DRIVERS OF THE BIOGEOCHEMISTRY AND FUNCTIONING OF THE AFROMONTANE-SAVANNA MARA RIVER, KENYA

Abstract

In Afromontane-savanna rivers, data on the role of land use, stream size and large mammalian herbivores (LMH, both wildlife and livestock) and their interactions on nutrient cycling, dissolved organic matter (DOM) composition, and ecosystem metabolism are still limited. This study investigated the role of land use, stream size LMH and their interactions on the biogeochemistry of the Mara River, Kenya. Specifically, the influence of land-use change, stream order and density of LMH on water physico-chemical, quantity and composition (quality) of DOM and ecosystem metabolism in the river was investigated. A total of 82 sites were selected for sampling during the beginning of the dry season in January 2018 for nutrients, dissolved organic matter (DOM), and modelling of whole-stream/ ecosystem metabolism. Sites were grouped into five broad categories defined by land use and density of LMH: forested (19), agricultural (26), low-density livestock (15), high-density livestock (12), and wildlife (i.e., hippopotami- hippos) (10) sites. There were significant (PERMANOVA, p<0.05) spatial variations in water physico-chemical parameters, nutrients, DOM quantity, and measures of ecosystem metabolism among land uses and not among stream orders. Sites influenced by LMH had high electrical conductivity, temperature, and dissolved oxygen levels (p<0.05). LMH inputs had varied effects on nutrients, ammonium and DOC associated with hippos and high-density livestock streams (p<0.05). Total dissolved nitrogen (TDN) and nitrates were higher in agricultural sites and low livestock density streams than in the high livestock and hippo-influenced sites. Hippos and high-density livestock streams had high aromaticity and high molecular weight DOM (p<0.05). In contrast, low-density livestock and agricultural streams had photodegraded and low molecular weight DOM. The ratio of gross primary production (GPP) and ecosystem respiration (ER) was greater than one (GPP/ER >1) in low-density livestock and agriculture sites. At forested sites, the ratio was less than one, which is an indication of heterotrophic conditions. The findings of this study show that irrespective of their location on the fluvial continuum, LMH strongly influence the biogeochemistry of Afromontane-savanna rivers, which overrides the influence of stream size on the physico-chemical, DOM composition and ecosystem metabolism in the Mara River. Therefore, physico-chemical, DOM composition and ecosystem metabolism must be included in continuous monitoring for better management of LMH density in Mara River. In addition, pre-existing riverine models need to be updated to suit the Afrotropical savanna river functioning and biogeochemistry.