Improved biovolume estimation of Microcystis aeruginosa colonies: A statistical approach
The Microcystis aeruginosa complex (MAC) clusters many of the most common freshwater and brackish bloom-forming cyanobacteria. In monitoring protocols, biovolume estimation is a common approach to determine MAC colonies biomass and useful for prediction purposes. Biovolume (μm3 mL−1) is calculated m...
Guardado en:
| Publicado: |
2018
|
|---|---|
| Materias: | |
| Acceso en línea: | https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_01677012_v151_n_p20_Alcantara http://hdl.handle.net/20.500.12110/paper_01677012_v151_n_p20_Alcantara |
| Aporte de: |
| id |
paper:paper_01677012_v151_n_p20_Alcantara |
|---|---|
| record_format |
dspace |
| spelling |
paper:paper_01677012_v151_n_p20_Alcantara2023-06-08T15:16:48Z Improved biovolume estimation of Microcystis aeruginosa colonies: A statistical approach Cyanobacteria monitoring Harmful algal blooms Machine learning Phytoplankton counting brackish water fresh water sea water Article autumn bacterium colony bacterium detection biomass environmental monitoring Microcystis aeruginosa nonhuman priority journal seasonal variation spring statistical model winter The Microcystis aeruginosa complex (MAC) clusters many of the most common freshwater and brackish bloom-forming cyanobacteria. In monitoring protocols, biovolume estimation is a common approach to determine MAC colonies biomass and useful for prediction purposes. Biovolume (μm3 mL−1) is calculated multiplying organism abundance (orgL−1) by colonial volume (μm3org−1). Colonial volume is estimated based on geometric shapes and requires accurate measurements of dimensions using optical microscopy. A trade-off between easy-to-measure but low-accuracy simple shapes (e.g. sphere) and time costly but high-accuracy complex shapes (e.g. ellipsoid) volume estimation is posed. Overestimations effects in ecological studies and management decisions associated to harmful blooms are significant due to the large sizes of MAC colonies. In this work, we aimed to increase the precision of MAC biovolume estimations by developing a statistical model based on two easy-to-measure dimensions. We analyzed field data from a wide environmental gradient (800 km) spanning freshwater to estuarine and seawater. We measured length, width and depth from ca. 5700 colonies under an inverted microscope and estimated colonial volume using three different recommended geometrical shapes (sphere, prolate spheroid and ellipsoid). Because of the non-spherical shape of MAC the ellipsoid resulted in the most accurate approximation, whereas the sphere overestimated colonial volume (3–80) especially for large colonies (MLD higher than 300 μm). Ellipsoid requires measuring three dimensions and is time-consuming. Therefore, we constructed different statistical models to predict organisms depth based on length and width. Splitting the data into training (2/3) and test (1/3) sets, all models resulted in low training (1.41–1.44%) and testing average error (1.3–2.0%). The models were also evaluated using three other independent datasets. The multiple linear model was finally selected to calculate MAC volume as an ellipsoid based on length and width. This work contributes to achieve a better estimation of MAC volume applicable to monitoring programs as well as to ecological research. © 2018 Elsevier B.V. 2018 https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_01677012_v151_n_p20_Alcantara http://hdl.handle.net/20.500.12110/paper_01677012_v151_n_p20_Alcantara |
| institution |
Universidad de Buenos Aires |
| institution_str |
I-28 |
| repository_str |
R-134 |
| collection |
Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA) |
| topic |
Cyanobacteria monitoring Harmful algal blooms Machine learning Phytoplankton counting brackish water fresh water sea water Article autumn bacterium colony bacterium detection biomass environmental monitoring Microcystis aeruginosa nonhuman priority journal seasonal variation spring statistical model winter |
| spellingShingle |
Cyanobacteria monitoring Harmful algal blooms Machine learning Phytoplankton counting brackish water fresh water sea water Article autumn bacterium colony bacterium detection biomass environmental monitoring Microcystis aeruginosa nonhuman priority journal seasonal variation spring statistical model winter Improved biovolume estimation of Microcystis aeruginosa colonies: A statistical approach |
| topic_facet |
Cyanobacteria monitoring Harmful algal blooms Machine learning Phytoplankton counting brackish water fresh water sea water Article autumn bacterium colony bacterium detection biomass environmental monitoring Microcystis aeruginosa nonhuman priority journal seasonal variation spring statistical model winter |
| description |
The Microcystis aeruginosa complex (MAC) clusters many of the most common freshwater and brackish bloom-forming cyanobacteria. In monitoring protocols, biovolume estimation is a common approach to determine MAC colonies biomass and useful for prediction purposes. Biovolume (μm3 mL−1) is calculated multiplying organism abundance (orgL−1) by colonial volume (μm3org−1). Colonial volume is estimated based on geometric shapes and requires accurate measurements of dimensions using optical microscopy. A trade-off between easy-to-measure but low-accuracy simple shapes (e.g. sphere) and time costly but high-accuracy complex shapes (e.g. ellipsoid) volume estimation is posed. Overestimations effects in ecological studies and management decisions associated to harmful blooms are significant due to the large sizes of MAC colonies. In this work, we aimed to increase the precision of MAC biovolume estimations by developing a statistical model based on two easy-to-measure dimensions. We analyzed field data from a wide environmental gradient (800 km) spanning freshwater to estuarine and seawater. We measured length, width and depth from ca. 5700 colonies under an inverted microscope and estimated colonial volume using three different recommended geometrical shapes (sphere, prolate spheroid and ellipsoid). Because of the non-spherical shape of MAC the ellipsoid resulted in the most accurate approximation, whereas the sphere overestimated colonial volume (3–80) especially for large colonies (MLD higher than 300 μm). Ellipsoid requires measuring three dimensions and is time-consuming. Therefore, we constructed different statistical models to predict organisms depth based on length and width. Splitting the data into training (2/3) and test (1/3) sets, all models resulted in low training (1.41–1.44%) and testing average error (1.3–2.0%). The models were also evaluated using three other independent datasets. The multiple linear model was finally selected to calculate MAC volume as an ellipsoid based on length and width. This work contributes to achieve a better estimation of MAC volume applicable to monitoring programs as well as to ecological research. © 2018 Elsevier B.V. |
| title |
Improved biovolume estimation of Microcystis aeruginosa colonies: A statistical approach |
| title_short |
Improved biovolume estimation of Microcystis aeruginosa colonies: A statistical approach |
| title_full |
Improved biovolume estimation of Microcystis aeruginosa colonies: A statistical approach |
| title_fullStr |
Improved biovolume estimation of Microcystis aeruginosa colonies: A statistical approach |
| title_full_unstemmed |
Improved biovolume estimation of Microcystis aeruginosa colonies: A statistical approach |
| title_sort |
improved biovolume estimation of microcystis aeruginosa colonies: a statistical approach |
| publishDate |
2018 |
| url |
https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_01677012_v151_n_p20_Alcantara http://hdl.handle.net/20.500.12110/paper_01677012_v151_n_p20_Alcantara |
| _version_ |
1768544906788208640 |