A Method for Determining and Improving the Horizontal Accuracy of Geospatial Features

“A Method for Determining and Improving the Horizontal Accuracy of Geospatial Features” is the title of a paper I am writing and will be presenting at the 2012 South Florida GISExpo presentation. Other posts on this research:

2. Coordinate Sample Builder

Here is part of the abstract:

  • Many geospatial data sets stewarded by GIS professionals are based on cadastre (parcel) boundaries.
  • Many of these same data sets lack positional accuracy measures that could be generated by comparing these features to parcels.
  • A simple positional accuracy measure for linear features developed by Goodchild and Hunter (1997) is used to generate positional accuracies for polygons representing geographic features closely related to parcels.
  • The method is then extended to extract and replace lower accuracy polygons with higher accuracy polygons derived from parcels.

The South Florida Water Management District (SFWMD) regulates water supply, water quality, groundwater withdrawals, and surface water runoff through the issuance of permits for these activities on specific land parcels. In the images below you can see a large (left) and small (right) Environmental Resource Permit.

Our GIS consists of ~ 85,000 permit polygons

  • 40,000 Environmental Resource Permits that never expire
  • 45,000 Water Use Permits that are valid for 20 years

These are used by engineers, hydrologists, and environmental scientists to make informed decisions during the application review process and post permit compliance.

We have been digitizing permits since the 1980’s and during that time we have used many different base maps.

  • 1980 to 1987 – USGS 1:24,000 Quad Map (7 years)
  • 1987 to 1995 – 20 Meter multi-spectral and 10 Meter SPOT panchromatic imagery (8 years)
  • 1995 to 1999 – 1 Meter USGS DOQQs (4 years)
  • 1999 to Present – Cadaster/Parcels in select counties

Which basically boils down to having 15 years of pretty awfully data. Since 1999 we have been manually “correcting” our older permits to parcels  but we have not made this task into a project (too many other things to do) instead we fix them when we run across them. In order to speed things up we decided to automate and in the following posts I will describe how we are going about this automation.

If you would like to view the permits on Google Earth go to our e-permitting page and for an “Application #” enter “120117-3” you should get the screen shown below where you can click on the “View in Google Earth” link. If your on a Mac you will have to File – Open the downloaded KML.

SFWMD Help for Viewing Permits through Google Earth

From 1980 to 1995 we digitized permits with a less then optimal base map. We had 7 years of data digitized with USGS 1:24,000 Quad Maps, and then we went 8 years using 20 Meter multi-spectral and 10 Meter SPOT panchromatic imagery.

In 1999 we began working with parcels and at the time the 16 counties in our jurisdiction had their parcels at different states of accuracy. The large counties with longer established GIS operations like Miami-Dade, Broward, and Palm Beach had very accurate parcel data. In the less populated counties we had to tolerate a certain amount of parcel drift as these counties improved their GIS.

In addition, we were compiling the parcels into a seamless mosaic at the District but we were not subjecting the final output to a reliable QA/QC. On one yearly update we found that the entire parcel fabric for one county had shifted tens of feet and it turned out that the previous years data had been incorrectly projected. After this we started visually inspecting the parcels in a more systematic way by zooming in to 30 random locations and noting any abnormalities.

In 2008, when we began working on how to improve the accuracy of our permits the first thing that we needed to do was objectively quantify their accuracy against the parcels. We knew that we could get a RMSE by comparing the vertices of permits and parcels but to do this effectively we need a tool. More in future posts.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s