The contribution of volunteer-based monitoring data to the assessment of harmful phytoplankton blooms in Brazilian urban streams

- Data from Brazilian urban streams were successfully collected by trained volunteers.
- Relatively high nutrient and turbidity levels were observed in the streams.
- Pollution sources and riparian vegetation were related to phytoplankton abundance.
- Phytoplankton densities and phosphate (not nitrate) were positively correlated.
- Thresholds for Cyanobacteria blooms were established based on volunteers' data.

Urban streams are vulnerable to a range of impacts, leading to the impairment of ecosystem services. However, studies on phytoplankton growth in tropical lotic systems are still limited. Citizen science approaches use trained volunteers to collect environmental data. We combined data on urban streams collected by volunteers with data obtained by professional scientists to identify potential drivers of phytoplankton community and determine thresholds for Cyanobacteria development. We combined datasets (n = 117) on water quality and environmental observations in 64 Brazilian urban streams with paired data on phytoplankton. Sampling activities encompassed dry (July 2013 and July 2015) and warm (February and November 2014) seasons. Volunteers quantified phosphate (PO43 −), nitrate (NO3−) and turbidity in each stream using colorimetric and optical methods and recorded environmental conditions in the immediate surroundings of the sites through visual observations. We used non-parametric statistics to identify correlations among nutrients, turbidity and phytoplankton. We also looked for thresholds with respect to high Cyanobacteria abundance (> 50,000 cells/mL). The streams were characterized by relatively high nutrient concentrations (PO43 −: 0.11 mg/L; NO3−: 2.6 mg/L) and turbidity (49 NTU). Phytoplankton densities reached 189,000 cells/mL, mainly potentially toxic Cyanobacteria species. Moderate but significant (p < 0.05) correlations were observed between phytoplankton density and turbidity (ρ = 0.338, Spearman) and PO43 − (ρ = 0.292), but not with NO3−. Other important variables (river flow, temperature and light) were not assessed. Volunteers' observations covaried with phytoplankton density (p < 0.05, Kruskal-Wallis), positively with increasing number of pollution sources and negatively with presence of vegetation in the riparian zone. Our results indicate that thresholds for PO43 − (0.11 mg/L) can be used to separate systems with high Cyanobacteria density. The number of pollution sources provided a good indicator of waterbodies with potential cyanobacteria problems. Our findings reinforced the need for nutrient abatement and restoration of local streams and highlighted the benefits of volunteer-based monitoring to support decision-making.

289