You are here:
Forecasting Microcystis aeruginosa population dynamics from bacterioplankton DNA
Citation:
Bagley, M., J. Santodomingo, A. Banerji, Joel Allen, J. Shoemaker, AND Dan Tettenhorst. Forecasting Microcystis aeruginosa population dynamics from bacterioplankton DNA. 2018 North American Lake Management Society, Cincinnati, OH, October 30 - November 02, 2018.
Impact/Purpose:
The ability to forecast harmful algal blooms (HABs) in freshwater lakes and rivers is becoming critical for management of drinking water treatment and recreational activities. In Harsha Lake, located in southwestern Ohio, Microcystis aeruginosa is believed to be the dominant cyanobacterium responsible for production of harmful cyanotoxins and therefore implicated in previous recreational public health advisories. We used random forest regression modeling to evaluate the ability of bacterioplankton community dynamics, as measured by 16S rRNA gene DNA metabarcoding, together with nutrients and other environmental factors (e.g., water gauge data, wind, temperature, rainfall), to predict relative abundance of Microcystis at 4 different locations in this lake over varying timeframes. When dividing the 2015 bloom season into 4-day intervals, 81- 84% of variation in Microcystis relative abundance could be explained from community and environmental data one to four intervals (~ 4-16 days) in advance. Of 6009 operational taxonomic units identified by DNA sequencing analyses, 69 were highly significant components of the models. Water gauge data (i.e., lake inflow, outflow, and height) also were highly significant model features while, surprisingly, total nitrogen, nitrate, ammonia and total phosphorus concentrations were not. These data provide preliminary evidence that genetic monitoring combined with environmental data can help support management of HAB health risks.
Description:
The ability to forecast harmful algal blooms (HABs) in freshwater lakes and rivers is becoming critical for management of drinking water treatment and recreational activities. In Harsha Lake, located in southwestern Ohio, Microcystis aeruginosa is believed to be the dominant cyanobacterium responsible for production of harmful cyanotoxins and therefore implicated in previous recreational public health advisories. We used random forest regression modeling to evaluate the ability of bacterioplankton community dynamics, as measured by 16S rRNA gene DNA metabarcoding, together with nutrients and other environmental factors (e.g., water gauge data, wind, temperature, rainfall), to predict relative abundance of Microcystis at 4 different locations in this lake over varying timeframes. When dividing the 2015 bloom season into 4-day intervals, 81- 84% of variation in Microcystis relative abundance could be explained from community and environmental data one to four intervals (~ 4-16 days) in advance. Of 6009 operational taxonomic units identified by DNA sequencing analyses, 69 were highly significant components of the models. Water gauge data (i.e., lake inflow, outflow, and height) also were highly significant model features while, surprisingly, total nitrogen, nitrate, ammonia and total phosphorus concentrations were not. These data provide preliminary evidence that genetic monitoring combined with environmental data can help support management of HAB health risks.
