Linking land use to stream pollution : pollutant dynamics and management implications

Abstract

The Njoro River Catchment, Kenya was systematically studied between February 2005 and May 2008. The overall objective of the study was to link current land use/cover to stream pollution. Eleven sampling sites along the stream were routinely monitored. Stream flow was intermittent and seasonally periodic with complete cessation of flow at the lower reaches during dry periods. Pollution levels were spatially and temporally variable. Analysis of the net flux of pollutants at a reduced spatial scale identified the densely settled and heavily farmed areas around Egerton University and Njoro Township as the most vulnerable sub-catchments. The least vulnerable was the predominantly forested Upper Njoro River Catchment. The principal sources of pollutants were WTP outfalls, surface runoff from farmland and the in-stream activities of people and livestock. The potential health risk (PHR) for bathing and drinking at water abstraction points along the stream exceeded acceptable health risk (AHR) levels. PHR was 2-3 times higher with the Cabelli (1983) Intestinal Enterococci (IE) model compared to the U.S. EPA (1994) Escherichia coli (EC) model. Reducing the current PHR levels of >4% of the exposed population to U.S. EPA bathing AHR levels of 0.7% (EC) and 1.9% (IE) would require at least 2-4 log reductions of IE and EC densities and further reductions to achieve the WHO recommended drinking water AHR level of 0.1% of the exposed population. It was concluded that the hydrology and water quality of the stream is critically affected by various land use types and the response of the stream to human interventions was principally governed by current land use/cover, the periodicity of activities and the seasonality of the hydrological regime.