change impacts to the area include earlier snow meltout, which changes the timing and amount of runoff; this is important because the Green Lakes Valley provides water to the city of Boulder. Documentation of changes in land cover over time in our study will become especially important in the future as the current mountain pine beetle outbreak in the Colorado Front Range intensifies (http://www.rockymountainnews.com/news/2009/jan/17/colorado-pine-beetle-i...). The expected widespread mortality of pines will have a pronounced impact on land cover in the region, which in turn is likely to have a strong effect on future land use decisions.
NWT land cover mapping and analysis
At the beginning of the MALS project, Niwot had little in the way of orthorectified imagery, and no satisfactory land classification that covered the area of interest. We have put a lot of effort into acquiring and orthorectifying imagery for a number of time slices ranging from 1938 to 2008. However, they differ in type, resolution, and quality, and there is a lack of consistency in pixel values within and between images. For this reason, we are generating land-cover polygons by digitizing them. We are in the process of acquiring aerial photographs for Nederland to add to existing imagery. Polygon digitizing is underway for three time periods in each location to produce the land-cover categories forest, non-forest vegetation, rock/soil/ice, water, and anthropogenic features. Anthropogenic features and water in the area surrounding NWT have increased in cover over the period 1938/1946 to 1988/1990 due to dam construction and enlargement (Figure 2). MALS spatial statistical methods (Pontius et al. 2004) will be applied to analyzing changes in land cover over time for categorized land-cover maps in both NWT and Nederland.
Komarkova, V., A. Peters, G. Kamani, W. Jones, V. Howard, H. Gordon, and K. Southwick. 1988. Natural recovery of plant communities on disturbance plots and history of land use in the Niwot Ridge/Green Lakes Valley, Front Range, Colorado. Niwot Ridge Long-Term Ecological Research Working Paper 88/1.
Pontius, R.G., Jr. and M. Millones. 2011. Death to kappa: Birth of quantity disagreement and allocation disagreement for accuracy assessment. International Journal of Remote Sensing 32:4407-4429.
Pontius, R.G., Jr., E. Shusas, and M. McEachern. 2004. Detecting important categorical land changes while accounting for persistence. Agriculture, Ecosystems and Environment 101:251-268.