Estimating phytoplankton stoichiometry from routinely collected monitoring data

• Published in: Biogeochemistry 2022-06, Vol.159 (2), p.251-264
• Main Authors: Yuan, Lester L., Jones, John R.
• Format: Article
• Language: English
• Subjects: Bayesian analysis; Bayesian theory; Biogeochemical cycle; Biogeochemical cycles; Biogeosciences; Earth and Environmental Science; Earth Sciences; Ecosystems; Environmental changes; Environmental Chemistry; Environmental conditions; Impoundments; Lakes; Life Sciences; Mathematical models; Mineral nutrients; Monitoring; Nitrogen; Nutrient concentrations; Nutrient content; Nutrients; Particulate matter; Phosphorus; Phytoplankton; Plankton; Probability theory; Sediment; Stoichiometry; Suspended particulate matter; Suspended sediments
• ISSN: 0168-2563; 1573-515X
• DOI: 10.1007/s10533-022-00926-8

Accurately estimating the elemental stoichiometry of phytoplankton is critical for understanding biogeochemical cycles. In laboratory experiments, stoichiometric ratios vary among species and with changes in environmental conditions. Field observations of total phosphorus (P) and total nitrogen (N) collected at regional and national scales can supplement and expand insights into factors influencing phytoplankton stoichiometry, but analyses applied to these data can introduce biases that affect interpretations of the observed patterns. We introduce an analytical approach for estimating the ratio between phytoplankton N and P from the particulate fraction of nutrient pools in lake samples. We use Bayesian models to represent observations of particulate P and N as the sum of contributions from nutrients bound within phytoplankton and nutrients associated with non-phytoplankton suspended sediment. Application of this approach to particulate nutrient data collected in Missouri impoundments yields estimates of the mass ratio of N:P in phytoplankton ranging from 8 to 10 across a variety of lakes and seasons. N:P in particulate matter ranged from 6 to 70, a variability driven by differences in nutrients bound to non-phytoplankton suspended sediment. We adapted the Bayesian models to estimate N:P using more commonly available measurements of total P and total N and applied this model to a continental-scale monitoring data set. We compared phytoplankton nutrient content estimated from the two analyses and found that when datasets lack direct measurements of particulate nutrient concentrations, the model estimate of phytoplankton nutrient content includes contributions from nutrients within phytoplankton and dissolved nutrients that are associated with changes in phytoplankton biomass.