GreenSeas Earth Observation Data

All data: http://thredds.nersc.no:8080/thredds/greenseas/greenseas-EO.html

Chlorophyll-a concentratiom

Chlorophyll-a is a green pigment found in plants and phytoplankton cells. Chlorophyll pigments allow the cells to harvest light and to transform it into energy through the reaction of photosynthesis. Chlorophyll-a concentrations are widely used as a proxy of phytoplankton biomass and it is expressed in milligrams of Chlorophyll-a per cubic meter. The algorithms for estimating near-surface chlorophyll-a concentration are empirical or semi-empirical and they are based on retrieval measurements of water leaving radiance (O'Reilly et al., 2000).

To construct the chlorophyll-a time-series, data were retrieved for the period 1997-2010 from NASA Ocean Color Web http://oceancolor.gsfc.nasa.gov. The R2010.0 reprocessing of Level 3 Mapped chlorophyll-a concentration from the Sea-viewing Wide Field-of-view (SeaWiFS) sensor was downloaded at 9 km spatial resolution and monthly temporal resolution. Then, linear interpolation was applied to map the chlorophyll-a concentration onto a 1degreex1degree fixed grid and the data were compiled into a Netcdf file.

Key limitations: Chlorophyll concentrations are only available from the first optical depth. The spatial and temporal coverage of remotely sensed data are limited by clouds, atmospheric aerosol, sensor saturation over ice or snow masses, and high solar zenith angle. The global chlorophyll products derived from NASA's ocean color satellite programs have a nominal uncertainty of ± 35% (Moore et al., 2009).

Key strength: High-resolution data, global spatial coverage, and continuous long-term observations.

Primary production

Chlorophyll-a is a green pigment found in plants and phytoplankton cells. Chlorophyll pigments allow the cells to harvest light and to transform it into energy through the reaction of photosynthesis. Chlorophyll-a concentrations are widely used as a proxy of phytoplankton biomass and it is expressed in milligrams of Chlorophyll-a per cubic meter. The algorithms for estimating near-surface chlorophyll-a concentration are empirical or semi-empirical and they are based on retrieval measurements of water leaving radiance (O'Reilly et al., 2000).

To construct the chlorophyll-a time-series, data were retrieved for the period 1997-2010 from NASA Ocean Color Web http://oceancolor.gsfc.nasa.gov. The R2010.0 reprocessing of Level 3 Mapped chlorophyll-a concentration from the Sea-viewing Wide Field-of-view (SeaWiFS) sensor was downloaded at 9 km spatial resolution and monthly temporal resolution. Then, linear interpolation was applied to map the chlorophyll-a concentration onto a 1degreex1degree fixed grid and the data were compiled into a Netcdf file.

Key limitations: Chlorophyll concentrations are only available from the first optical depth. The spatial and temporal coverage of remotely sensed data are limited by clouds, atmospheric aerosol, sensor saturation over ice or snow masses, and high solar zenith angle. The global chlorophyll products derived from NASA's ocean color satellite programs have a nominal uncertainty of ± 35% (Moore et al., 2009).

Key strength: High-resolution data, global spatial coverage, and continuous long-term observations.

Phytoplankton phenology

A number of indices have been developed to characterize quantitatively the seasonality of phytoplankton (Platt and Sathyendranath, 2008). Specifically, indices that relates to the study of timing of periodic biological events as influenced by the environment are referred to as phytoplankton phenology. These indices include: timings of initiation, peak, and termination as well as the duration of the phytoplankton growing period. Changes in phytoplankton phenology (triggered by variations in climate) can profoundly alter: (1) the efficiency of the biological pump, with inevitable impact of the global carbon cycle; and (2) the interactions across trophic levels, which can engender trophic mismatch with major impacts on the survival of commercially important fish and crustacean larvae.

Phenology indices were estimated using the R2010.0 SeaWiFS chlorophyll-a time-series (see section 2.1) for the period 1997-2008 at 9 km spatial resolution and 8-day temporal resolution. Linear interpolation was applied to map the chlorophyll-a concentration onto a 1degreex1degree fixed grid. Missing data were reduced from the time series prior to estimating the timing of ecological events. Missing values were filled by interpolating spatially adjacent values (average of 3 × 3 pixels on the 9km grid), when these were available. Any remaining missing values were filled by interpolating temporally adjacent values (average of previous and following 8-day composites), when these were available. Otherwise the value was not filled. A 3-week running mean was applied to remove small peaks in chlorophyll-a. The timings of initiation and end of the phytoplankton growing period were detected as the weeks when the chlorophyll concentration in a particular year rose above the long-term median value plus 5% and later fell below this same threshold (Racault et al., 2012). The duration of the growing season is defined as the number of weeks between initiation and end. The phenology indices were compiled into a Netcdf file.

Key limitations: The present method of analysis captures only the characteristics of the main phytoplankton growing season that is associated with the timing of the maximum amplitude of the chlorophyll signal in a given year. The method does not account for sub-surface chlorophyll maxima or secondary blooms. The spatial and temporal coverage of remotely sensed data are limited by clouds, atmospheric aerosol, sensor saturation over ice or snow masses, and high solar zenith angle. In subpolar regions, errors associated with persistent missing data on the estimation of timings of phytoplankton bloom initiation and peak have been evaluated to be 30 and 15 days, respectively, from the GlobColour time-series (Cole et al., 2012).

Key strength: High-resolution data, global spatial coverage, and continuous long-term observations.

Phytoplankton size fractions

Phytoplankton cell sizes can be delineated into phytoplankton size fractions, which can be used to characterize major functional groups. The phytoplankton size fractions are typically split into picoplankton (<2 ?m), nanoplankton (2–20 ?m) and microplankton (>20 ?m). The rates of photosynthesis vary with the phytoplankton community structure with the larger phytoplankton size fraction (microphytoplankton) generally having lower maximum photosynthetic rates per unit of biomass compared to nano- and picophytoplankton. However, microphytoplankton have a higher carbon to chlorophyll ratio and furthermore they tend to have faster sinking rates, making them more efficient to export carbon from the surface layer to the deep ocean. Environmental variations in wind and temperature can drive changes in the size structure of the phytoplankton community, which can alter trophic interactions with significant impacts on fisheries recruitment. A variety of bio-optical models have been designed for use in remote sensing to map phytoplankton size fractions on global and regional scales (Sathyendranath et al., 2001; Ciotti and Bricaud, 2006; Devred et al., 2006; Uitz et al., 2006; Hirata et al., 2008). These models typically detect either the dominant phytoplankton size fractions or the fractional contribution of each phytoplankton size fraction at a given satellite pixel.

The three-component model developed by Brewin et al. (2010) was used to calculate the fractional contributions of the three phytoplankton size fractions (micro-, nano- and picoplankton) to the overall chlorophyll-a concentration. The SeaWiFS chlorophyll-a time-series (see section 2.1) for the period 1997-2010 at 1degree spatial resolution and monthly temporal resolution was used as input to implement the model of Brewin et al. (2010). The phytoplankton size fractions were compiled into three Netcdf files.

Key limitations: The model has been designed to capture the major trends in phytoplankton community structure in the open ocean (Case I waters). The results cannot be used to interpret variations in phytoplankton community size structure in more complex waters (coastal regions, Case II waters). The spatial and temporal coverage of remotely sensed data are limited by clouds, atmospheric aerosol, sensor saturation over ice or snow masses, and high solar zenith angle.

Key strength: High-resolution data, global spatial coverage, and continuous long-term observations.

Ratio of euphotic depth to mixed layer depth

In the water column, the depth at which there is sufficient light level for photosynthesis to occur is defined as euphotic layer (Zeu). When Zeu is shallower than the mixed layer (Zmld), production exceeds respiration and net phytoplankton growth can occur. Zeu depends on factors attenuating the light penetration through the water column including water transparency, suspended particulate matter and phytoplankton concentrations. Seasonal variability in Zmld is driven by temperature and wind. Deep water mixing allows for nutrient replenishment within the layer.

Zeu can be estimated from remote-sensing reflectance following two main approaches: either a quasi-analytical algorithm based on Inherent Optical Properties (IOPs) or an empirical relationship based on chlorophyll-a concentrations (Lee et al., 2007; Morel et al., 2007). Zmld can be estimated based on a threshold criterion of water temperature or density (e.g. from a near-surface value at 10m depth: Temperature difference = 0.2°C or Density difference = 0.03 kg m-3; de Boyer Monte?gut et al., 2004).

Euphotic depth derived from remote-sensing reflectance using the Inherent Optical Properties approach (Lee et al., 2007) was retrieved from the evaluation products available at NASA Ocean Color Web http://oceancolor.gsfc.nasa.gov. The R2010.0 SeaWiFS reprocessing of Level 3 Mapped Zeu data was downloaded at 9 km spatial resolution and monthly temporal resolution. Then, linear interpolation was applied to map Zeu onto a 1degreex1degree fixed grid. Temperature data were downloaded from Simple Ocean Data Assimilation (SODA) model output (Carton and Giese, 2008; http://iridl.ldeo.columbia.edu/SOURCES/.CARTON-GIESE/.SODA/.v2p1p6/) for the period 1997-2007 at monthly temporal resolution and 0.25degreex0.4degreex40-level spatial and vertical resolutions. Zmld was defined as the temperature criterion of ±0.2°C change compared with the temperature at 10m depth (de Boyer Monte?gut et al., 2004). To remain coherent with the resolution of the Zeu data, Zmld data were then averaged to one-degree spatial resolution. Finally, the ratio Zeu:Zmld was calculated and the data were compiled into a Netcdf file.

Key limitations: The reanalysis output of temperature should not be equated with "observations". The availability of observations used in the reanalysis can considerably vary depending of the location and time period, hence possibly affecting the reliability of the reanalysis output. Furthermore the different sources of observations, and their potential biases, may introduce spurious variability and trends in the reanalysis output. The quasi-analytical algorithm used to estimate Zeu has been developed for the open ocean (Case I waters). The spatial and temporal coverage of remotely sensed data of Zeu are limited by clouds, atmospheric aerosol, sensor saturation over ice or snow masses, and high solar zenith angle.

Key strength: SODA temperature reanalysis incorporate millions of observations. The data set is available at high-resolution, global spatial coverage, and long-term temporal coverage (1 decade).

References

Marie-Fanny Racault, Trevor Platt and Shubha Sathyendranath: D6.4: Time-series of ecosystem status for the Atlantic and Southern Oceans, produced from remotely sensed data

Racault M.-F. (2014). Global ocean atlas of phytoplankton phenology indices from SeaWiFS ocean-colour time-series during the decade 1998-2007. British Oceanographic Data Centre - Natural Environment Research Council, UK. doi:10.5285/005c027f-2d88-087c-e053-6c86abc04e4c