It Must Be Something in the Water

Exploring the Link Between River Chemistry and Land-Cover Traits in the Congo River Basin


  • GIS was used to identify major waterways through a diverse set of land-cover types.
  • ArcGIS brought together the sample points/locations and the spatial data on the fly.
  • With ArcGIS Online, researchers can share and distribute commonly used data.

Scientists from the Woods Hole Research Center (WHRC) embarked on an expedition to collect water data from the Congo River Basin, the world's second-largest river system and one of utmost importance for understanding the global carbon budget. Greg Fiske, WHRC's GIS manager and researcher, was part of the team. He was there, armed with vast quantities of spatial data, to ensure that the team stayed on the route during the expedition and to contribute to the sampling goals.

ArcGIS was used to monitor dissolved organic carbon along the Congo River prior to and during this trip.

ArcGIS was used to monitor dissolved organic carbon along the Congo River prior to and during this trip.

Supported by the National Science Foundation, the Global Rivers Project is a collaboration of several institutions around the world and focuses on six globally significant river systems: the Congo, Yangtze, Brahmaputra, Ganges, Kolyma, and Fraser. Fiske contributes his GIS skill and expertise to work with an array of scientists, including geologists, geochemists, hydrologists, engineers, and remote-sensing experts, to explore the relationship between river chemistry and large-scale land-cover characteristics.

Within the Republic of Congo, the team traveled by four-wheel-drive truck on a southwest–northeast transect, covering 1,400 kilometers (roughly the distance from Massachusetts to North Carolina) and enduring rough terrain, poor roads, insect infestations, and days of soggy weather countered by days of stifling heat. November is the rainy season in the Congo, which was the main reason the trip was planned for that time. High-flowing rivers and wetlands at their peak created an ideal contrast between the sampled water chemistry from this trip and that from previous excursions taken in the dry season.

One of the key measurements was dissolved organic carbon (DOC), which is a general term for those organic components dissolved in marine and freshwater ecosystems. It is a key indicator of land disturbance and land-cover changes worldwide. As a doctor may take a sample of your blood to divulge information about the health and well-being of your internal systems, so too can scientists characterize the conditions of the land within a watershed by taking detailed chemical samples of rivers and streams. And DOC isn't alone. Scientists have an elaborate quiver of water chemical metrics that may define land-related changes within our world's major watersheds. In addition to DOC, water samples were analyzed for a variety of other constituents, including dissolved and particulate forms of nitrogen and phosphorus, as well as temperature, salinity, pH, and a selection of dissolved gases.

As with any statistical modeling, diversity is important in the sample set. GIS was used to find easily accessible major waterways where the areas of contribution (upstream watershed) covered the most diverse set of land-cover types available in the Republic of Congo. Latitude-longitude coordinates for each sample point were captured using Garmin GPS and mapped regularly.

The Congo River Basin has the largest swamp forest in the world. It is where the team spent the majority of time collecting data. Team members traversed these forests in a pirogue (wooden dugout canoe), sampling water along the way. Despite the hardship of the journey, they were happy to encounter a variety of land-cover types: grasslands and croplands in the south; sparse forest areas in the nation's midsection; dense humid forests in the north; and, finally, swamp forests in the northeastern area of the country.

Over the side of a pirogue, researchers collect water samples in the swamp forest.

Over the side of a pirogue, researchers collect water samples in the swamp forest.

The team's goal was to collect data on each land-cover type in the basin, as well as samples from tributaries that feed those areas. In preparation for the trip, Fiske loaded spatial data onto his laptop, which the team used for indicating specific land-cover types that were important to the analysis. He also created some GIS scripts to assess the upstream area and produce a selection of land-cover metrics within that area based on key remotely sensed GIS layers.

At the end of each day, Fiske uploaded the geocoded water sample data to his laptop, running ArcGIS software. He overlaid water-sample and land-cover type attributes on the remote-sensing data so the team could immediately see the results of its work and affirm that it was in the appropriate location. On the fly, GIS brought together the sample points/locations and the spatial data. GIS displayed an area's tree cover, its biomass/carbon, high-resolution natural color imagery, and more. For the Congo area, important data layers include the percentage of swamp forest and seasonal inundation. (Other watersheds around the world may need different GIS data layers.)

Fiske also made good use of ArcGIS software's hydrology toolbox for a lot of the work on the project. This made work in the field easier. Using these tools and the custom scripts, at the river's edge, the team could click a button and query the GIS about the size of the upstream contributing area or the percentage of tree cover or other key land-cover types.

The challenge that made mapping difficult, and sometimes made fieldwork impossible, was the intense rain and the river waters. Fiske and other members of the team were constantly concerned about dropping gear into the river or losing it in the bottom of a flooded pirogue. Sensitive gear was protected in waterproof bags and hard cases. Data was backed up on a USB-powered external hard drive that was secured in a watertight case.

The team had other duties besides data collection. As part of the National Science Foundation grant, some project funding is designated for outreach and education. Building on a successful partnership with schools and communities in the Russian Arctic, the team visited schools in two communities in the Congo River Basin on this trip, and art supplies were distributed to the students. The children were asked to put pen to paper and describe, through art, the importance of their local river. Their pictures and drawings will become part of the My River, My Home exhibition, joining nearly 100 other pieces from students in Siberia and Canada. The exhibition will travel to galleries, and a virtual collection will be posted on the WHRC website.

In addition, the team worked with local people and trained them to collect water samples and metrics. This would allow the scientists to retrieve critical time-series information on the tributaries of the Congo. WHRC staff and partners worked with students, teachers, and community leaders who live in the towns and villages within watersheds, and they built partnerships for sustainable solutions.

Upon returning to the United States, the team provided data and samples to associates who would use them to further study the attributes and relationships within the basin. Collectively, the members hope to reveal just how important dissolved organics and other key elements can be in detecting the impact that humans have on the land of the Congo River Basin. Their findings will provide the basis for understanding the health of a watershed and directing future watershed management.

Fiske is currently the GIS adviser for the Global Rivers Project. Using tools such as ArcGIS Online, he has been able to share data with others and distribute commonly used base layers. He chose ArcGIS Online because it is free and user-friendly, so even those with no GIS experience can use it. He posts project-wide base layers—such as stream networks, sample locations, and watershed boundaries—to ArcGIS Online so that scientists and others can download maps and data directly into their own GIS projects. It is essential that the various river projects all use the same version of these layers. The ArcGIS Online map viewer is also used to make maps of sample locations and other important layers that are shared among colleagues.

Fiske and others are also designing spatial models to show correlations of certain land-cover types and water chemistry variables in hopes of being able to extrapolate the rules defined by the Congo analysis to other portions of tropical Africa—places that would be much more difficult to go to and physically sample.

For more information, contact Greg Fiske, research associate/GIS manager, Woods Hole Research Center (e-mail:

Contact Us | Privacy | Legal | Site Map