Analysis of landscape - level variation in Normalized Difference Vegetation Index

s available online at http://esa.snre.umich.edu/classes/nre540/index.html TOP Data Archive... Spatializing a Dataset to Examine Progress on the Digital Divide Tracey Hughes ([email protected]) Application of spatial analysis for community informatics is an emerging paradigm, and this GIS dataset is one of the first in the field. ArcGIS 9.1 is used here to analyze the effects of the Technology Opportunities Program (TOP). From the mid-1990s to early 2000s, TOP granted over two hundred thirty million dollars ($230,000,000) to 606 entities across the United States and its territories. Funded sites range from medical centers and universities to public utility agencies, arts organizations and tribal councils. A great deal of time and effort were required to bring the TOP GIS analysis to this point. Beginning with an Excel database containing non-standardized addresses for TOP Sites, all sites were geocoded using a combination of StreetMap USA and manual latitude/longitude identification. Census 2000 variables were identified and extracted from the Geolytics Neighborhood Change Database. TOP tract and TOP site attributes were merged using ArcGIS’s Identify feature and new data layers were created through Select and Buffer actions. The newly created TOP Sites GIS data layer has forty-seven (47) attributes relating to ethnicity, income, language and density. One of these attributes, the UMID, is common across all TOP Data Archive databases and will allow for spatial analysis on all of these databases. Mapping traditional characteristics of the Digital Divide – ethnicity and income – against TOP data reveals interesting patterns when we employ a majority/minority research methodology. We identify twenty ‘Core Sites’, with Core Sites being those located in Census tracts with majority ethnic population and majority of population with income of one-half of the average household income. The stage has been set for rich analysis of TOP. Further points of study could include: • Exploration into the characteristics and success of the Core Sites • Analysis of sites by region or year of funding • Comparison of ethnicity, income and density from 1990 to 2000 • Incorporate University of Michigan coding data re project specifics iCommunity Informatics is an emerging field studying how communities can be enabled through use of Information and Communications Technology (ICTs). It focuses on all aspects of community life including economics, sociability, culture and politics. For more information, the Community Informatics Research Network at http://www.ciresearch.net/ is a good starting point. ii The impetus for this dataset is based on a larger collaborative effort between the University of Michigan School of Information, TOP, and the University of Illinois Urbana-Champaign School of Library and Information Science to archive the TOP initiative undertaken by the US government. Collectively our team is known as the TOP Data Archive. TOP is now effectively defunct, and the TOP Data Archive is capturing all associated data for research and analysis. Analysis of landscape-level variation in Normalized Difference Vegetation Index (NDVI) at the University of Michigan Biological Station (UMBS). Iryna Dronova [email protected] Understanding the landscape-level variation in Normalized Difference Vegetation Index (NDVI) is critical for applying vegetation indices derived from remote sensing data to analysis of ground-based ecological variables related to vegetation activity at the ecosystem scale. This study compares the variation in NDVI from a leaf-on Landsat ETM+ image among a number of ecological landscape classification systems for the forested area at the University of Michigan Biological Station (UMBS): landscape ecosystem classification, glacial landform maps, soil texture map and vegetation cover type map derived from ground-based survey. The NDVI values were summarized for each category within a particular landscape classification system, and then variation across these categories was analyzed using descriptive statistic measures, the F-test and multiple comparisons procedures. Even though leaf-on NDVI exhibits considerable variation across all classification systems, the most pronounced variability corresponds to ground survey-based vegetation cover type map. Major glacial landform and soil texture classifications also were sensitive to variation in NDVI. Interestingly, although NDVI overall considerably varied across landscape ecosystem types, many pairs of landscape ecosystem categories were not significantly different in mean NDVI. Aggregation of landscape ecosystems into more general hierarchical types eliminated some of the nonsignificant differences between categories but did not considerably improve the sensitivity of this classification to NDVI. Similar results were found when variation in another vegetation index – WDRVI (Wide Dynamic Range Vegetation Index) was compared among different classifications. These findings will have important implications for the choice of landscape classification for selection of ground-based sites for validation of remote sensing analysis. Western Snowy Plover Habitat Analysis at Coal Oil Point Reserve in Santa Barbara, CA Michele Tobias [email protected] The Western Snowy Plover (Charadrius alexandrinus nivosus) is a federally threatened shore bird that makes its home on the sandy beaches of the west coast of North America. One factor contributing to this species’ decline is loss of breeding grounds. This project looks at the historical change in Western Snowy Plover habitat at Coal Oil Point Reserve in Santa Barbara, CA, using aerial photographs that range from 1989 to 2004. Coal Oil Point Reserve has high quality breeding grounds rivaled by few other available areas in the birds’ range. Two specific factors are of concern in this study: overall available habitat area and habitat structure. On-screen digitizing and analysis of one photo from each pair using ESRI’s ArcMap will determine the overall habitat area, and on-screen digitized transects will determine specific areas of change. Three-dimensional analysis of each stereo pair using Leica Photogrammetry Suite will identify areas of changes in the habitat’s physical relief. Photos were restricted to only those taken within the months when the beach is most stable (May to October) to avoid including winter storm erosion events in the data. Variation in the size of the beach area is a given, but this study will determine whether there is a directional change (either increasing or decreasing) in the birds’ breeding habitat. Visual inspection of the photographs suggests that significant changes in area have occurred over time. Identifying Potential Pine Barrens and Wetlands in the High Plains Region of Northern Lower Michigan Stephanie Pendergrass [email protected] The Nature Conservancy of Michigan has identified the High Plains region of northern Lower Michigan as an important conservation priority that adds to the ecological and biological diversity of the Great Lakes Ecoregion. The unique physiography, surficial geology, and climate of the High Plains district have contributed to the development of unique ecosystems that support many endemic, rare, or endangered species. The project team is using elements of The Conservancy’s Five-S Conservation Planning process to develop a Conservation Plan aimed at conserving whole ecosystems rather than single species. Pine barrens and wetlands have been identified as the key conservation targets, and individual species are nested within these target systems. The locations of these targets need to be identified in order to assess their quality and develop strategies for their conservation and management. An 1800 “presettlement” land cover dataset was used to determine the historical locations and acreage of pine barrens and wetlands. One goal of the Conservation Plan is to conserve or restore a certain percentage of their historical extent, so current land use, land cover, and land owners/managers must be factored into the Conservation Plan. To determine the current locations and quality of existing or potential barrens and wetlands, we are conducting interviews with land managers, assessing the Michigan Natural Features Inventory’s (MNFI) database of species and community occurrences, and evaluating a “change grid.” This “change grid” is a raster dataset that depicts how much an area deviates from historical land cover, and it also contains information on the type of change between 1800 and 2000 (for example, from oak/pine barrens to aspen). The MNFI developed this dataset by overlaying an 1800 vegetation layer on the 2000 IFMAP landcover grid. Change levels include unchanged, minor change, natural or anthropogenic major change, unrestorable, and water. A set of appropriate change classes and land cover combinations was used to ascertain potential locations of current or restorable pine barrens and wetlands. These new layers have been overlaid with a land manager layer to determine whether potential locations are in public or private ownership. Land managers are being interviewed to assess the quality of any existing barrens and wetlands and the feasibility of restoration on other potential sites. Appropriate conservation strategies for existing and potential barrens and wetlands will be developed, and high quality sites, areas that cross ownership boundaries, or sites falling mainly on private land will be highlighted for special consideration. Groundtruthing of the potential barrens and wetlands identified through this analysis will be needed in order to implement any of the Conservation Plan. An assessment of nutrient trends for the River Raisin Watershed in southeast Michigan Leslie Ramirez [email protected] The River Raisin watershed drains an area of approximately 2780 km across five counties in southeastern Michigan before flowing into Lake Erie. It is home to numerous species of fish and aquatic invertebrates, as well as rare species of mussels. In 2000, agriculture accounted for approximately 65% of land use and growing ex-urban development is also contributing to changing land use practices within the watershed. Consequently, the river is heavily impaired with increased loadings of chemicals, nutrients, and sediments. These pollutants are products of human activities and pose a threat to the watershed, potentially degrading water quality and affecting critical wildlife habitat. Nutrient concentrations for nitrate and soluble reactive phosphorus (SRP) at 17 sites on the main stem and tributaries of the River Raisin obtained from a study conducted in 1996 were used in this assessment. The goal here was to compare this data to water quality standards developed by the Michigan Department of Environmental Quality (DEQ). First, several trends were observed during the analysis. Nitrate concentrations for all streams ranged from low undetectable levels to 18 mg/L; SRP ranged from 3.5 μg/L to 184 μg/L. Tributaries in the upper watershed showed lower nitrate concentrations whereas those in the lower watershed tended to have higher concentrations, especially during the spring and summer seasons. Nitrate also increased from headwaters to downstream on the main stem of the river. SRP concentrations generally followed the same trend from headwaters to downstream for the tributaries, however a significantly high peak (184.21 μg/L) was observed on Evans Creek. The main stem showed seasonal variation throughout the year but SRP was found to be at higher concentrations downstream than at the headwaters. Second, streams of high and low quality were identified based on this data. DEQ established a maximum contaminant level of 10 mg/L nitrate and a target of 0.30 mg/L (30 μg/L) total phosphorus for Michigan water bodies. Stream segments were scored based upon these standards and percent land cover was tabulated to relate nutrient concentrations to 1995 land use practices. Although non-point and point sources of pollution were not a part of this assessment, they also influence water quality in streams and affect the health of the watershed. 1 Castillo, M.M., Allan, J.D., & Brunzell, S. 2000. Nutrient concentrations and discharges in a Midwestern agricultural catchment. Journal of Environmental Quality, 29(4): 1142-1151.