[an error occurred while processing this directive] [an error occurred while processing this directive]
 

 

Visualizing Megatransect Data


ArcUser October-December 2001
 
logo

Megatransect leader J. Michael Fay returned from his 15-month expedition across central Africa with 300 hours of video, 100,000 numerical records, 100 hours of DAK tapes, and hundreds of pages of text. This data was Fay's effort to "quantify the stroll."

Michale Fay and  John Grayson

Esri staff member John Grayson and Michael Fay demonstrate the prototype GIS developed for the Megatransect data.

The Megatransect, like the Lewis Clark expedition across the interior of the United States in the early 1800s, was a scientific survey as well as an adventure. (See the accompanying article, "Real Hero Gives Keynote Address," for more information about the expedition.) Using a GPS unit, Fay captured location readings every 20 seconds. In numerous "Rite in Rain" notebooks, he continually recorded animal sightings, evidence of human activity, characteristics of animal dung, vegetation structure, and many other field observations. Video and audio recordings also documented the route and condition of animal populations.

Advances in technology made in the 200 years between the Lewis and Clark expedition and the Megatransect allow Fay to correlate his copious field observations with the very specific GPS location data he gathered by using GIS. The times recorded for field observations can be related to the time stamps for the GPS readings so that the location of each observation can be derived and mapped. The dynamic segmentation capability in ArcGIS makes this possible. Dynamic segmentation models linear features using routes and route events. A route is a linear feature that defines discrete locations along that feature using a measurement system. A route's geometry is a polyline with x, y, and m values. Data is modeled along the route using events.

While preparing for his presentation to the 2001 Esri International User Conference, Fay worked with several Esri staff members on a GIS using a small portion of the data collected during the Megatransect. This data describes the route through the Odzala National Park. This area of the Congo basin rainforest contains an exceptionally diverse ecosystem.

Over a period of a few weeks John Grayson, Witold Fraczek, and Lenny Kneller from the Redlands campus and Liz Sarow of the Denver regional office created a prototype GIS using ArcGIS. The team received two kinds of data from the Megatransect-field notes and GPS readings. The field notes Fay collected were entered in Microsoft Excel spreadsheets that included a field showing the time each observation was made. The GPS data was in decimal degrees.

Distance from populated places.

This map, generated using the ArcGIS Spatial Analyst extension, illustrates the distance from populated places. The white polygon shows the existing national park boundaries. The black line shows the Megatransect route.

Both types of data were processed to create a standardized time field so that field notes and location data gathered with GPS could be correlated. After identifying data outliers, the attribute data was averaged into five-minute increments. A polyline for the Megatransect route was generated with its m values set to the time. The polyline was parsed into five-kilometer segments, a standard linear measurement for studies of this type. Tables were created for each type of attribute data. This was done so that data for animal tracks, creeks, animal dung, human impacts, monkey observations, and vegetation structure could be analyzed separately. Each type of attribute data type was summarized for each five-kilometer segment. When the data was taken into ArcScene, each attribute (e.g., evidence of human impacts, monkey sightings) became the elevation factor for the route.

The Geostatistical Analyst extension was also used to model the relationship between areas of high elephant population density and high human population density. This analysis revealed an inverse relationship between the areas of density for the two attributes that was striking when the data was modeled in three dimensions. Adding the original park boundaries from 1935 and the present day boundaries that encompass an area four times greater than the original park provided another dimension for analysis. Photo and video events were also created to link multimedia files to locations on the route.

The GIS, illustrating some of the things discovered about the dataset, was demonstrated during Fay's presentation. It gave the audience a tantalizing look at the possibilities for modeling the dataset using ArcGIS. Fay said he hopes to continue working with the dataset and would like to make the results available on the Web to all types of users from schoolchildren to NASA researchers.

Learn how you can contribute to saving a portion of the Congo.

Table of Contents for the October–December 2001 issue

[an error occurred while processing this directive]