Phytoplankton Phenology

1
Phytoplankton Phenology

• Presented by
• Marco Vargas

2
Requirement, Science, and Benefit

• Requirement / Objective
• Ecosystems
• Increase number of regional coastal and marine
  ecosystems delineated with approved indicators of
  
• ecological health and socioeconomic
  benefits that are monitored and understood.
• Increase portion of population that is
  knowledgeable of and acting as stewards for
  coastal and marine
• ecosystems.
• Climate
• Understand and predict the consequences of
  climate variability and change on marine
  ecosystems
• Science
• Can we model the annual cycles of phytoplankton
  and describe the timing and magnitude of
  phytoplankton
• biomass in the open ocean?
• Can we relate changes in phytoplankton biomass to
  oceanic variables?
• Are there any trends in the timing and magnitude
  of phytoplankton biomass?
• Benefit
• National Ocean Service and their customers, Ocean
  Color Community
• Assessing marine ecosystem response to climate
  change
• Ability to monitor and detect changes in
  distribution of phytoplankton in marine
  ecosystems

3
Challenges and Path Forward

• Science challenges
• Improve atmospheric correction
• Need longer time-series
• Next steps
• Explore different time/spatial resolutions to
  maximize resolved features
• Extend study to global oceans
• Develop equivalent algorithms to process
  NOAA-generated
• VIIRS/NPP
• Transition Path
• Generate experimental version from MODIS/SeaWIFS
• Generate operationally from MODIS/SeaWIFS for
  global ocean
• Include algorithm development/validation for
  VIIRS/NPP to provide
• product continuity
• Our goal is to transition the phytoplankton
  phenology product to operations within the next
  few years.
• End users National Ocean Service and their
  customers, Ocean Color Community

4
Modeling of Bloom Data

• Daily SeaWIFS global chlorophyll (4 km res) is
• aggregated to pentad (five-day) means with a
• spatial resolution of 3x3 lat/lon
• Highly non Gaussian
• Bloom data are non-negative
• GLMs (Generalized Linear Models) for
• Gamma distributed data.
• The chlorophyll amount Y is modeled using a
• Gamma GLM with the canonical log link
• Spatial distribution of the eight models
  constructed
• to represent the annual cycle of chlorophyll in
  the
• study area

Distribution of example grid-box D with estimated
distributions overlaid. The black line represents
a Gaussian distribution and the red line
represents a Gamma distribution
5
Generalized Linear Models (GLMs) for Gamma
distributed data

6
Determination of Phenological Markers
Spatial distribution of (A) bloom onset, (B)
bloom maturity, (c) start of bloom decay, and (d)
bloom termination