Nitrogen Fixation and Denitrification in Sediments of Two Kenyan Lakes

Abstract

Parallel measurements of nitrogen fixation using acetylene reduction assays and loss of denitrified nitrogen using acetylene inhibition assays have been made upon detrital material from two tropical Kenyan lakes. The mean range of N2 fixation was 0.224 to 4.48 ng N g (dry weight) -1 hr-1. Assuming a 2 cm interaction depth these values are low to very low compared to other studies. They are about four orders of magnitude lower than the only other, but indirect, previous estimation for one of the lakes. Denitrification rates ranged between 20.1 and 1160 ng N g (dry weight)-l hr-1. These exceed the N2 fixation values by more than two orders of magnitude, but are within the ranges given in other studies. FOR TROPICAL AFRICAN INLAND WATERS there has been much interest in nitrogen as a limiting nutrient for organic growth (e.g., Moss 1969, Prowse and Talling 1958, Viner 1975, further references in Talling and Talling 1965). Blue-green algal nitrogen fixation can be of great importance to the nutrient budgets of lakes (Ashton 1980, Horne and Viner 1971), and in subtropical southern Africa there is evidence for a shift from P-limited to N-limited conditions due to man's influence (Toerein et al. 1975, Walmsley and Ashton 1977). Although nitrogen dynamics clearly are of considerable ecological and economic importance for African waters, there are very few relevant studies. Bacterial fixation of N2 into an ecosystem and the reverse process of loss of N2 by bacterial denitrification are virtually unmeasured for any of the continent's aquatic ecosystems. Yet the environmental conditions of high organic content, substantial dissolved nitrate, low or absent dissolved oxygen, and optimal temperatures which would encourage these processes are common, particularly around the margins of lakes where fringing communities can markedly influence the economy of the entire water body. There is only one other comparable study, that of Gaudet (1979) for the swamps of Lake Naivasha, which is an indirect assessment of the rate of nitrogen fixation by subtractioin of other components of a gross nitrogen budget. These present results report an attempt at more direct estimations of lake detritus nitrogen fixation and of denitrification by acetylene reduction and inhibition, respectively. As such, they are the first for Africa. They were carried out on Lake Naivasha (115 kMi2, mean depth 4.6 m, conductivity in ,uS cm-' 353) at a similar site as used by Gaudet (loc. cit.), and upon sediment from near the north bank of Lake Sonachi which is small (0.18 kiM2, mean depth4 m) and highly saline (u-S cm-l 62709030 in surface water) and immediately to the west of Lake Naivasha. The lakes are at the bottom of the Kenyan part of the East African Rift Valley (0?47'