Colby College - Long Pond South Basin Report 51 ANALYTICAL PROCEDURES AND RESULTS GIS Introduction A Geographic Information System (GIS) is a combination of computer hardware and software that uses spatially referenced and georeferenced information to create maps and models (Chang 2006). Spatially referenced information is data that are shown in the correct location in relation to other features on a map; by definition, data displayed on a map are spatially referenced. Georeferenced data are data that have been linked to the map by a location using a coordinate system such as latitude and longitude or a Universal Transverse Mercator (UTM). The UTM system divides most of the earth into 60 different zones, omitting areas that are above 84° N latitude and below 80° S latitude. The point of reference for each zone is the intersection of its central meridian with the equator. Points within a zone are located by a northing and easting coordinate pair, which refers to how far east and north a point is from the point of reference of the zone. The UTM system is often preferable to latitude and longitude when working with maps because it minimizes the distortion that occurs when a circular object such as the earth is represented on paper. Spatially referenced information is created by adding a layer on ArcGIS® 9.2 and then editing that layer to add the desired information, such as points, lines, or polygons. Common sources of georeferenced information include GPS points or a digitized pre-existing map (Chang 2006). A GIS can accomplish up to four different tasks: data input, data storage and retrieval, data analysis and manipulation, and data display. CEAT used ArcGIS® 9.2 software, published by ESRI, to create maps and spatial models. ArcGIS® 9.2 uses two forms of data called spatial data structures: vector and raster. Vector data are comprised of points, lines, or polygons. Vector data were used to determine land-use in the watershed of Long Pond South. Satellite images were digitized into polygons, with each polygon denoting the area of a different land-use. Raster data are obtained by transforming either vector data or photographic images into uniformly sized cells. Each cell in raster format is associated with a datum. Temperature gradients are anColby College - Long Pond South Basin Report 52 excellent example of a use of raster data. A raster layer of temperature gradients is made up of uniformly sized cells that are assigned a color based on the datum they are associated with, in this case a numerical temperature. Images are frequently used as base layers for GIS maps because they provide familiar reference points to viewers. However, they can also be transferred into raster or vector format to increase clarity or for use in data analysis. For example, it may be easier to see the location of streams when they are colored lines instead of looking at satellite photos of a large area. Transferring image data into vector or raster format can be helpful in data analysis because these formats are required for all quantitative analysis, such as determining land-use areas and generating models. A GIS layer contains data pertaining to the same theme, and must be of the same spatial data structure and type. Examples of layers include temperature gradients, land use classes, soil types, and roads. Each layer is associated with relevant information such as its name, location, and date sampled. The compilation of associated information contributes to the attributes of a layer, and the attributes can be viewed or edited. Most maps are a combination of several layers that are needed to present the desired information. Models are created by ranking the categories within a georeferenced layer based on the impact the categories have on what is being investigated. For example, in the erosion potential model categories of land use are ranked based on how susceptible each land-use type is to erosion. The ranked layers are then weighed and multiple layers are combined to create a new layer, which displays the output from the model. This method is analogous to finding a weighted average; the weighted average is the model. The models created on GIS include an erosion potential model, erosion impact model and a septic suitability model. Water and watershed studies require data that are typically georeferenced, allowing for a GIS to be used. GIS is a powerful tool because it allows the user to analyze information and visually represent complex data. The maps and models created as part of this study are designed to provide a generalized view of the watershed from which future improvements can be suggested.Colby College - Long Pond South Basin Report 53 WATERSHED LAND-USE PATTERNS Introduction A comprehensive survey of watershed health, status, and trends must identify existing land-uses and quantify their effect upon water quality. Each land-use impacts the watershed to a different extent based on surface sensitivity to runoff, percentage of impervious surface, and level of nutrient-loading activity; the combination of these land-use qualities generates unique erosion properties that ultimately influence nutrient flux in the receiving water body. For example, those land-use types characterized by dense persistent vegetation (e.g., coniferous forest, deciduous forest, or mixed forest) absorb rainfall and reduce the erosion generated by runoff. The root structure of vegetation lowers the erosion potential of the underlying soil by stabilizing it. Due to the water absorption and erosion prevention provided by highly vegetated coverage, these land-uses add minimal amounts of nutrients to the associated water body. On the other hand, land-uses with minimal vegetation (e.g., commercial land, parks, pasture, or residential land) increase the impervious surface of the watershed and destabilize surrounding soil, increasing runoff potential and erosion. These land-use types contribute more nutrients to the receiving water body (Dennis 1986). When conducted across time, land-use surveys provide valuable insight into historical land patterns, which may be used to predict future trends in land activity. The current land-use survey compares Long Pond South watershed data collected in 2003 to data from 1966. By identifying land-uses and analyzing their modification throughout time, this study elucidates historical land-use trends and facilitates future predictions