My research focuses on marine optics and its application to the interpretation of ocean-colour radiometry (OCR). The data are measured in situ (ship) or from remote sensors (aircrafts or satellites). In other words, I am interested in the fate of those photons, emitted from the sun in the visible part of the spectrum, that enter our atmosphere and reach the surface of the ocean, and then penetrate the water. Some of the photons will be reflected back into the atmosphere without interacting with the ocean. The ones that enter the ocean will either get absorbed or leave the ocean where they can be collected by various optical sensors. The ratio of light entering and leaving the ocean carries useful information about the nature of the particles and dissolved material present in the seawater. A major part of my work deals with decoding this information to retrieve important biogeochemical characteristics of the oceans.

Below is a brief overview of some of the ocean colour activities I have investigated. I have tried to use simple concepts to avoid inundating the uninitiated with too many details or too much information. Please let me know if you have any questions.

Phytoplankton bio-optical model

Phytoplankton, ubiquitous in the oceans, form the basis of the marine trophic foodweb and play a major role in absorbing carbon dioxide (CO2) from the atmosphere. They also play a major role in the optical characterisation of sea water. Without any phytoplankton, the ocean would be blue, but like leaves on the trees, phytoplankton contain chlorophyll pigments that absorbs light in the blue and red region of the visible spectrum. Changes in concentration of phytoplankton explain the various shades, from blue to green, of the oceans.

phytoplankton absorption spectrum

Example of phytoplankton absorption coefficient.

Bio-optical models, based on in situ measurements, are a mathematical formulation that relates phytoplankton absorption to phytoplankton concentration (indexed by its chlorophyll-a content). Those models are wavelength dependent, since absorption by phytoplankton is not constant in the visible spectrum. Part of my work has been dedicated to develop phytoplankton absorption models (aka bio-optical model). Recently, in collaboration with Dr. Shubha Sathyendranath, we have developed an absorption model that, in addition to relating absorption as a function of chlorophyll-a concentration, infers phytoplankton size cell (Devred et al. 2006, JGR). I also have an interest in species-dependent phytoplankton absorption. In my primary area of interest, the Northwest Atlantic, diatoms are dominant during bloom events. Their specific spectral signature (absorption) compared to that of other phytoplankton species allowed us to develop an algorithm to identify diatom-dominated waters using remote sensing data(Sathyendranath et al., 2004, MEPS).

It would be easy to determine water optical properties if only phytoplankton (and its degradation products) were responsible for absorption and scattering in the water column. This assumption holds in open oceans, far from terrigeneous and anthropogenic sources, but this is not the case in coastal waters. Sediments and dissolved organic matter (DOM) add a level of complexity to the problem. To account for the other optically active components, optical models have been developed, and in some cases models have been validated and are commonly used by the ocean colour community. So far, I have not have the time, or data, to study these relationships. Once all the marine components have been identified and the optical properties computed, it is possible to compute the water-leaving radiative field.

Reflectance model and inversion of Inherent Optical Properties (IOP)

Absorption and backscattering of marine components (i.e., pure seawater, phytoplankton, sediments and dissolved organic matter) are referred to as inherent optical properties, because these optical properties are independent of the light field and depend only on the nature of the particles or molecules. Unlike IOPs, apparent optical properties (AOP) do depend on the light field (sun geometry, proportion of direct to diffuse radiation), the marine interface state and other variables. Reflectance, R, defined as the ratio of upwelling to downwelling radiation, is an apparent optical property.

Reflectance can be inferred from inherent optical properties using a simple ratio: R=k bb/(a+bb), where a represents the total absorption coefficient of the seawater, bb represents the total backscattering coefficient of seawater and k is a constant that depends on the sun geometry and other parameters. Inversion of reflectances uses optimisation methods to find the best set of IOPs that corresponds to a given set of reflectance measurements. For example, a radiometric sensor measures reflectances R412, R443, R490, R510, R555, R670 at 412, 443, 490, 510, 555 and 670 nm (i.e., SeaWiFS wavelengths). Using an appropriate reflectance model and wavelength dependence models for the absorption and backscattering marine constituents, the backscattering and absorption properties of the marine constituents at a reference wavelength (often 443 nm) can be derived for the measurements R443, R490, R510, R555, R670 cited above. I have been involved in a working group mandated by the International Ocean Colour Coordinating Group to deliver a report on the various methods to deal with inversion of IOPs (IOCCG report #5).

Identification of ecological provinces in the NW Atlantic using Ocean Colour data

static versus dynamic assignment of the provinces' boundaries

Static versus dynamic assignment of the provinces' boundaries. The right image was computed for the last two weeks of April 2003.

Definition of biomes and ecosystems in the ocean has been studied for decades and a major work of Alan Longhurst (Ecological geography of the sea, Academic Press, 2007, 542p.) has helped the oceanographic community in defining ecological provinces. Based on the work of Longhurst et al. (1995), I have developed a method to delineate ecological provinces in the Northwest Atlantic.

This method uses phytoplankton biomass fields (Chl-a) and sea surface temperatures (SST) measured by the Moderate Resolution Imaging Spectroradiometer (MODIS). A statistical analysis (cluster analysis) was performed on two-week composite images of SST and Chl-a to account for the geographical variation of the province boundaries with time (manuscript accepted in MEPS).

The Arctic Ocean

chlorophyll, sea-ice and sea surface temperature in the Beaufort Sea The right panel shows chlorophyll-a concentration (MODIS Aqua) and sea-ice concentration (NSIDC), the left panel shows sea-surface temperature (MODIS Aqua) and sea-ice concentration. The black arrows correspond to the wind and the solid red line correspond to the 200 m isobath.

The arctic and its ocean is warming at a faster rate than the rest of the world. Summer-ice extent is receding and records-low are more and more frequent. The decrease in ice cover and the warming of the sea water is changing the primary producers (i.e. phytoplankton). Recent studies show that the arctic ocean is becoming more productive, however the number of small phytoplankton seems to increase due to increase stratification (a result of ice melt). Satellite sensors, and ocean-colour satellite in particular, allow a synoptic monitoring of the Arctic ocean (e.g., sea-ice cover, sea-surface temperature, chlorophyll-a concentration, dissolved organic carbon, wind stress).

At Takuvik, I am responsible of the remote sensing component of the laboratory. With the help of two colleagues, I provide support to students and researchers, and I also pursue my own research, focusing on phytoplankton functional types and retrieval of marine inherent optical properties (mainly, phytoplankton absorption). I have also interest on the fate of phytoplankton and its response to physical forcing.