Testing the Accuracy of DEMs in GIS

The first few weeks have flown by here in Williamsburg! The project is fully underway and progress was slow at first but with my methods smoothed out and the data downloaded I’ve been able to pick up speed. As mentioned in my introductory post the majority of my research is being done using Geographic Information Systems (GIS). Beginning the next few days I will be analyzing numerous agricultural fields in the Coastal Plain.

However, before I could start I needed to be sure the data I would be using was of high enough resolution to accurately determine runoff patterns on such a small scale. To determine this I used very high resolution LIDAR (Light Detection and Ranging) which was available online for the state of Maryland. Using these 2m resolution Digital Elevation Models (DEMs) steps in GIS were taken to determine flow patterns on 3 fields and accumulation of flow at the field margin. These steps were also used to analyze the same 3 fields however using lower resolution 10m DEMs available on the USGS Seamless website. Once I had determined flow patterns and flow accumulation I was able to analyze the results of each DEM resolution in Excel. Graphs of cumulative fraction of field drained vs. distance around the field were created showing both the 2m and 10m series on the same plot. Immediately it was clear that the two DEMs matched exceptionally well, with spikes in fraction drained and fraction entering the fields very close in distance value and fraction value. I am currently working on a method to quantitatively determine the level of fit the plots possess as well as a detailed description of methods used to get a head start on my thesis writing.

This process of testing the 10m resolution DEMs allows the project to confidently determine drainage patterns on the fields of study, knowing that higher resolution data gives very similar results. Also, my project will be looking at flow through a channel in field in Charles City County. The past few weeks I have practiced and adjusted the methods I will be using to determine channel discharge on streams around campus. This method utilizes the addition of a tracer to the flow and the dilution of the tracer downstream. I look forward to using the method during the next large storm and the analysis of more fields in GIS throughout the next week.