Characterization of Potential Threats from Cyanobacterial Toxins in Lake Victoria Embayments and during Water Treatment

Abstract

Africa’s water needs are often supported by eutrophic water bodies dominated by cyanobacteria posing health threats to riparian populations from cyanotoxins, and Lake Victoria is no exception. In two embayments of the lake (Murchison Bay and Napoleon Gulf), cyanobacterial surveys were conducted to characterize the dynamics of cyanotoxins in lake water and water treatment plants. Forty-six cyanobacterial taxa were recorded, and out of these, fourteen were considered potentially toxigenic (i.e., from the genera Dolichospermum, Microcystis, Oscillatoria, Pseudanabaena and Raphidiopsis). A higher concentration (ranging from 5 to 10 µg MC-LR equiv. L−1) of microcystins (MC) was detected in Murchison Bay compared to Napoleon Gulf, with a declining gradient from the inshore (max. 15 µg MC-LR equiv. L−1) to the open lake. In Murchison Bay, an increase in Microcystis sp. biovolume and MC was observed over the last two decades. Despite high cell densities of toxigenic Microcystis and high MC concentrations, the water treatment plant in Murchison Bay efficiently removed the cyanobacterial biomass, intracellular and dissolved MC to below the lifetime guideline value for exposure via drinking water (<1.0 µg MC-LR equiv. L−1). Thus, the potential health threats stem from the consumption of untreated water and recreational activities along the shores of the lake embayments. MC concentrations were predicted from Microcystis cell numbers regulated by environmental factors, such as solar radiation, wind speed in the N−S direction and turbidity. Thus, an early warning through microscopical counting of Microcystis cell numbers is proposed to better manage health risks from toxigenic cyanobacteria in Lake Victoria.

Keywords: Dolichospermum; Microcystis; drinking water; exposure routes; microcystins; rapid sand filtration; recreational areas.

Figure 1

Location of the two study areas in Murchison Bay (MB) and Napoleon Gulf (NG) (1) and close up with tributaries and wetlands (2) and the location of sampling sites (Inshore, Recreation, WTP abstraction points (Abs. point) and open lake) in MB (3) and NG (4). The sampling depths in MB were as follows: Inshore = 1.4–2.5 m (1.84 ± 0.34), Recreation = 2–2.7 m (2.3 ± 0.3), WTP abstraction points (Abs. point) = 9.4–18.4 m (14 ± 2.33) and Open Lake = 12–14 m (12.7 ± 0.5). In NG, the sampling depths were as follows: Inshore = 5.4–6.5 m (6.1 ± 0.3), Recreation = 5.7–7.2 m (6.5 ± 0.5), WTP abstraction points (Abs. point) = 11.4–15.8 m (13.1 ± 1.2) and Open Lake = 16.7–21.5 m (18.1 ± 1.3).

Figure 2

Potentially toxigenic and non-toxigenic cyanobacteria recorded from the lake survey at sampling sites: (i) inshore, (ii) recreational area, (iii) WTP abstraction point (Abs. point), (iv) open lake, in Murchison Bay and Napoleon Gulf. Among the 46 cyanobacteria species, 14 were considered potentially toxigenic. Circle sizes are proportional to the total biovolume (mm³ L⁻¹) of each taxon (data collected between November 2017 and October 2018, n = 120).

Figure 3

Cyanobacterial biovolume and composition at the WTP Abs. point and in raw water (RW) from MB-Gaba and NG-Walukuba water treatment plants. (A) Temporal dynamics of cyanobacteria genera; (B) Proportion of potentially toxigenic and non-toxigenic cyanobacteria (data collected between November 2016 and January 2017, n = 90).

Figure 4

Spatial variation in the intracellular MC concentrations (mean ± SD), (A) at (i) inshore stations in the bays, (ii) recreational areas, (iii) WTP abstraction points (Abs. point) and (iv) open lake in Murchison Bay (MB) and Napoleon Gulf (NG), (between November 2017 and October 2018, n = 120), (B) during the water treatment from lake water to final water in MB-Gaba and NG-Walukuba (between November 2016 and January 2017, n = 90). Note the different scales in the y-axis. (C) Relative frequency of the structural variants of MC detected during the lake survey and (D) from the WTP abstraction points and raw water entering the WTPs (n = 18), from MB-Gaba and NG-Walukuba. Abbreviations: Abs. point = WTP abstraction point, RW = Raw water, CFW = Coagulated and flocculated water, SFW = Sand filtered water, FW = Final water, nd = not detected.

Figure 5

Concentration of dissolved MC (mean ± SD) as detected by ELISA (A) and LC-MS (B) from WTP abstraction points in MB and NG, and during the treatment process in MB-Gaba and NG-Walukuba. (C) Relative frequency of the structural variants of MC detected in the dissolved water from MB-Gaba (n = 45). Note: there was no detection of dissolved MC in NG-Walukuba. Abbreviations: Abs. point = WTP abstraction point, RW = Raw water, CFW = Coagulated and flocculated water, SFW = Sand filtered water and FW = Final water.

Figure 6

Decision support tree for prediction of MC concentrations in lake water from Murchison Bay and Napoleon Gulf, northern Lake Victoria. Note: LOQ for MC = 0.04 µg L⁻¹; MeanWind_D5 = the mean wind speed for the five days (km h⁻¹) (N–S direction) and MeanSolar_10 = mean solar irradiance for the ten days prior to sampling (W m² day⁻¹). (Data collected between November 2017 and October 2018, n = 120). Colors: Red = 0.08–14.8 µg MC-LR equiv. L⁻¹, Yellow = LOQ–4.5 µg MC-LR equiv. L⁻¹, Brown and Blue = LOQ–0.91 µg MC-LR equiv. L⁻¹.