Data Gathering and Quality Assessment

- Introduction -

To analysis quartz sand mining in Wisconsin, six datasets will be downloaded from websites, geospatial gateways and online map viewers. This data includes: transportation (railways), landcover, a digital elevation model (DEM), cropland, a soil survey, and a Trempealeau county geodatabase containing various land record information for the county. Each dataset originates from a different source leading to a difference in projections, data accuracy, and metadata documentation. In this post, an overview of the data collection process is presented along with a quality assessment of the associated data.

- Methods -

Obtaining the data:

• Land records for Trempealeau County were downloaded from the Trempealeau County Land Records Department's website as a geodatabase. This geodatabase was renamed TMP.gdb and will be used as the primary location for all data gathered for this analysis. As such, each of the following datasets will need to be projected in the same projection used in the TMP.gdb and clipped to the Trempealeau County boundary.

• An updated railroad polyline shapefile was downloaded from the United States Department of Transportation Bureau of Transportation Statistics website.

• A landcover raster and a 1/3 arc Second DEM in ArcGrid format was downloaded from the National Map Viewer.

• A cropland raster was downloaded from the USDA geospatial data gateway.

• SSURGO data was downloaded from the USDA NRCS Web Soil Survey application as a polygon shape file.

Each shapefile was imported into the TMP.gdb and given the proper projection and clipped by the Trempealeau County boundary feature class. Each raster was projected, clipped(extract by mask), and imported into the TMP.gdb via a python script. The exact python script and discussion of python use can be viewed in my Python Page by clicking here or the Python tab at the top of the page.

- Results -


Map 1: Comparison of the data downloaded for the sand mining analysis

- Discussion -

Data accuracy assessment:

To assess the accuracy of the data, eight different measures of accuracy were collected from the metadata files associated with each data set. These attributes were recorded in Table 1 and include: source, scale, effective resolution, minimum mapping unit (MMU), planimetric coordinate accuracy, lineage, temporal accuracy, and attribute accuracy.


Table 1: Accuracy comparison of the data downloaded and imported into the TMP.gdb


The data originator was designated the source. Data sourced from accredited organizations is preferred because a certain level of quality is implied (though it is still important to check). For lineage, only the sources were recorded for sack of table space, though the processes also need to be considered. For temporal accuracy, either the range of data collection or the last update was recorded. For attribute accuracy, only a broad statement found in the metadata was recorded, not the entirety of the text. For some data sets, certain measures of accuracy were missing such as effective resolution, scale, planimetric coordinate accuracy, and minimum mapping unit. In these instances Table 2, Table 3, and Table 4 were consulted in an attempt to infer some of the values. In some cases the values could not be found or inferred and the value of N/A was recorded.



Table 2: Relationship between scale, MMU, and effective resolution for raster data. (Nagi)
http://blogs.esri.com/esri/arcgis/2010/12/12/on-map-scale-and-raster-resolution/

When attempting to infer values, the definitions of scale, effective resolution, MMU, and planimetric coordinate accuracy and their relationships needed to be understood. Scale is the relationship between map distance and real world distance. The scales recorded in Table 1 are representative fractions. Effective resolution is commonly denoted as the square root of the area of one pixel of raster data (30x30m pixel size equals an area of 900m^2 making the effective resolution 30m). MMU is the size of the smallest depictable feature for raster data and the size that features will retain their geometry for vector data(e.g. polygon to line or point, or whether polygons will merge or remain separate). The MMU of raster data is commonly 4 pixels. Planimetric coordinate accuracy refers to the root-mean-square (RMS) error corresponding to how well features are positioned and represented in relation to the real world.

Table 3: Relationship between planimetric accuracy(ft) and scale. (ASPRS)
http://www.asprs.org/a/society/committees/standards/1990_jul_1068-1070.pdf



Table 4: Relationship between planimetric accuracy(m) and scale. (ASPRS)
http://www.asprs.org/a/society/committees/standards/1990_jul_1068-1070.pdf


- Sources -

American Society of Photogrammetry and Remote Sensing (ASPRS). ASPRS Accuracy Standards for Large-scale Maps (1990). Accessed via web: http://www.asprs.org/a/society/committees/standards/1990_jul_1068-1070.pdf

Nagi, Rajinder. On Map scale and raster resolution (2010).  ESRI blog. Accessed via web: http://blogs.esri.com/esri/arcgis/2010/12/12/on-map-scale-and-raster-resolution/

Trempealeau County Land Records. Geodatabase. Accessed via web: http://www.tremplocounty.com/landrecords/

United States Department of Agriculture National Resource Conservation Service. Cropland raster. Accessed via web: http://datagateway.nrcs.usda.gov/

United States Department of Agriculture National Resource Conservation Service. Soil survey shapefile. Accessed via web: http://websoilsurvey.sc.egov.usda.gov/App/HomePage.htm

United States Geological Survey, National Map Viewer. Landcover raster and DEM. Accessed via web: http://nationalmap.gov/viewer.html

United States Department of Transportation. Railway shapefile. Accessed via web: http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_atlas_database/index.html