Mind the gap: The impact of missing data on the calculation of phytoplankton phenology metrics

• Published in: Journal of Geophysical Research: Oceans Vol. 117; no. C8
• Main Authors: Cole, Harriet, Henson, Stephanie, Martin, Adrian, Yool, Andrew
• Format: Journal Article
• Language: English
• Published: Washington, DC Blackwell Publishing Ltd 01.08.2012; American Geophysical Union
• Subjects: Anthropogenic factors; biogeochemical model; Biological oceanography; Chemical oceanography; Data processing; Earth sciences; Earth, ocean, space; Exact sciences and technology; Geobiology; Geophysics; Marine; Marine ecosystems; NOBM; ocean color remote sensing; Oceanography; Phenology; Phytoplankton; phytoplankton phenology; Pneumoviridae; Remote sensing; Seasonal variations; seasonality metrics; spring bloom; Trophic levels; Zooplankton; color; aerosols; clouds; annual variations; Winter; Trophic level; carbon; phytoplankton; plankton; export; models; NASA; Time series; zooplankton; Satellite observation; Subpolar zone; Sun; Interannual variation; Missing data; Phenology; ecosystems; Timing; SeaStar satellite; seasonal variations; errors; Comparative study
• ISSN: 0148-0227; 2169-9275; 2156-2202; 2169-9291
• DOI: 10.1029/2012JC008249

Annual phytoplankton blooms are key events in marine ecosystems and interannual variability in bloom timing has important implications for carbon export and the marine food web. The degree of match or mismatch between the timing of phytoplankton and zooplankton annual cycles may impact larval survival with knock‐on effects at higher trophic levels. Interannual variability in phytoplankton bloom timing may also be used to monitor changes in the pelagic ecosystem that are either naturally or anthropogenically forced. Seasonality metrics that use satellite ocean color data have been developed to quantify the timing of phenological events which allow for objective comparisons between different regions and over long periods of time. However, satellite data sets are subject to frequent gaps due to clouds and atmospheric aerosols, or persistent data gaps in winter due to low sun angle. Here we quantify the impact of these gaps on determining the start and peak timing of phytoplankton blooms. We use the NASA Ocean Biogeochemical Model that assimilates SeaWiFS data as a gap‐free time series and derive an empirical relationship between the percentage of missing data and error in the phenology metric. Applied globally, we find that the majority of subpolar regions have typical errors of 30 days for the bloom initiation date and 15 days for the peak date. The errors introduced by intermittent data must be taken into account in phenological studies. Key Points Global maps of seasonality metrics and the associated uncertainty are presented Bloom start and peak date errors are 30 and 15 days respectively in most regions The error in bloom start date has a directional bias that changes with latitude