Using Advanced Spatial Technologies for Gorilla Habitat Analysis - DFGFI
For the past ten years, Dian Fossey Gorilla Fund International (DFGFI) scientists and collaborators have been studying the habitat and behavioral ecology of the mountain gorilla (Gorilla beringei beringie), using images of the Virunga environment taken from space, Global Positioning System (GPS) units to record gorilla movement and human activity, and historic maps. These data are then combined in a computerized Geographic Information System (GIS) to display and analyze the data for research and conservation purposes. DFGFI’s work is a pioneering application of these tools for primatology, and it has great potential for other conservation work throughout Africa and the world.
Africa remains the most poorly mapped continent on Earth. Less than 15% of Africa has been mapped to modern standards, and this problem is made even worse in the Virungas, due to its unique climate, geography and history. This remote and cloud covered region is where three very different African nations with different colonial histories meet, and it is the most distant portion of their borders. To the gorillas it is a single place, their only home; but to politicians, cartographers, and governments it is a place where distant borders end in an unstable and dangerous world. In many ways, it truly is the ends of the Earth.
Ten years ago at Rutgers University with a small USAID grant to DFGFI for the creation of mountain gorilla habitat maps. Dieter Steklis, a primatologist at Rugers and then director of DGFGI’s Karisoke Research Center in Rwanda, contacted his Rutgers colleague Scott Madry, who was teaching remote sensing, GIS, and aerial Sensing and Spatial Analysis, to see if we could produce some basic maps of the entire Virunga Conservation Area.
First, DFGFI needed to find out what existing maps were available for the three countries (Rwanda, Uganda, Democratic Republic of Congo) that compromise the Virunga Conservation Area. Several problems quickly became apparent: No maps of Zaire (now the Democratic Republic of Congo) were available at all, as there was a civil war underway in the region and possession of topographic maps or aerial photographs was politically highly sensitive. Unfortunately, the Zaire area accounts for more than half of the gorilla habitat. The best maps of Rwanda and Uganda were made in the 1950s and '60s during the colonial era. Even worse, it was clear that the Rwandan maps (produced by the Belgians and French) and Ugandan (British) maps were the products of their respective traditions and used different mapping projections, coordinate systems, and (worse yet) did not seem to match up at all when put together. Creating a unified regional map was clearly going to be more difficult than it first appeared.
Next, Scott Madry contacted a former graduate student from Belgium, to see if there were any historic maps that could be useful in filling in the missing Zaire (DRC) areas. He contacted the Belgian Royal Museum for Central Africa in Turvuren and eventually DFGFI received a complete set of 1938 1:100,000 and 1:50,000 maps of the region that had been produced by the Belgian Colonial Service. These maps covered the entire Virunga volcano range and surrounding area. These maps also served an important purpose of showing the region before modern development and population encroachment. These old maps came to serve as a "base map," which is still the most recent comprehensive map of the entire area with significant detail.
It soon became apparent that a digitized base map could serve as an important foundation for a variety of more sophisticated analyses concerning the relationships between the characteristics of the gorillas' environment and their behavior. GIS technology is a perfect tool for the DFGFI to meet their objectives. GIS can be thought of as a database of computerized maps, but it is much more. It allows each class of feature (such a roads, hydrology, political boundaries, vegetation, etc.) from different maps to be used together to create separate layers of digital information. This allows data from many different sources, scales, and dates, to be re-combined in a powerful display and analytical environment. Over the years the DFGFI has gathered an array of sources of geographic information, including regional maps, remote sensing images, GPS data, and historic records.
A first objective was to create a three-dimensional base map. In order to complete this DFGF digitized the various features from existing maps and combine them together in a GIS. Each contour from the 1938 Belgian maps was manually digitized to create what we call a Digital Elevation Model, or DEM. This allows us to visualize the Virunga volcano region in three dimensions. Main roads, political and park boundaries, hydrology, and other map features that were available and would be used in future analyses of human impact on the habitat were also added to the map.
'Remote sensing imagery' describes photos taken from afar, such as aerial photographs and satellite images. In order to generate accurate land use and vegetation data for the entire region, and to track the changes over time, satellite images were used. There are many environmental satellites that have been orbiting the Earth daily since the 1970’s, and search of all available satellite imagery sources was done quickly by DFGFI. Unfortunately, virtually all of the satellite images were cloud covered. Many were clear except over the Virungas, due to the effect that high mountains have on the local microclimate. Satellite radar data was the next choice, because these systems send their own burst of energy down to the Earth and record the reflected energy. Satellite radar systems can operate day or night, rain or shine. NASA was flying a research radar system on the Space Shuttle in 1994, so Scott Madry contacted former colleagues there (he worked at a NASA research facility), and arranged for the Virunga Conservation Area to be imaged during the shuttle flight. This produced the first cloud-free remotely sensed view of the entire region, which was used to create an initial vegetation map. The two shuttle flights in April and September of 1994 were during and after the terrible upheaval in Rwanda, so the DFGFI was able to record the deforestation and other effects of the many refugee camps that were near the Virungas.
Nick Faust (Georgia Institute of Technology), is a world-class specialist in remote sensing and image processing. He was asked to join DFGFI and has contributed enormously to their work in image acquisition and classification. His lab has focused on analysis of hyperspectral imagery of the Virungas acquired during a fly-over in 1999. Details of this collaborative project with Earth Search Sciences, Inc. and The National Geographic Society were featured in the documentary "Gorillas on the Edge." More recently, his lab acquired a nearly cloud-free Landsat satellite image in January 2003, and this has been used to create a more accurate and recent vegetation map of the region.
An important aspect of DFGFI work is technology transfer to the people of Rwanda. As a result of a grant from the Georgia Research Alliance, DFGFI in partnership with Georgia Tech, Clark Atlanta University and the National University of Rwanda, has established a GIS Center in Rwanda, with emphasis on remote sensing technology. Scientists from the Center and DFGFI are using these new tools not only for mountain gorilla research, but also for other applications in Rwanda. This will make it possible for the Rwandan people themselves to harness these powerful tools to create a better future.
The GPS is a system of satellites that allows researchers to determine their position, altitude, and time anywhere in the world using a small hand-held receiver. The DFGFI received a donation of Trimble GPS receivers and trained DFGFI field researchers and staff in their use. Additional funding from the MacArthur Foundation and the Daniel K. Thorne Foundation allowed the DFGFI to purchase more GPS units. Gorilla movements were logged, including nesting locations over extended periods of time. Equally important, the Karisoke anti-poaching patrols also recorded evidence of poaching activity. The GPS units were also used to map park boundaries and other major features in the area. This work is ongoing, and provides the DFGFI with the ability to include field observations directly into the GIS.
While GPS data is being used to record recent years of gorilla movements and poaching activity, to understand patterns over time, historical research must be incorporated. Netzin Gerald Steklis, Director of the DFGFI Scientific Information Resources, has been an invaluable member of the DFGFI team, providing archival maps, photographs, field notes, and other relevant data acquired in the field by numerous researchers, including Dian Fossey's hand- drawn maps of daily gorilla group locations. Fossey's gorilla ranging maps, along with long-term records from our Karisoke trackers, are being digitized and will be used to compare the current gorilla ranging patterns with those of 20 years ago. The DFGFI historic database is one of the longest set of primate observations in the world, and hence ideal for revealing long-term patterns of habitat use.
The goal of the DFGFI is to understand the dynamic relationship between mountain gorillas and their habitat over time. The DFGFI will try to understand the changing patterns of gorilla behavior, their relationship to their environment, and to quantify and understand the impacts of poaching and encroachment. All of this work is designed to support the larger scientific and in-country capacity building goals of DFGFI. GIS allows data such as: historic maps, aerial photographs, field results, GPS locations, satellite images, and more, to be combined so that DFGFI can begin to explore the data and ultimately, test hypotheses in ways that were not possible only a decade ago. Another significant benefit of the GIS database lies in the cumulative nature of DFGFI's field research. As field patrols continue and new research is conducted, the data are entered into the system, constantly enlarging the database, thereby allowing a comparison of data over time. Hypotheses can be tested, and patterns can be tracked in ways that were not possible using traditional field notes and hand-drawn maps.
This project has stretched the boundaries of the application of advanced technologies for regional primatological research. It is being conducted in an extremely remote and uncharted region of the world in the face of great political and social, unrest. It has taken the collaborative effort of a team of dedicated people with different skills working together for a common goal. The DFGFI is confident that their efforts will make an important contribution to the understanding of mountain gorillas and their habit.
Madry, S. et al. “Using Advanced Spatial Technologies for Gorilla Habitat Analysis.” The Dian Fossey Gorilla Journal Fall 2003.