Mapping soil-landscape elements and the wetland in dambos and estimating CH4 and CO2 emissions from a dambo-terminated catena

Jerome Lugumira Sebadduka

Dambos are seasonally saturated grassy valleys mainly found on the central African plateau. They are usually sub-divided into three catenary units - gleyed and frequently inundated bottoms, flat and rarely inundated floors, and sloping sandy margins - fringed at the interfluve by the well-drained uplands. Since dambos constitute ~11% of Africa's arable land, soil information is required to guide sustainable use of the land. Further, it is important to determine the extent of the wetland environment in these landscapes so as to avoid miss-use of the land, which could arise because of varying definitions of the wetland in these landscapes. In addition, lack of knowledge about the true nature of dambo wetlands limits our understanding of their greenhouse gas (GHG) source and sinks strengths, which prevents projection of future GHG scenarios accompanying dambo use. This study was conducted so as to address these inadequacies, and is guided by the following specific objectives: (i) delineate dambo soil-landscape elements using aerial gamma-ray and terrain data; (ii) characterize a dambo wetland; and (iii) determine CH₄ and CO₂ sources and sinks in a dambo landscape. The area Hansen et al. (2009) studied was revisited. For objective 1, their model training and validation data were used. For objectives 2 and 3, data (e.g., soil, water table and gas samples) were collected from experimental plots in the four landscape positions, geographically constrained around dambo bottom pixels. Data used were collected during the main (March to July 2008) and short rainfall (October and November 2009) seasons in the area. An ANOVA analysis showed landscape position to have a proportionate influence on the variability of eU (46%), K% (28%) and eTh (27%); owing to the differences in soil properties along dambo cross-profiles. The results based on random forests (RF) and multinomial-ISODATA modeling, where gamma-ray and derivatives of a Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM) were used to classify dambo catenary units, were accurate, but only slightly better than the method which made no use of gamma-ray (e.g., conditional inference tree) . It was concluded that dambo landscape elements can be mapped by using these two data sources; although terrain data provides more information. Based upon a combination of hydrology and soil properties, dambo bottoms were the only element shown to constitute the dambo wetland. This zone is inundated for at least three-quarters of the main rainfall season and soils are hydric. Using the landscape map created by Hansen et al. (2009), the wetland was found to constitute only ~15% of the dambo. This is smaller than what was mapped by FAO-Africover and the Department of Survey and Mapping, Uganda (DSM). The wetland was also found to be the main source of CH₄ and sink of CO₂, with additional strengths attributed to the neighboring floor. Given that these constitute less than 20% of the landscape, dambo net contribution to the regional CH₄ budget is insignificant because 80% of the landscape is a sink. The worry, though, is the ongoing degradation, with the impact this has on the release of CO₂.

Mapping soil-landscape elements and the wetland in dambos and estimating CH4 and CO2 emissions from a dambo-terminated catena

Files and links (1)

Abstract

Metrics

Details