Gathering data

To begin this project we need to have data (figure 1) on our area of interest. The data we are interest in will be for Trempealeau County, Wisconsin and the surrounding area and will be associated with varying aspects of frac sand mining. This will require gathering data from several different sources which will result in varied formats and procedures associated with gathering and implementing the data. Some of the data of interest are land use, railroad lines, local geology and soils and the location of sand mines.

The first data we downloaded was on the railroads from the national atlas website (http://www.nationalatlas.gov/). The file was an eOO format, railroad map for the 48 conterminous states. The eOO format is an arc info coverage map. Next we visited the United States Geologic Survey (USGS) national map website (http://nationalmap.gov/viewers.html.) where we obtained land cover data for Trempealeau County and the national elevation dataset. The elevation dataset is a 1/3 arc second arcgrid digital elevation model. It was received as two individual tiles that we were able to mosaic together in arcmap. Next we went to the United States Department of Agriculture’s geospatial data gateway (http://datagateway.nrcs.usda.gov/) where we gathered data on cropland for Wisconsin. Last we used the Natural Resources Conservation Service SSURGO (http://soildatamart.nrcs.usda.gov/) site to gather soils data. All of these data were received as zip files which were saved into a project file, here they were unzipped. After all of the data were downloaded and unzipped we opened Arccatalog where we built a new file geodatabase and imported the data. The SSURGO data was a problem here. This data came as a database not compatible with arc. To overcome this we used Microsoft access to read the file into arc. Now that we have all of our data in Arccatalog we can project the data into the most appropriate format for the area of interest. I chose to use UTM zone 15N for my data due to the majority of the data being centrally located in this region. I did not use the state system because of the location of Trempealeau County and the surrounding area in relation to the borders of the state system.   

Fig.1. Data layers gathered from various sources for use
in our study of the effects of frac sand mining on local roads. 

    

The next step was to locate the mines. We have record of about 120 locations of mines or other related facilities that are either in use or proposed obtained from Wisconsinwatch.org. To minimize the time to locate all of these locations we formed teams of three to four people and divided the locations among us. We began by removing all of the sites with good addresses that should geocoder. We imported a Bing baseman with road names. Then we attempted to use a geocoder to locate the rest of the sites, by connecting to a local server. The first attempt yielded what appeared to be good results, of the 29 addresses that I began with I had 22 matches and 7 that did not match. However, on analysis all of the site were placed within a town. We went back to our excel table and attempted to normalize the address data, but when we ran the geocoder again we got a similar result. After doing some research we found a fusion table with latitude and longitude locations for all of the mines. We downloaded the table saved it in an excel format and added it to Arcmap as an event theme. We then had to save it to Arcmap as a feature class and set its projection, now we were able to use this as a reference in locating the sites. To locate the sites we used the review/rematch tool in the geocoder. We zoomed to each individual site to verify it from our original table, then we would zoom out until we found a mine, when we found one we could verify its location with either its street address or by the PPLS system, which we overlaid on the map by connecting to a remote server.  This process took several hours to complete. After completion, I was able to identify 35 of 39 sites (figure 2); the remaining four sites did not have enough information to locate them. There may be some uncertainty due to a large number of sites being proposed and not completed and also outdated aerial imagery. 

Fig.2. This map, with a Bing base map, shows the location of all 35 sites that
 I was able to match. The symbolization of the sites varies because the
 matching was done in two smaller groups. 

Fig.3. The data table of all the sites I was involved in locating. I matched
35 sites while 4 had insufficient data to make a match.