DRAFT
November 17, 2010 – V1.6
Geo-CAPE Coastal Ocean Ecosystem Dynamics White Paper (long version for HQ)
White Paper Objectives:
1. Provide a detailed explanation of the Science Traceability Matrix – THE MISSION.
2. Justify the science (referring to Decadal Survey and scientific literature).
3. Explain the societal benefits and advancement in scientific understanding.
4. Provide an update on the work-to-date on the mission.
5. Describe the pre-launch activities and resources necessary to advance the mission.
6. Provide a plan for post-launch cal/val (resources and activities).
Background (2-3 pages)
a. Earth Science Decadal Survey
The U.S. National Research Council (NRC), at the request of NASA, NOAA and the U.S.
Geological Survey, conducted an Earth Science Decadal Survey review to assist these agencies
in planning the next generation of Earth Science satellite missions. The final report
recommended 17 missions including the Geostationary Coastal and Air Pollution Events (GeoCAPE) mission that focuses on measurements of tropospheric trace gases and coastal ocean
color from geostationary orbit (NRC, 2007). The NRC placed Geo-CAPE within the second tier
of mission launches, which NASA plans to launch after 2020 (NASA 2010). “A primary
objective for observing coastal ocean regions is to determine the impact of climate change and
anthropogenic activity on primary productivity and ecosystem variability (NRC 2007).” The
oceans component of the Geo-CAPE focuses on coastal ecosystem dynamics and would provide
upper ocean observations of water-leaving radiances, chlorophyll, primary productivity,
particulate and dissolved organic carbon, particulate inorganic carbon, turbidity, sediment fluxes,
land-ocean carbon fluxes, and phytoplankton community structure (ibid). The NRC DS report
presented several coastal ocean science objectives for Geo-CAPE:
• To quantify the response of marine ecosystems to short-term physical events, such as passage
of storms and tidal mixing.
• To assess the importance of high temporal variability in coupled biological-physical coastalecosystem models.
• To monitor biotic and abiotic material in transient surface features, such as river plumes and
tidal fronts.
• To detect, track and predict the location of sources of hazardous materials, such as oil spills,
waste disposal, and harmful algal blooms.
• To detect floods from various sources, including river overflows.
I.
Over the past several years, NASA has initiated planning efforts and science and engineering
studies for a geostationary ocean color mission. NASA’s Earth’s Living Oceans: The Unseen
World: An advance plan for NASA’s Ocean Biology and Biogeochemistry Research document
(NASA 2006) presents a description for a geostationary ocean color mission. This planning
document describes the mission science and a preliminary set of instrument requirements for a
geostationary hyperspectral imaging radiometer to study coastal ocean processes. The
considerations for ocean color retrievals in coastal waters entails significant improvements in
current ocean color sensor capabilities, which include: high frequency sampling each day,
higher spatial and spectral resolution than current sensors, broad spectral coverage including
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
UV-VIS-NIR and SWIR bands, high signal-to-noise ratio (SNR) and dynamic range, cloud
avoidance, minimal polarization sensitivity or change (<0.2%), minimal stray light with narrow
field-of-view (FOV) optics and low scatter gratings (<0.1%), no image striping or image latency,
solar and lunar on-orbit calibration, etc. Details on the proposed instrument requirements are
described in the NASA planning document (2006; http://oceancolor.gsfc.nasa.gov/DOCS/). The
authors recommended that NASA (or in partnership with other U.S. federal agencies or
international space agencies) contribute a regional sensor to a broader international effort to
provide global coverage of the coastal oceans from geostationary orbit.
Among the challenges posed by satellite observations of water leaving radiances from coastal
waters is their small contribution to the total flux at the top of the atmosphere (TOA). Typical
signals from oceans contribute <10% to the total flux, but the presence of colored dissolved
organic material (CDOM) in coastal waters can reduce this reflectance even further to <1% of
the total signal (see section V.a). As a result, it is imperative to adequately correct the signal for
various atmospheric contributions to the total signal (see section II.d). Indeed, this was the
motivation of the NRC (2007) for combining the air quality (AQ) and ocean color (OC)
objectives from geostationary orbit into one mission: to enable optimal aerosol corrections to the
OC retrievals. In highly urbanized coastal zones correcting for trace gases such O3 and NO2 are
also critical, especially to avoid an atmospheric signature imposing a false impression of
temporal and spatial variability within the coastal waters (see sections II.d and IV.a). By
combining AQ and OC observations into a single mission, a unique opportunity arises for
studying coupled atmospheric and coastal ocean processes, including terrestrial linkages via
watershed processes. A discussion of such interdisciplinary topics are beyond the scope of the
present white paper, however, the potential science that may be enabled by GEO-CAPE is
addressed in a separate white paper (Jordan et al., in preparation).
b. Geo-CAPE activities
In August 2008, NASA convened a broad community workshop on the Geo-CAPE mission to
refine the scientific goals, objectives and requirements for this mission and define the necessary
investments to advance the mission concept for a Phase A mission start. The workshop report is
posted at http://Geo-CAPE.larc.nasa.gov/documents.html. NASA assembled formal atmospheric
and ocean science working groups between August 2008 and April 2009 and convened a
working meeting in September 2009 to discuss Geo-CAPE mission science objectives and
requirements. The agenda and presentations given at this meeting are posted at http://geocape.larc.nasa.gov/events-SEP2009SWGM.html. In March 2010, NASA convened a working
meeting of the Geo-CAPE ocean and atmospheric science working groups. The meeting focused
on preparing for mission design studies with goals of: endorsing draft Science Traceability
Matrices (STMs) as starting points for mission design studies; identifying any needed instrument
design studies; and specifying initial mission design studies and preparations. During this
meeting, draft Geo-CAPE Atmospheric and Coastal Ocean Ecosystems Dynamics Science
Traceability Matrices were endorsed by voice consensus as a sufficient, although not immutable,
starting point for preliminary mission planning work. As living documents, the STMs are
expected to further evolve pending the outcome of ongoing science pre-formulation studies. By
the close of the meeting, the Science Working Groups (SWG) and Mission Design Coordination
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Group agreed to conduct a sequence of mission design studies. Currently, NASA is funding
science and engineering studies for the Tier 2 Earth Science Decadal Survey missions including
Geo-CAPE.
Other Geostationary Ocean Color Observations
Previous mission concept: GOES-R coastal waters imager
Current ocean color sensors, for example SeaWiFS and MODIS and in the future VIIRS, are
well suited for sampling the open ocean. However, coastal environments are spatially and
optically more complex and require more frequent sampling and higher spatial resolution sensors
with additional spectral channels. To address those issues, NOAA considered including a Coastal
Waters imaging capability (HES-CW) as part of the Hyperspectral Environment Suite (HES) on
the next generation Geostationary Operational Environmental Satellite (GOES-R). From 2004 to
2006, NOAA supported the COAST program to assess requirements for geostationary imaging
in the coastal ocean. Based on the results of the first experiment in September 2006 a spatial
sampling of 300 m or better and the MERIS channel set or better were recommended for coastal
imaging (Davis, et al. 2007). In October 2006 due to budget and engineering concerns, NOAA
dropped HES from the GOES-R program and the COAST program was terminated. NOAA still
recognizes that it has strong requirements for coastal waters imaging, but it has no specific plan
to meet those requirements at this time.
Present Geostationary Ocean Color Instruments: GOCI
On June 26, 2010, the Korean Geostationary Ocean Color Imager (GOCI) was launched on
Ariane 5 as part of the Communication, Ocean and Meteorological Satellite (COMS-1)
spacecraft, which was developed jointly by the Korean Aerospace and Research Institute (KARI)
and EADS Astrium. The GOCI sensor is an 8-band (412, 443, 490, 555, 660, 680, 745, 865 nm)
staring frame capture sensor with a targeted coverage area of 2500x2500km centered on 130˚E
and 36˚N. GOCI’s 1 hour imaging frequency (8 times/day) and 500x500m local spatial
resolution will permit unprecedented retrievals of coastal ocean dynamics. The Korea Ocean
Research and Development Institute (KORDI) is planning a follow-on mission to GOCI (called
GOCI-2) with a launch date of January 2018.
Future Vision: constellation of Geostationary Imagers
Geo-CAPE should be considered in the context of an emerging international constellation of
ocean color radiometry sensors that will significantly improve our understanding of ocean
biology, biogeochemistry and ecology in coastal and offshore waters. Given the regional nature
of geostationary ocean color radiometry observations, a geographically distributed constellation
of regionally-focused imagers is a crucial need in order to provide the desired global coastal
coverage accompanied by high temporal revisits for dynamic regions (IGOS, 2006). NASA, via
Geo-CAPE, could contribute one or more sensors/platforms (alone or as part of a mission of
opportunity partnership with another U.S. or international space agency) to an international
global effort that will join other regional efforts such as the recently launched GOCI on the
COMS-1 platform from South Korea. The International Ocean Colour Coordinating Group
(IOCCG) is working to facilitate international coordination and cooperation in this context
through establishment of a Working Group on “Ocean Colour Observations from the
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Geostationary Orbit” that will articulate drivers, needs, requirements and evaluate present and
planned capabilities with regard to geostationary ocean color observations in support of both
research and applications.
Science Traceability Matrix
The main advantage afforded by a geostationary platform versus a low-earth polar orbit is the
capability to image the same regions multiple times per day. Such a capability is vital to study
coastal oceans where the physical, biological and chemical processes react on short time scales
from seconds to a few days. From a geostationary vantage point, a sensor can stare at an
instantaneous field-of-view (iFOV) to gain sufficient signal-to-noise to retrieve ocean
reflectances during low light conditions such as at high solar zenith angles (early morning and
late afternoon) and at high satellite view angles (e.g., high northern or southern latitudes).
Furthermore, the flexibility of scanning throughout the day allows for greater opportunity to
obtain non-cloudy pixels at any given location due to diurnal variability in cloud cover (e.g.,
morning fog along the Pacific northwest and afternoon clouds off the Florida coast).
In spring 2009, a NASA Geo-CAPE Oceans Science Working Group (SWG) was assembled to
help define a science traceability matrix (STM) for the oceans component of the Geo-CAPE
mission. The STM summarizes the science questions, approach, measurement requirements and
instrument requirements for the mission to the scientific community (see Table A1 in the
appendix for a complete STM). The SWG endorsed the current draft of the coastal ecosystems
dynamics STM at the March 2010 Geo-CAPE working meeting. The STM was presented to the
ocean color community at the NASA Ocean Color Research Team meeting in May 2010
(http://oceancolor.gsfc.nasa.gov/MEETINGS/OCRT_May2010/). The SWG proposed the
following set of science questions traceable to the NRC DS (2007) and NASA’s Ocean Biology
and Biogeochemistry (OBB) Program long-term planning document (NASA 2006) that GeoCAPE can address.
● How do short-term coastal and open ocean processes interact with and influence larger scale
physical, biogeochemical and ecosystem dynamics?
● How are variations in exchanges across the land-ocean interface related to changes within the
watershed, and how do such exchanges influence coastal and open ocean biogeochemistry and
ecosystem dynamics?
● How do natural and anthropogenic changes including climate-related forcing impact coastal
ecosystem biodiversity and productivity?
● How do airborne-derived fluxes from precipitation, fog and episodic events such as fires, dust
storms & volcanoes significantly affect the ecology and biogeochemistry of coastal and open
ocean ecosystems?
● How do episodic hazards, contaminant loadings, and alterations of habitats impact the biology
and ecology of the coastal zone?
Q1. How do short-term coastal and open ocean processes interact with and influence larger
scale physical, biogeochemical and ecosystem dynamics?
The large-scale response of ocean circulation, biogeochemistry and ecosystems to atmospheric,
climatic and anthropogenic forcing is the integral of processes occurring on smaller scales.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Examples include vertical mixing, upwelling, primary production, grazing, as well as turbulent
kinetic energy processes that can occur on inertial and semi-diurnal tidal frequencies (1-2 orders
of magnitude higher than surrounding portions of the energy spectrum). Some of these processes
are not easily discernible by the current generation of polar orbiting ocean color satellite sensors.
Satellite missions such as Geo-CAPE, with associated field campaigns, will help determine
precisely, how these small scale processes operate and impact biology, allowing for
parameterization in larger scale predictive models. The interplay of these dynamic physical,
chemical, and biological processes drives the transfer of matter and energy on regional and
global scales, affecting Earth’s climate as well as human health and prosperity. Natural and
anthropogenic perturbations alter these processes at spatial scales ranging from microscopic to
basin-wide, and on time-scales ranging from minutes to decades and longer. Microorganisms at
the base of the food web play a key role in integrating these scales, translating short-lived events
to more easily recognized patterns and cycles of productivity observable at the ecosystem level.
Gaining a more detailed understanding of how microorganisms respond to highly ephemeral,
spatially varying processes is therefore central to attaining a predictive understanding of the
longer term and larger scale consequences of environmental change.
Q2. How are variations in exchanges across the land-ocean interface related to changes within
the watershed, and how do such exchanges influence coastal and open ocean biogeochemistry
and ecosystem dynamics?
Exchanges of waterborne materials from land to ocean are a function of seasonal discharge
dynamics, atmospheric deposition and land surface attributes that are influenced by a host of
natural and anthropogenic processes. Land to ocean subsidies of nutrients, labile carbon, light
attenuating substances and pollutants can have profound effects on estuarine and coastal
ecosystems. Wetlands, estuaries and river mouths at the land-ocean interface are regions of
vigorous biogeochemical processing and exchange, where land-derived materials are
transformed to other compounds, affecting fluxes of carbon and nutrients to both the coastal
ocean and the atmosphere. Changing climate, land use practices and air quality have the potential
to alter coupled hydrologic-biogeochemical processes and the associated movement of water,
carbon and nutrients through various terrestrial reservoirs. Such changes will ultimately
influence the delivery of dissolved and particulate materials from terrestrial systems into rivers,
estuaries, and coastal ocean waters.
Watershed processes must be considered over spatial extents of hundreds to millions of square
kilometers and varying time scales (hours to decades) in order to adequately characterize
relationships among climate forcing and use practice/land cover change, and transport of
materials through watersheds and, ultimately, to coastal regions. Within the coastal domain,
similar spatial and temporal scales are required to characterize dynamics of terrestrial
constituents and their effect on ecosystems. The objectives of describing processes controlling
fluxes on land, their coupling to riverine systems, delivery of materials to estuaries and the
coastal ocean and corresponding coastal ecosystem responses, necessitate the use of an
integrated suite of models and remotely sensed data and targeted in situ observations.
Q3. How do natural and anthropogenic changes including climate-related forcing impact
coastal ecosystem biodiversity and productivity?
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
How climate variability and global change will impact the biodiversity and productivity of
coastal ecosystems is still the subject of significant debate. Coastal ecosystems (out to 150 km
from shore) account for about 15% of global primary production but their importance is that they
provide the great majority of marine living resources that are harvested for human consumption.
Coastal ecosystems also receive the great majority of anthropogenic inputs (except CO2) due to
their proximity to coastal human populations. The impacts on coastal ecosystems are
hypothesized to occur via straightforward bottom-up nutrient supply processes. Coastal PP,
diatoms and fish and other consumers all should decrease when (1) upwelling or other nutrient
supply processes decrease, (2) nutrient levels in the thermocline/nutricline decrease, and (3) the
thermocline/nutricline deepens. While these biogeochemical links are currently observable at
longer time scales using polar orbiting satellites such as MODIS and SeaWIFS, Geo-CAPE
observations will provide critical data linking the inertial and semi-diurnal frequency variability
in ocean processes to the spectrum of biological response. Moreover, it will resolve strictly
biological variability at sub-diurnal time scales (e.g. dinoflagellate vertical migration,
phytoplankton growth) currently not resolvable using sun-synchronous polar orbiting satellites.
Thermal stratification over a very shallow thermocline/nutricline for instance can foster coastal
dinoflagellate blooms. These hypotheses require careful observation and testing that can be
addressed with Geo-CAPE and associated field support.
Q4. How do airborne-derived fluxes from precipitation, fog and episodic events such as fires,
dust storms & volcanoes significantly affect the ecology and biogeochemistry of coastal and
open ocean ecosystems?
Atmospheric fluxes influence marine ecosystems in two ways, via direct deposition to the
surface of marine waters and via indirect deposition to the watersheds emptying into those
waters. Watershed processes have been discussed above in Science Question 2, here we will
focus on direct deposition that can be in the form or wet or dry deposition. Although the
atmosphere is often referred to as a passive vector in delivering material from terrestrial activity
to the ocean, it has been shown that “cloud processing” – the transformation of materials due to
interactions in the atmosphere – is important to understand bioavailability and effects on the
ocean ecosystem (Hand et al 2004). For example, two key nutrients, nitrogen and iron, are
known to have significant atmospheric depositional sources that are highly episodic. Dust
storms are known to deposit significant amounts of iron both to open ocean and coastal ocean
waters via dry deposition of dust aerosol particles. Similarly, recent work has indicated volcanic
ash may also be a significant source of iron in some ocean waters via aerosol deposition. Unlike
aeolian deposition of iron, nitrogen deposition is more important in coastal waters than open
ocean areas due to the proximity of coastal ecosystems to anthropogenic source regions.
Precipitation and fog are known to efficiently scavenge both inorganic and organic forms of
nitrogen from the atmosphere enhancing their deposition by several orders of magnitude over
that which occurs via dry deposition processes alone. However, the extent of the influence of
atmospheric deposition of any of these materials depend strongly on many other factors within
the aquatic ecosystem and it is not a simple matter to translate atmospheric wet deposition fluxes
into ecological importance. In addition to nutrient delivery, other atmospheric constituents may
be deposited to ocean waters from both natural (e.g., biomass mass burning) and anthropogenic
(e.g., agricultural, industrial, residential, and combustion) sources. As with nitrogen deposition,
various compounds from anthropogenic sources are expected to exert a greater influence in
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
coastal waters than the open ocean due to their proximity to sources. Recently, it has been
shown that atmospheric deposition of nitrogen and sulfur could also lower pH of coastal waters
worsening the coastal ocean acidification problem (Doney et al 2007). Atmospheric constituents
of greatest concern from such sources in coastal regions include toxic organic compounds (e.g.,
polycyclic aromatic hydrocarbons), trace metals (e.g., mercury), and persistent organic pollutants
(e.g., polychlorinated biphenyls (PCBs)). Much work remains to understand how all the various
compounds delivered to marine ecosystems from the atmosphere are biogeochemically cycled
within those systems determining their ultimate fates.
Q5. How do episodic hazards, contaminant loadings, and alterations of habitats impact the
biology and ecology of the coastal zone?
The effects of episodic hazards, such as hurricanes and other extreme storms, floods, tsunamis,
chemical spills, harmful algal blooms, which occur often without warning and frequently are of
short duration, are especially challenging to observe. Yet it is these same events that have the
most severe and lasting effects on coastal ecosystems. Other severe impacts resulting from loss
of coastal marshlands due to development and sea level rise occur so gradually over such long
periods of time that they are likewise difficult to observe. In both cases, a geostationary
observation platform provided by Geo-CAPE will permit more detailed assessment of the extent
and duration of damage to coastal habitats from disasters.
The recent Deepwater Horizon oil disaster, which has both episodic and long-term effects on the
environment, is one example where the Geo-CAPE mission would have been extremely
valuable. Effective response and prediction relies on accurate and timely information that is
updated frequently. Assessment of impacts on coastal and open ocean communities requires
both standing stock and rate measurements over many years.
Data products
The coastal ocean ecosystem dynamics data products that will be generated from Geo-CAPE
observations are described in Table 1. The data products are classified as mission critical or
highly desirable and also in terms of the maturity of the products based on current ocean color
retrievals: climate data record (CDR), candidate CDR, research products, and exploratory
products.
Table 1. Classification of Satellite Data Products for GEO-CAPE Coastal Ocean Ecosystem
Dynamics. Mission critical: products that drive measurement and instrument requirements.
Highly desirable: products relevant to addressing mission science questions but not critical
because the retrieval algorithm and/or field/lab measurement is not mature. The color code
denotes the maturity of the satellite product.
Mission Critical
Normalized Spectral Remote Sensing
Reflectances (and normalized water-leaving
radiances)
Chlorophyll a
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
Highly Desirable
Particle size distributions & composition
(biogenic, mineral, etc.)
Physiological properties (fluorescence quantum
yields, etc.)
DRAFT
November 17, 2010 – V1.6
Diffuse attenuation coefficient (490 nm)
Inherent optical properties & products:
Colored Dissolved Organic Matter
(CDOM) absorption
Particle absorption & scattering,
Phytoplankton and detritus absorption
& scattering
Euphotic depth
Primary production
Particulate Organic Carbon (POC)
Particulate Inorganic Carbon (PIC)
Photosynthetically available radiation (PAR)
Fluorescence line height (FLH)
Total suspended matter (TSM; coastal)
Trichodesmium concentration
Dissolved Organic Carbon (DOC; coastal)
Phytoplankton carbon
HAB detection and magnitude
Other plant pigments (carotenoids,
photoprotective pigments, photosynthetic
pigments, phycobilins, etc.)
Net community production of POC
Net community production of DOC
Export production
Terrigenous DOC
Black carbon
pCO2(seawater)
Air-Sea CO2 fluxes
Photooxidation
Detection of vertically migrating phytoplankton
Respiration
Functional/taxonomic group distributions
Petroleum detection, type & thickness
CDR = Climate Data Record, Candidate CDR, Research products, Exploratory Products
Approach
The SWG has recommended an ocean sensor capable of accomplishing three observational
approaches: (1) targeted, high-frequency, episodic event-based monitoring and evaluation of
tidal and diurnal variability of upper ocean standing stocks, rate measurements and hazards from
river mouths to the coastal ocean, (2) survey mode that measures diurnal, seasonal and interannual variations in the standing stocks, rate measurements and hazards for estuarine and
continental shelf regions with linkages to open-ocean processes at appropriate spatial scales (Fig.
1), and (3) to observe coastal regions at sufficient spatial scales to resolve near-shore processes,
coastal fronts, eddies, and track carbon pools and pollutants. The measurement and instrument
requirements necessary to address the science questions, observational approaches, and
measurements are summarized in Table 2. These requirements are consistent with the
requirements for Geo-CAPE recommended by the NRC panel (NRC 2007) and a geostationary
ocean color mission described in the OBB planning document (NASA 2006).
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Figure 1. Planned coastal coverage and conceptual scan scenes for the Geo-CAPE coastal ecosystems sensor.
Other ocean scenes can be scanned as necessary. Location shown is 90W at equator; figure will be revised to
center at 95W.
Measurement and Instrument Requirements
Coastal oceans are the most productive ecosystems in terms of primary, secondary, and tertiary
production as they receive nutrient supplies from river discharges, non-point runoff, upwelling,
and atmospheric deposition. Coastal oceans are more dynamic in time and more heterogeneous
in space than the open ocean, therefore more demanding in measurement frequency (temporal)
and resolution (spatial). The objective of differentiating the various phytoplankton functional
groups (PFGs) also demands more spectral bands. While existing polar-orbiting multi-band
instruments provide near-daily observations of the surface ocean at ~1-km resolution, the key
science questions of the Geo-CAPE mission can only be answered through enhanced
measurements of the spectral radiance with sufficient temporal frequency and spatial resolution.
This measurement requirement further puts strict requirement on the instrument design.
The threshold and goal requirements constitute the trade space within which the science
objectives can be achieved. An optimal spatial resolution to resolve coastal ocean geophysical
features (and hence in-water constituents) would be <200 to 100 m (ground sample distance;
GSD) for turbid waters within 10km of the shore (Bissett et al. 2004; Davis et al. 2007).
However, beyond 10km from shore, a GSD of 1km would be sufficient to resolve geophysical
features (ibid). Since spatial resolution represents one of the principal drivers of instrument size
and mass, a compromise must be made between resolving in-water constituents within the near
shore and developing a geostationary satellite sensor that is both reasonable in size and mass and
technologically feasible. The nadir threshold spatial resolution of 375 m represents the coarsest
resolution acceptable to image estuaries and their larger tributary rivers (e.g., Chesapeake Bay
and the Potomac River) as well as resolve eddies, coastal fronts, and moderately sized
phytoplankton patches (Dickey et al. 1991).
High frequency satellite observations are critical to studying and quantifying biological
and physical processes within the coastal ocean. Current satellite-based products of ocean
primary production, which represents the rate of carbon fixation in phytoplankton or
photosynthesis, currently rely on no more than a single satellite observation per day of
chlorophyll and other ancillary products. Due to cloud cover and gaps in coverage of LEO
sensors such as SeaWiFS and MODIS, the number of satellite observations over an ocean region
is reduced to only a few measurements per week at best. Since phytoplankton blooms develop
over the course of a few days to a week, the complete dynamics of the blooms (such as the onset,
peak and termination of the bloom are not captured by individual LEO sensors). Yet, the in situ
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
derived primary production (PP) measurements used to validate this satellite product quantify PP
over a 6 to 24 hour period. Furthermore, the physiology of phytoplankton cells (chlorophyll
content, nutrient uptake, etc.) varies on a diel cycle, and this has a significant impact on the their
growth rate and hence PP. Therefore, multiple observations per day over several days are
necessary to derive robust satellite-based estimates of PP. Since tidal currents reverse within ~6
hours for semi-diurnal (and ~12 hours for diurnal tidal cycles), tracking hazards such as oil slicks
or harmful algal blooms using a satellite sensor requires a minimum of three observations per
day distributed 3 hours apart (Davis et al. 2007).
The instrument requirements necessary to meet the science objectives and measurement
requirements for the coastal ecosystem dynamics component of the Geo-CAPE mission are
shown in Table 2. The broad spectral range permits atmospherically-corrected retrievals of
spectral remote sensing reflectances from the ultraviolet (UV) to near-infrared (NIR). The
contribution of water-leaving radiances to the top-of-the atmosphere (TOA) is typically <10%.
This requires atmospheric correction to account for molecular scattering (Rayleigh), gaseous
absorption (ozone, water vapor, oxygen, nitrogen dioxide), and aerosol scattering and absorption
to the TOA radiances. The spectral range and resolution requirements have been established to
provide appropriate atmospheric correction to the TOA radiances. Coastal data products will
require spectral remote sensing reflectances from 350-760nm with a high signal-to-noise ratio
(Table 3).
Table 2: Measurement and instrument requirements for the Geo-CAPE coastal ecosystems
dynamics sensor.
Measurement Requirements
Temporal Resolution
Threshold
Goal
Targeted Events
1 hour
0.5 hour
Routine Coastal U.S.
0.5 hour
3 hours
Region of Special Interest
1 RSI at 3 scans/day
multiple RSI at 3 scans/day
(RSI)
Other Coastal waters 50°N to
4 times/year
3 hours
45°S
375 m x 375 m
250 m x 250 m
Spatial Resolution (nadir)
Field of Regard for Ocean
50°N to 45°S;
Color retrievals
162.5°W to 32.5°W
Coastal Coverage (distance
375 km
500 km
from coast to ocean)
Instrument Requirements
345-900 nm; 3 SWIR bands
340-1100 nm; 3 SWIR bands
Spectral Range
1245, 1640, 2135 nm
1245, 1640, 2135 nm
UV-VIS: 0.5 nm FWHM; NIR: 0.25 nm FWHM; NIR: 0.5 nm;
Spectral Resolution
1 nm; SWIR: 20-50 nm
SWIR: 20-50 nm
Signal-to-Noise Ratio (SNR) 1000:1 for 10 nm FWHM (380- 1500:1 for 10 nm (380-800
800 nm); 600:1 for 40 nm
nm); 600:1 for 40 nm FWHM
FWHM in NIR; 300:1 to 100:1
in NIR; 300:1 to 200:1 for
for SWIR bands (20-50nm
SWIR bands (20-50nm
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
FWHM)
FWHM)
±9° N to S & E to W imaging
Field of Regard
capability from nadir for Lunar
& Solar Calibrations
±70°
±75°
Solar Zenith Angle Sensitivity
High
sensitivity
but
nonDetector Sensitivity &
saturating
Saturation
<0.5%
Polarization
<25% of pixel size during within 10% of pixel size during
Pointing Stability
single exposure
single exposure
(pixel tracking)
1%
through
mission
lifetime
<0.5%
through
mission lifetime
Relative Radiometric
Precision
Complete characterization to
Pre-launch Characterization
achieve relative radiometric
precision on orbit
Monthly at ≤7° phase angle
Lunar Calibration
3 years
5 years
Lifetime Design
Table 3: Minimum requirements for spectral bands and SNR to retrieve mission critical data
products and necessary atmospheric corrections.
Band
Center1
Bandwidth1
Minimum
SNR
Application/Comments1
350
15
500
Absorbing aerosol detection
360
15
500
CDOM-chlorophyll separation; strong NO2 absorption
385
10
1000
CDOM-chlorophyll separation; strong NO2 absorption; avoid
precipitous drop in solar spectrum at 400 nm
410
20
500
NO2 absorption
412
10
1000
CDOM-chlorophyll separation; SeaWiFS (20 nm) & MODIS (15
nm) bands; strong NO2 absorption
425
10
1000
CDOM-chlorophyll separation, strong NO2 absorption
443
10
1000
Chlorophyll-a absorption peak; SeaWiFS (20 nm) & MODIS (10
nm) bands; strong NO2 absorption
460
10
1000
Assessory pigments & chlorophyll
475
10
1000
Assessory pigments & chlorophyll
490
10
1000
SeaWiFS (20 nm) & MODIS (10 nm) bands; chlorophyll bandratio algorithm
510
10
1000
SeaWiFS (20 nm) band; chlorophyll-a band-ratio algorithm; strong
O3 absorption
532
10
1000
Aerosol lidar transmission band; MODIS (10 nm) band; strong O3
absorption
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
540
10
1000
Phycoerythrobilin pigment
555
10
1000
Bio-optical algorithms (e.g., band-ratio chlorophyll); MODIS-547
nm, SeaWiFS-555 nm; strong O3 absorption
583
10
1000
Phycoerythrin, strong O3 absorption
623
10
1000
Strong O3 absorption; Phycocyanin, cyanobacteria,
625
10
1000
Phycocyanin
640
10
1000
Between O3 & water vapor absorption peaks
655
10
1000
Chlorophyll a&b, strong O3 absorption, weak water vapor
absorption
665
10
1000
Fluorescence line height baseline, bandwidth constrained by water
vapor absorption line & 678 band
678
10
1000
Fluorescence line height; HABs detection; band center offset from
fluorescence peak by O2 absorption line
710
10
1000
Fluorescence line height; HABs detection; terrestrial "red edge";
straddles water vapor absorption band
748
10
600
Fluorescence line height baseline; Atmospheric correction-open
ocean; MODIS band, between O2 A-band & water vapor
absorption peaks
765
40
600
Atmospheric correction-open ocean; SeaWiFS band, O2 A-band
absorption
820
15
600
Water vapor concentration/corrections. There are other water
vapor absorption features that could be used.
865
40
600
Atmospheric correction-open ocean; SeaWiFS band (40 nm
bandwidth); MODIS band-869 (15 nm bandwidth)
1245
20
300
Atmospheric correction-turbid water; MODIS band; bandwidth
constrained by water vapor & O2 absorption peaks
1378
20
300
Detection and correction of cirrus clouds and high-altitude aerosols
1640
40
250
Atmospheric correction-turbid water
2135
50
100
Aerosol properties, turbid water aerosol correction
1
Band centers, Bandwidth and Applications obtained from C. McClain per ACE Ocean requirement; Revisions
applied per Geo-CAPE SWG.
Platform requirements
The requirement of multiple views per day can only be achieved on a geosynchronous platform.
Further, to observe coastal waters from 50°N to 45°S over the entire continental U.S., Central
America and South America, the platform needs to be located near 95oW on the equator.
Ancillary data requirements
A successful ocean color mission requires accurate pre-launch calibration/characterization and
post-launch (in-orbit) vicarious calibration as well as extensive field measurements to validate
the satellite measurements. Further, for accurate atmospheric and other corrections, the
following measurements, often from other satellite missions, are required: (1) ozone; (2) total
water vapor; (3) surface wind velocity; (4) surface barometric pressure; (5) NO2 concentration.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
To address the key science questions of the Geo-CAPE mission, the following measurements are
also required from other satellite missions, field measurements, and/or numerical models:
(1) sea surface temperature (SST); (2) sea surface height (SSH); (3) ultraviolet radiation; (4)
mixed layer depth (MLD); (5) atmospheric and surface ocean CO2; (6) pH; (7) ocean circulation;
(8) tidal and other coastal currents; (9) aerosol and dust deposition; (10) run-off loading in the
coastal zone; (11) wet deposition in the coastal zone.
Benefits to society / societal impacts
The societal benefits of ocean color have been extensively detailed in Report #7 (2008), “Why
Ocean Colour? The Societal Benefits of Ocean Colour Technology” of the International Ocean
Colour Coordinating Group (IOCCG), as well as in IOCCG Report #8 (2009), “Remote Sensing
in Fisheries and Aquaculture”. As addressed in those previous reports and numerous other
documents cited therein and elsewhere, ocean color can be utilized to support a number of
important research and applied or operational efforts such as: assessments of climate variability
and change through improved understanding of biogeochemical cycles (e.g., carbon pools and
fluxes) and food web impacts; integrated ecosystem assessments and living marine resource
management, e.g., marine protected areas, fisheries, aquaculture and threatened/endangered
species; monitoring of coastal and inland water quality, e.g., pollutant and pathogen–laden runoff
plumes and spills; assessments of natural and anthropogenic hazards, e.g., harmful algal blooms,
oil and sewage spills, sediment resuspension events; improved understanding of ocean and
coastal dynamics, e.g., eddies and blooms; development of robust indicators of the state of the
ocean ecosystem, and, ecological modeling and forecasting activities.
In support of these efforts, ocean color observations from a geostationary platform such as GeoCAPE will provide significantly improved temporal coverage of nearshore coastal, adjacent
offshore and inland waters, and likely improved spatial and spectral coverage relative to current
LEO sensors, which are generally more focused on global observations of open ocean waters.
The higher frequency observations from geostationary will help mitigate the effects of cloud
cover, as well as better resolve the dynamic, episodic, and/or ephemeral processes, phenomena
and conditions commonly observed in coastal regions. A denser and more comprehensive ocean
color data set will result, resulting in the further development, use and operational
implementation of more timely and accurate products, e.g., harmful algal bloom forecasts, which
in turn will provide better information to users in support of management and decision/policy
making needs.
Instrument and mission design studies
Geo-MDI
In preparation for the release of the NRC report, NASA conducted a series of mission planning
and engineering studies in Autumn 2006 – Advanced Earth Science Mission Concept studies.
The instrument developed for this study, Geostationary Multidisciplinary Imager (Geo-MDI),
had similar requirements to those shown in Table 2 with the exception of the nadir spatial
resolution of 250 m and the incorporation of additional requirements for retrieval of atmospheric
trace gases and terrestrial vegetation products. A mission design study called Geostationary
Multi-discipline Observatory (GMO) integrated mission #5 followed the instrument design
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
studies and integrated three sensors (Geo-MDI, GeoMac and CISR) to simulate a Geo-CAPE
mission payload (see http://geo-cape.larc.nasa.gov/documents.html for a summary).
Figure 4. Volume comparison of the Geo-MDI (geostationary multi-disciplinary imager) instrument
developed in Fall 2006 in the Instrument Design Lab (IDL) study at Goddard Space Flight Center with the
Coastal Ecosystem Dynamics Imager (CEDI) instrument design concept developed in the IDL study
conducted at GSFC in January 2010. The two primary factors that resulted in the reduction in volume and
mass between the two studies is (1) coarsening of the spatial resolution from 250x250m for Geo-MDI to
375x375m for Geo-CEDI and (2) removal of a secondary UV-Vis focal plane designed for atmospheric
retrievals of trace gases.
Geo-CEDI
A follow-on Instrument Design Lab (IDL) study was conducted at GSFC in January 2010 to
design an instrument that meets or exceeds the specific threshold requirements established by the
Oceans SWG and with a reduced size and mass than Geo-MDI (Fig. 2). The Coastal Ecosystems
Dynamics Imager (CEDI) design developed from this IDL study meets or exceeds nearly all of
the threshold requirements for the mission documented in the STM. CEDI has two viewing ports
one with a 2-sided diffuser plate for solar calibrations and a second viewing port for lunar
calibrations and nadir science (Fig. 3). Fig. 4 shows the layout of the optics and associated major
subsystems.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Figure 3. Mechanical drawing of exterior of (CEDI) showing the solar and lunar calibration views and nadir
science view.
Figure 4. Optical layout of the Coastal Ecosystem Dynamics Imager (CEDI) instrument design concept
developed in the Instrument Design Lab (IDL) study conducted at Goddard Space Flight Center in January
2010.
There are three separate spectrographs in Offner configurations that cover the spectral regions
required for the various ocean color products and atmospheric correction algorithms. Table 3
provides the spectral coverage and resolution parameters for each band.
Table 3. Geo-CEDI spectral design parameters.
Spectral Design Parameters
Coverage [nm]
Resolution [nm]
Band 1
345-600
0.5
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
Band 2
600-1100
0.5
Band 3
1200-2200
2.5
DRAFT
November 17, 2010 – V1.6
Each spectrograph band has an associated 2-dimensional focal plane optimized for sensitivity
and dynamic range for that particular portion of the spectrum. One dimension of the focal plane
is used for spatial discrimination while the second is used for spectral discrimination. In our
baseline design all three focal planes are hybridized photodiode/readout integrated circuit
(ROIC) structures using silicon (Si) detectors for the UV/VIS/NIR Bands and mercury cadmium
telluride (HgCdTe) detectors for the SWIR Band. The typical integration times for Band 1 and
Band 2 will be 200 ms and 12.5 ms for Band 3 to avoid saturation under cloudy conditions. The
spectrographs share a common telescope via the use of two beam-splitters and the 2-axis scan
mirror then projects a co-aligned image of the three slits onto the Earth’s surface. The current
design has been optimized to obtain diffraction-limited performance in all three Bands and meets
the spatial ground sampling and coverage requirements outlined in the following section.
Table 4. Geo-CEDI scanner timing and spatial sampling.
Sample Scale
Scan Line
Field of Regard
Sample angular Sample Time Comments
size
2.2 arc-sec x 1.25 0.5 to 1.0 sec E-W stepping (e.g., 0.8 sec Dwell for
deg.
image co-adding + 0.2 sec step and
settle)
0.625 x 1.25 deg. 8.5 to 17.1
Relocate/Revisit Coastal Scenes
minutes
(Nominally 72 to 84 scenes/day)
The basic geometry and timing of the data accumulation is given in Table 4. Under threshold
science requirements a single pixel field-of-view is 2.2 arc-sec, and this resolution must be
maintained over an image sample dwell period of 0.8 seconds. CEDI has a north-to-south slit
orientation (2048 pixels) and therefore scans east-to-west through step movements of the scan
mirror. One of the challenges in designing an instrument capable of meeting the ocean threshold
requirements from a geostationary orbit is the balance between achieving the required SNR
(1000:1 in UV-Vis) over a large dynamic range (Lmax/Ltyp of 3 to 70 from UV to SWIR) while
avoiding saturation of the detectors over bright scenes (e.g. imaging over clouds) and
maintaining a short integration time of 1 second (0.8 sec scanning plus 0.2 sec for scan mirror
stepping) for each line of pixels scanned. This challenge was overcome by co-adding
consecutive multiple scans per line of pixels. For example, at typical ocean radiances at 70
degree solar zenith angle (SZA) would require co-adding two 0.4 sec scans for Bands 1 and 2
and 46 scans for Band 3 yielding a total integration time of 17.1 minutes for a scene size at nadir
of 384 km x 768 km. Other scenarios can be envisioned which would reduce or increase the scan
line dwell period (e.g. special events imaging of high impact that would require faster revisit
times or unique processes which require more precision and longer dwell periods).
SUMMARIZE 2010 MISSION STUDY
Sensitivity analyses
Atmospheric corrections M Wang, Tzortziou, Herman
NEED INPUT FROM MENGHUA ON OZONE AND AEROSOLS
II.
a.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
From Jay Herman and Maria Tzortziou
In order to retrieve underwater composition it is necessary to measure radiances at the top of the
atmosphere (TOA) with a precision of 0.1% (in wavelength) and to within the NIST transfer
standard for absolute radiometric accuracy (2 to 3%). However, absolute radiometric accuracy
affects retrieval algorithms much less than loss of spectral precision. The 0.1% spectral precision
is based on prior experience with MODIS and SeaWiFS, and is the requirement needed for the
retrieval algorithms. Problems with measurement precision arise from instrumental sources
(intrinsic signal to noise ratio based on the detector technology, readout noise, optical
aberrations, and stray light) and from errors in atmospheric corrections needed to derive the
underwater leaving radiances. For the rest of this discussion, it will be assumed that the
instrumental characterization meets the measurement requirements (SNR > 1000:1, stray light
less than 0.1%).
The main sources of atmospheric correction arise from ozone absorption, aerosol scattering, and
aerosol absorption. In coastal regions, NO2 absorption is an equally important source of retrieval
error. O3 absorption and aerosol scattering can be handled by standard methods based on GEO
CAPE measurements. Absorbing aerosol amounts are presently unknown for UV wavelengths,
but can be derived from GEO CAPE measurements for cloud-free scenes. Aerosols and trace gas
(mainly NO2) amounts require ground-based validation, which will be discussed in another
section.
Absorbing aerosols in coastal regions typically have an optical depth of a few tenths with weak
wavelength dependence. In cloud-free scenes, the aerosol refractive index and optical depth
could be derived. However, the underwater signal introduces a wavelength dependence that
interferes with the aerosol retrieval. In the absence of independent ground-based estimates of
aerosol absorption, the aerosols may introduce an error in the retrieval of underwater
composition.
In coastal waters, in addition to aerosols, atmospheric correction is needed for the presence of
tropospheric trace gases such as NO2. Nitrogen dioxide has a strong absorption spectrum in the
range 330 to 500 nm that varies rapidly with wavelength. Tropospheric NO2 amounts are highly
variable in time and space, caused by both natural and anthropogenic emissions. Based on
satellite observations (e.g. from OMI, GOME, SCIAMACHY), NO2 amounts in coastal waters
near urban areas typically range from 0.2 to 0.8 DU (1 DU = 2.67 x 1016 molecules/cm2), but
they can exceed 1 DU. These numbers apply to the east coast of the US from Boston, Mass. to
Norfolk, Va., and to specific sites between Norfolk and the tip of Florida.
One of the main challenges using current satellite NO2 data for atmospheric correction of coastal
ocean color is their coarse spatial resolution (e.g. 12 km x 24 km at nadir view for OMI and even
coarser for GOME and SCIAMACHY) relative to satellite ocean color measurements (e.g. 1 km
x 1 km for MODIS) and expected variability at the land-ocean interface. Much higher-resolution,
higher-frequency data from Geo-CAPE could contribute to understanding small-scale, short-term
variability in coastal ocean processes and atmospheric composition. The question then arises
about the accuracy needed in the NO2 measurement. Figure 5 shows radiative transfer
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
calculations of the percent error in water leaving radiances per DU of NO2 for NO2 altitude
profiles 0 – 2 km, 0 – 3 km, and 0 – 4 km. If NO2 is mostly in the lower 2 km, then the error at
400 nm ranges up to 17% per DU for large solar zenith angle plus viewing zenith angle (SZA +
VZA), which define the air-mass for photon transmission through the atmosphere. If we have an
error of 0.2 DU, the error will be about 3.4% at 400 nm, and higher (up to 8%) if the NO2 is
distributed between 0 and 4 km. In general, we should know the NO2 amount to within 0.2 DU to
provide an adequate atmospheric correction.
A slight error in the wavelength independent estimate for Fresnel reflection arising from an
incomplete model of ocean surface conditions and wind speed will affect all wavelengths in a
narrow range equally and not affect retrieval algorithms substantially. One component of
wavelength dependent error arising from an incorrect estimate of Fresnel reflection is the amount
of Rayleigh scattering back into the water with a l-4 wavelength dependence. This correction is
dependent on the estimated wave structure derived from wind-speed estimates.
Figure 5. Radiative transfer calculations of the percent error in water leaving radiances per DU of NO 2
for NO2 altitude profiles 0 – 2 km, 0 – 3 km, and 0 – 4 km.
Uncertainties for lab/field measurements (summarize ACE/Geo-CAPE activity)
A joint activity between the Geo-CAPE Oceans SWG and the Aerosol, Cloud and ocean
Ecosystem (ACE) Oceans SWG is underway to quantify uncertainties for field and laboratory
measurements.
b.
c.
Uncertainties of satellite products
Products generated from Geo-CAPE, because of its remote-sensing nature, inherently contain
some degree of uncertainties. The sources of uncertainties range from imperfect sensor
engineering to retrieval algorithms [Antoine et al., 2008; Boss and Maritorena, 2006; Chami and
Defoin-Platel, 2007; Salama and Stein, 2009]. Conventionally, the uncertainty (sometimes called
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
“error”) of a product (such as Chl or Kd) has been evaluated statistically by comparing retrieved
values from remote sensing with those from water samples measured concurrently, and an
averaged “error” for the entire dataset [Darecki and Stramski, 2004; O'Reilly et al., 1998], or
subgroups [Moore et al., 2009], has been derived. This quantity of “error” provides a general
picture of the consistency between the remotely-sensed and the measured properties. It does not,
however, isolate individual sources that contribute to the total uncertainty of a remotely-sensed
data product.
Derivation of ocean products from ocean-color measurements is a complex process [McClain et
al., 2000], and is generally divided into two separate, but consecutive, steps: one is atmosphere
correction [Gordon and Wang, 1994]; the other is ocean-color inversion [Carder et al., 1999;
IOCCG, 2006; O'Reilly et al., 1998]. Each one of the two steps involves many sub-steps and/or
variables, and contains different sources and magnitudes of uncertainties. To fully understand the
quality of the ocean products derived from Geo-CAPE and then to characterize the spatial
distribution of the uncertainties, it is necessary to account for all sources of uncertainties and
quantify their contributions to uncertainties of the final products pixel wise. Specifically, it thus
requires the development of a system to 1) analyze/quantify uncertainties of Geo-CAPE
produced water-leaving radiance (the key product for derivation if in water properties) and 2)
analyze/quantify uncertainties associated with the generation of ocean products from waterleaving radiance, per pixel [Lee et al., 2010; Maritorena et al., 2010].
Calibration & Validation
Background/ Calibration Strategy
III.
a.
Unlike land surface or clouds, the ocean reflectance is very small (often < 1%) due to strong
absorption of the water molecules and other optically active constituents (OACs). Consequently,
of the satellite measured (i.e., top-of-atmospheric or TOA) signals, only a small portion comes
from the ocean. This portion varies with wavelength and concentrations of OACs but is typically
< 10% for most ocean waters, but can change substantially from <1% in the blue wavelengths for
CDOM-rich waters to > 50% in the green wavelengths for optically shallow waters. Because of
the GEO-CAPE emphasis on coastal waters where CDOM is often high due to either terrestrial
runoff or upwelling, the calibration and validation of GEO-CAPE is particularly more
demanding than previous and existing ocean color instruments.
Accurate estimates of the small ocean signal (spectral reflectance) require adequate
1. Calibration, including (a) Radiometric calibration, (b) Spectral calibration, (c) Spatial
calibration (geo-location)
2. Atmospheric correction
3. Extensive validation
Spectral calibration requires the use of standard light source in the laboratory (pre-launch), and
can be verified or adjusted using known spectral features (e.g., solar Fraunhofer lines) after
launch. Spatial calibration requires onboard GPS and the accuracy is often in the order of a half
pixel (Wolfe et al., 2002). The most critical yet challenging calibration for a successful ocean
color instrument is the radiometric calibration.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
From the experience and lessons learned from the proof-of-concept Coastal Zone Color Scanner
(1978-1986) and based on radiative transfer theory, Gordon (1987) provided a thorough review
on the calibration and methodology requirements of satellite ocean color instruments, followed
by a recent update by Dinguirard and Slater (1999). The calibration includes pre-launch
laboratory calibration (and sensor characterization) and post-launch in-orbit calibration. The
former covers radiometric calibration to a few percent, spectral (wavelength) calibration
including out-of-band response, polarization sensitivity, temperature response, and other
characterizations (see below). The latter covers sensitivity monitoring (using a constant source,
e.g., the moon) and vicarious calibration through tuning of the sensor gain by using well-defined
ground measurement and a radiative transfer model. Because the same model is used for
atmospheric correction, this latter calibration is extremely important because it represents a
system tuning that accounts for all known and unknown uncertainties in the entire measurement
system, including uncertainties in the solar extraterrestrial irradiance. Wang and Gordon (2002)
showed that as long as the system is well characterized, vicarious calibration is sufficient to
remove most errors, even if the pre-launch radiometric calibration is off by 10%. The pre-launch
radiometric calibration, however, is still important in balancing the sensitivity and dynamic
range.
The pre-launch and post-launch calibrations are discussed separately below. Briefly, pre-launch
calibration and characterization are conducted in the laboratory using NIST traceable equipment
(Johnson et al., 1999). Post-launch calibration (McClain et al., 2000) can be made using 1) a well
characterized and controlled measurement (e.g., MOBY); 2) another well calibrated in-orbit
satellite instrument; 3) modeled surface reflectance in the absence of the above two (e.g., Evans
and Gordon). All require a radiative transfer model to propagate the ocean surface signal to the
TOA signal. In addition, monitoring a stable target (e.g., the moon) is essential to provide longterm stability correction.
b.
Pre-launch Calibration & Characterization
Rigorous prelaunch characterization of the radiometric performance characteristics of a satellite
sensor cannot be over-emphasized. Once on orbit, it is very difficult or impossible to determine
quantitatively most sensor performance characteristics, e.g., out-of-band response, polarization
sensitivity, etc. The prelaunch characterization, calibration and performance validation is a
complex process, particularly as sensors become more sophisticated and uncertainty
requirements more stringent in response to expanding science objectives. The GeoCAPE Ocean
STM outlines the progression of requirements from science questions to sensor and mission
requirements. What is often overlooked is a parallel progression of sensor performance
requirements to test requirements, protocols, and metrology technology development needs.
There are many examples of missions that are delayed because testing was inadequate,
ambiguous, or simply inefficient. As the GeoCAPE formulation proceeds, calibration metrology
must be included to ensure that the necessary capability is mature when the sensor is ready for
testing.
The GeoCAPE ocean radiometer requirements are particularly demanding because the
water-leaving radiances are small, the science goals are broad, and the focus is on coastal and
estuarine waters where scales of variability are small and the range of geophysical properties is
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
greatest. Consequently, highly precise, accurate measurements of top-of-the-atmosphere (TOA)
radiances are required for mission science goals to be met. One critical activity that must be
undertaken early in the program is development of precise test requirements tied to performance
specifications; that is, detailed descriptions of the radiometric tests that must be performed to
verify that the sensor actually meets each performance specification. Once the tests are outlined,
then the procedures and test set-ups must be defined for the particular instrument design.
Different instrument designs will require, in general, tailored procedures and test fiducials.
Finally, test facility environmental requirements such as available space, air quality, and lighting
must be considered. Some tests may have special requirements that typical calibration
laboratories cannot accommodate.
Testing on both individual components as well as the full optical system is necessary;
either component-level testing or system-level testing alone is not sufficient. Component-level
testing is required for the development of a detailed instrument model that can estimate not only
the imaging and straylight characteristics, but also polarization attributes, relative spectral
response, and signal-to-noise ratios before system-level testing is undertaken. The model
provides insight on whether or not the design is adequate before final fabrication is undertaken.
Models require test data on individual components in the optical train (reflectance or
transmittance, polarization dependence, etc.) as well as information on the detectors (wellcapacity, quantum yield, dark current, read noise, etc.). Often, the specific information required
for detailed modeling, e.g., optical coating prescriptions, is not available because vendors
consider such data proprietary. This may limit the fidelity of the model. Component testing can
provide much of the necessary information for the instrument performance model. But
component-level testing by itself cannot reliably establish the sensor performance because
interactions between components (due to reflections, for example) cannot be tested. Systemlevel testing is required to validate the sensor performance requirements. Finally, all radiometric
systems and sub-systems must be fully characterized to understand the performance of the
system. If the sensor uses a solar diffuser or other device for tracking sensor degradation on
orbit, that apparatus must also be characterized at a component-level, for example the diffuser bidirectional reflectance function, and at the system level. As with the primary optical path, a
model must be developed for this element of the sensor.
In conclusion, GeoCAPE will benefit from the advances in sensor characterization
metrology as the other preceding Decadal Survey missions address the same issues. For
example, NIST will continue to refine the laser-based facility known as the Spectral Irradiance
and Radiance Responsivity Calibrations using Uniform Sources (SIRCUS) developed for the
characterization and calibration of irradiance and radiance meters (Brown et al., 2006) and the
hyperspectral image projector (HIP) to mimic the full spectral characteristics of a scene on-orbit
(Allan et al., 2009), i.e., scenes from the HIP will be used to evaluate the instrument’s pointspread correction algorithm and response to complex scenes.
c.
Post-Launch Calibration/Validation (list requirements)
Post-launch calibration and validation will at a minimum follow the approaches established for
SeaWiFS and MODIS (Barnes et al. 2001; Franz et al. 2007; Hooker et al. 2007; Eplee et al.
2010; Eplee et al. in prep.) and expand upon them due to the more advanced capabilities of the
Geo-CAPE sensor. The radiometric stability of the coastal ecosystem sensor will be monitored in
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
orbit by scanning stable targets such as the moon (Barnes et al. 2004; Stone et al. 2005) and
potentially the sun and locations on the earth. A hyperspectral sensor enables the use of solar
Fraunhofer lines for spectral calibration. Vicarious calibration of the coastal ecosystem sensor
will be accomplished with in situ radiometric measurements from several ideal sites
(oligotrophic, clean atmosphere, limited cloud cover, accessible) distributed at different latitudes
(Hooker et al. 2007). The heritage site for the Marine Optical Buoy (MOBY; Clark et al. 1997)
off the coast of Lanai, Hawaii, which has been used to vicariously calibrate MODIS and
SeaWiFS, is not a good candidate for Geo-CAPE because of the spatial distortion due the
extreme angle between the sensor and the site.
As has been proven with other ongoing ocean color satellite missions, continuous calibration and
validation of products are vital to the success of science quality data from satellites. Sustained
mission lifetime observations will occur at selected time series locations and will be augmented
by measurements made on cruises, moorings of opportunity, autonomous platforms and intensive
field campaigns. Time-series sites will be recommended based on the dynamic range of
variability of products and processes capable of being captured by Geo-CAPE. Additionally,
collaboration with current and planned calibration/validation sites for other ocean color missions
will be leveraged. Data from field observations will be archived in the NASA SeaBASS
database, which stores bio-optical and biogeochemical data collected concurrently with satellite
overpasses and uses these for continuous calibration and validation of satellite data products.
Numerous federal, state, and local government agencies are mandated to monitor the nation’s
coastal water bodies and Geo-CAPE can make a significant contribution towards effort. With its
high spatial and spectral resolution and the possibility of repeat visits each day, Geo-CAPE
overcomes some of the traditional hurdles faced by previous generation space based
measurements for routine coastal monitoring. The significant concerns about the accuracy and
validity of satellite-derived products can be addressed by using data obtained by routine
monitoring for validation of Geo-CAPE products. Chlorophyll concentrations, turbidity, and
suspended solids are all routine measurements using well established protocols on coastal ocean
observing programs conducted by federal agencies such as NOAA (e.g. Integrated Ocean
Observing System (IOOS; includes observatories in the North American East and West coasts,
Gulf of Mexico and Great Lakes [http://www.ioos.gov/catalog/]; the NowCOAST portal to a
variety of NOAA and non-NOAA data sets and forecasts [http://nowcoast.noaa.gov/]; Center for
Coastal Monitoring and Assessment [http://ccma.nos.noaa.gov/]); NSF (Long Term Ecological
Research sites (LTER; http://www.lternet.edu/sites/); NASA (e.g. Plumes and Blooms
[http://www.icess.ucsb.edu/PnB/PnB.html]; Gulf of Maine North Atlantic Time Series, GNATS
[www.bigelow.org/index.php/download_file/-/view/105];
http://seabass.gsfc.nasa.gov/seabasscgi/archive_index.cgi/BIGELOW/BALCH); EPA (e.g.
http://www.epa.gov/emap/nca/index.html and http://www.epa.gov/ncer/science/
globalclimate/); U.S Fish and Wildlife [http://www.fws.gov/coastal/], USGS (e.g. United States
Historical Climatology Network [http://cdiac.ornl.gov/epubs/ndp/ushcn/
ushcn_map_interface.html]), as well as a multitude of state environmental agencies, by local
governments, as well as private partnerships with non governmental organizations such as the
River Keeper (Waterkeeper’s alliances http://waterkeeper.org/). In addition to data available
from these coastal observing programs, high resolution validation data can also be obtained from
automated instrumentation on moorings and platforms such as those associated with the U.S.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Integrated Ocean Observing System. For example, the Gulf of Maine buoys, a part of the
Northeastern Regional Association of Coastal Ocean Observing System (NERACOOS) provide
high frequency chlorophyll concentrations and turbidity measurements.
VI. Geo-CAPE Enabling Activities (FY11 and beyond)
a. Orbital instrument design studies
Preparation for the Geo-CAPE mission requires instrument and mission design studies to ensure
that all critical technological requirements are met. For fiscal year 2011, the Geo-CAPE science
working groups have recommended that a pointing stability study be conducted to examine the
state of the technology in meeting this challenging requirement for the coastal ecosystem sensor.
Additional design studies are recommended to optimize the baseline CEDI design and to
examine other optical designs. Technology development studies funded through the NASA
Earth Science Technology Office (ESTO) can also
Mission planning scenarios – Mannino, Davis, Hu
The Geo-CAPE mission is unique in all its resolutions (spatial, temporal, spectral, and
radiometric). To fully realize its capacity for optimal measurements of different coastal
environments, mission planning is required to address the following aspects.
b.
Survey mode (“routine”) sampling
A sampling strategy will be developed to dermine when and how often to sample each
individual area shown in Fig. 1. Factors to consider include scientific questions to address
(e.g., diurnal change or temporal aliasing), historical cloud cover statistics, seasonality of
the study area, and engineering restrictions. A sampling strategy to use real-time cloud
cover measurements from concurrent weather satellites will also be pursued to determine
whether or not measurement is necessary. A data acquisition plan matrix will be
developed for optimal sampling.
Event sampling
Unexpected events, natural or accidental such as the 2002 Florida “black water” event or
the 2010 Gulf of Mexico oil spill, may occur during the lifespan of the Geo-CAPE
mission. A strategy will be implemented to deviate the routine sampling plan to target on
the events.
Download data
On-board processing
Will all spectral bands and high-frequency sampling, data volume may be a limiting
factor for real-time downlinking and processing. On-board processing options will be
studied for spectral binning [others?]
Cloud avoidance
Masking [what is this?]
c. Studies for Algorithm Development
To get the desired product for studies of ecosystem or environment, in addition to achieve
reliable/accurate measurements radiance at satellite altitude, robust algorithm(s) is the key to
convert measured optical quantities to biogeochemical products.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Historically, because ocean-color remote sensing was focused on the concentration of
chlorophyll (Chl, a proxy for biomass) in the global oceans, the algorithm to convert waterleaving radiance (Lw, the quantity after removing atmosphere contributions and effects) to Chl
has been heavily empirical in nature. Also, with a perspective of global mean (or total) Chl, one
standard (or default) algorithm (or one default sets of algorithm coefficients), which remains the
same for different locations and different phytoplankton life cycles, has been utilized for the
generation of Chl from Lw. Studies have found that this practice may not produce optimal Chl
products at some seasons, or for some specific locations, which in particular include many
coastal waters.
The Geo-CAPE mission, with its focus on US coastal waters, in particular trying to
study/understand daily variations/dynamics of community phytoplankton and other
biogeochemical properties, will thus strive to develop algorithms specifically tuned for such
objectives, instead of using one empirical Chl algorithm for all coastal waters, for example. To
achieve this goal, significant efforts are required in both theoretical evaluations and field
measurements. On the theoretical front, for instance, atmosphere correction requires information
of water vapor which is presently provided from climatological database. Geo-CAPE measures
coastal environments from morning to afternoon, which will thus face different amount of water
vapor in a day. It is then necessary to know the range of such daily variations and their likely
impacts on atmosphere correction. Otherwise false daily variation of water constituents could be
resulted.
On the derivation of biogeochemical properties in the water column (including Chl, POC, …), it
is important to obtain regional and temporal relationships between concentrations and their
corresponding spectral properties (such as spectral absorption and scattering/backscattering
coefficients). As articulated in detail in Zaneveld et al (2006), a remote sensor ultimately
measures a total signal (a spectral radiance), and this total signal is primarily controlled by
spectral total absorption and spectral total backscattering coefficients. For different coastal
environments, or at different time of a day, the water environment under Geo-CAPE observation
could have same spectral a and spectral bb, but they could be made of different components,
and/or different concentrations. Thus, how to accurately divide the total information into that of
desired products for the targeted coastal regions, and over different time of a day, remains a
challenge. To meet this challenge, for waters in the various coastal regions, not only Chl and Lw
need to be measured, the measurements (in the upper water column) should also include:
● spectral absorption of phytoplankton
● spectral absorption of dissolved organic matter
● Spectral scattering/backscattering coefficients of particles
● Particle composition and size distribution
● Suite of phytoplankton pigments
In particular, these measurements should best be made to cover daily cycles.
d. in-situ instrument development
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
GEO-CAPE pre- and post-launch algorithm development and validations goals require enabling
investments in development of in situ observational capabilities. The need for observations in a
variety of coastal environments, spanning a wide range of conditions, seasons, and times of day
will require a combination of high resolution time series sites and focused process studies. The
opportunity to take advantage of new coastal ocean observatory infrastructure should also be
exploited. Emerging optical and biogeochemical sampling technologies should be advanced as
appropriate for various platforms including moorings, piers and towers, autonomous vehicles,
ships, and small coastal boats.
i) radiometers
Hyperspectral radiometers capable of high frequency sampling will be needed both for above
water time series observations and high resolution vertical profiling in shallow waters. Above
water designs with automated solar-tracking capability and options for sky-radiance and sun
irradiance measurements, in addition to water leaving radiance, will enable necessary advances
in both in-water algorithms and atmospheric correction approaches. Wavelength ranges that
extend into the UV and infrared will also be essential.
Novel deployment strategies for hyperspectral and multispectral radiometers on aircraft and
high-speed coastal vessels will be valuable in providing high spatial resolution mapping of
representative coastal optical features and water mass types. Technologies also exist for
deployment of radiometers on autonomous profiling floats and gliders.
ii) IOP sensors
Advances in capabilities of in situ instrumentation for measurements of inherent optical
properties (IOP) permit the collection of ground truth information in support of algorithm
develop and validation of products relevant to GEO-CAPE. Measurements of fundamental
optical properties should include multispectral and hyperspectral absorption, scattering,
backscattering, and volume scattering function. Such information is critical in understanding the
basis for observed variations in remote sensing reflectance signatures and their relationship to
optical properties and associated constituents that influence them.
A key issue for long term deployment of IOP and other in situ sensors is the control of
biofouling. Progress in this area has resulted in successful deployment of IOP sensors for
extended periods between servicing. Progress in reducing the size and power requirements of
sensors has enabled longer deployments and smaller autonomous platforms including floats and
gliders.
iii) Biogeochemical sensors
In addition to optical measurements, Geo-CAPE algorithm development and validation demands
advances in sensors that provide direct or proxy information about products or constituents of
interest. Emerging technologies that characterize particles and plankton, nutrients and other
chemicals (hydrocarbons, etc.), carbon system components and pH, and rate processes
(respiration, primary production, nutrient uptake) have the potential to greatly expand the impact
of Geo-CAPE. Approaches to be explored include optical spectrometry, mass spectrometry, flow
cytometry, optical imaging, multi-isotope incubation, and other analytical techniques that can be
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
automated in a robust manner. These types of technologies will be important for developing
reliable algorithms, but also for new approaches that merge in situ and remotely sensed
information for more complete system characterization and ultimately improved predictive or
forecasting skill for complex coastal systems.
a.
directed field campaigns and immediate priorities for calibration/ validation
research
Enabling Activities
In addition, the new capabilities provided by Geo-CAPE, such as continuous coverage
over daylight hours, require field activities in addition to the standard calibration and validation
associated with other ocean color missions. Field observations with corresponding capability to
the Geo-CAPE sensor will be critical in the pre-launch phase for the purpose of developing
algorithms or identifying proxy measurements that will exploit the new capabilities of the GeoCAPE instruments. Investments that are made during the pre-launch phase will offer the
capability for calibration and validation efforts after launch. The pre-launch efforts will identify
critical in-water measurements that are required to take full advantage of the Geo-CAPE sensor.
The hyperspectral, high temporal and spatial resolution of Geo-CAPE will require
corresponding capabilities in the field at several sites in the coastal ocean. Given the
hyperspectral nature of the Geo-CAPE mission, hyperspectral radiometric measurements will be
vital. The value of the AERONET-OC network, operating SeaPRISM instruments has been
thoroughly discussed by the Science Working Group. A major drawback of the current
SeaPRISM design is the limited spectral resolution of the observed wavelengths. A
hyperspectral redesign of the SeaPRISM, or a new instrument will undoubtedly be required. In
addition to radiometry, hyperspectral observations of inherent optical properties (IOPs,
absorption and scattering) will be required to deconvolve the total radiometry measurements,
attributing the absorption and scattering contributions of the optically active constituents
(phytoplankton, non-algal particles, colored dissolved organic matter). Platforms of opportunity
and lower cost in situ time series of hyperspectral data should be investigated. Existing timeseries should be analyzed to determine if they meet requirements and recommendations provided
for improvements if needed.
The unique nature of Geo-CAPE lies in its ability to observe the same scene multiple
times per day. The assessment of standing stocks of phytoplankton and carbon pools will allow
for better quantification and constraint of process rate estimates. The multiple views per day
along with ancillary observations will lend insight into the physical, chemical, and biological
formation and evolution of water masses.
Discussion is underway regarding the prioritization of satellite products and processes
and the selection of sites to investigate these. The Science Working Group has identified the
products that are still in research or exploratory status as warranting investment at this stage in
the planning process. Several of the exploratory products have been demonstrated in open ocean
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
case 1 waters (phytoplankton dominated) but similar retrieval success has yet to be proved in
optically complex case 2 waters (optical properties vary independently) and in dynamic coastal
regimes. Some of these products include (in no particular order), primary production and
respiration, net community production of DOC/POC, fluorescence dynamics over the day,
seawater pCO2, differentiation of phytoplankton functional types, and detection of surface
manifestations of phytoplankton vertical migration.
The Science Working Group has pointed out the value of utilizing field sites associated
with an ongoing time series to allow a greater context of the measurements and reducing the
costs associated with enabling activity field programs. The selection of field sites for the
enabling investigations will depend on the prioritized products and processes. The selection of
the field sites will maximize the number of products/process that can be addressed. The site
selection will be based upon extent of existing time series measurements, the dynamic range of
variability for the parameters of interest and geographical coverage of different regimes (i.e
coastal upwelling, run-off dominated, stratified, etc.). In addition, if feasible, the coincident
observations of atmospheric aerosol and trace gas properties will provide important information
given the interdisciplinary nature of the mission with atmospheric and oceanic instruments being
flown on the same platform.
Interdisciplinary Science Studies
Among the recommendations made at the August 2008 workshop, was a suggestion that an effort
be made to delineate potential interdisciplinary research that would benefit from observations
anticipated from Geo-CAPE. The emphasis on atmospheric boundary layer observations for key
trace gases such as O3 and NO2, along with the high spatial resolution of the ocean color sensor,
and the high resolution permitted by a geosynchronous orbit is expected to provide key
information that may be combined with in situ data and models to help us better understand the
complicated dynamics and biogeochemical cycles along our urbanized coast lines. There are
myriad and complex interconnections between the atmosphere and coastal waters, with crucial
terrestrial linkages as well. For example, both anthropogenic and natural sources of nutrients
have atmospheric vectors (particularly for N and Fe) that contribute to the supply of these
elements in marine ecosystems, but their bioavailability depends upon the form in which they are
deposited and upon the organisms in the water that might make use of them [add refs]. The
carbon cycle in coastal waters is particularly complex involving CO2, CO, VOCs, and aerosols,
with coastal waters functioning both as sources and sinks. Much work remains to be done to
understand the processes involved in this cycle [add refs]. There is also an increasing body of
evidence that suggests marine ecosystems may play an important role in urban air quality by
providing halogen radicals that contribute to the oxidative capacity of the boundary layer along
the coastal margins [add refs]. For a more thorough discussion of interdisciplinary science topics
and the role Geo-CAPE may play in helping us to better understand these coupled atmosphereocean biogeochemical processes, please see the companion white paper Geo-CAPE:
Interdisciplinary Science Potential [Jordan et al., 20xx].
References
Allan, D. W., Rice, J. P., and Goodman, J. A., Hyperspectral projection of a coral reef scene
using the NIST hyperspectral image projector, Proc. SPIE,7334, 733415, 2009.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Antoine, D., F. d’Ortenzio, S. B. Hooker, G. Be´cu, B. Gentili, D. Tailliez, and A. J. Scott
(2008), Assessment of uncertainty in the ocean reflectance determined by three satellite
ocean color sensors (MERIS, SeaWiFS and MODIS-A) at an offshore site in the
Mediterranean Sea (BOUSSOLE project), J. Geophys. Res., 113, C07013,
doi:07010.01029/02007JC004472.
Barnes, R.A., R. E. Eplee Jr., F. S. Patt, H. H. Kieffer, T. C. Stone, G. Meister, and J. J. Butler C.
R. McClain (2004) Comparison of SeaWiFS measurements of the Moon with the USGS
lunar model, Applied Optics, 43:5838–5854.
Bissett, W.P., Arnone, R., Davis, C.O., Dickey, T., Dye, D., Kohler, D.D.R. and Gould, R. 2004.
From meters to kilometers- a look at ocean color scales of variability, spatial coherence,
and the need for fine scale remote sensing in coastal ocean optics, Oceanography, 17, 3243.
Boss, E. S., and S. Maritorena (2006), Uncertianties in the products of ocean-colour remote
sensing, in Remote Sensing of Inherent Optical Properties: Fundamentals, Tests of
Algorithms and Applications, edited by Z.-P. Lee, International Ocean-Colour
Coordinating Group, Dartmouth, Canada.
Brown, S. W., Eppeldauer, G. P., and Lykke, K. R., Facility for spectral irradiance and radiance
responsivity calibrations using uniform sources, Appl. Opt.,45, 8219-8237, 2006.
Carder, K. L., F. R. Chen, Z. P. Lee, S. K. Hawes, and D. Kamykowski (1999), Semianalytic
Moderate-Resolution Imaging Spectrometer algorithms for chlorophyll-a and absorption
with bio-optical domains based on nitrate-depletion temperatures, J. Geophys. Res., 104,
5403-5421.
Chami, M., and M. Defoin-Platel (2007), How ambiguous is the inverse problem of ocean color
in coastal waters?, J. Geophys. Res., 112, C03004, doi:03010.01029/02006JC003847.
Clark, D.K., Gordon H.R, Voss K.J., Ge. Y., Broenkow W, et al. (1997) Validation of
atmospheric correction over the ocean. J. Geophys. Res., 102, 17209–17.
Darecki, M., and D. Stramski (2004), An evaluation of MODIS and SeaWiFS bio-optical
algorithms in the Baltic Sea, Remote Sens. Environ., 89, 326-350.
Davis, C.O., M. Kavanaugh, R. Letelier, W.P. Bissett and D. Kohler. 2007. Spatial and spectral
resolution considerations for imaging coastal waters. Coastal Ocean Remote Sensing,
edited by Robert J. Frouin, ZhongPing Lee, Proc. of SPIE, 6680, doi: 10.1117/12.734288.
Dickey, T.D. 1991. The emergence of concurrent high-resolution physical and bio-optical
measurements in the upper ocean and their applications. Review of Geophysics, 29, 383413.
Dinguirard, M., and P. N. Slater (1999). Calibration of space-multispectral imaging sensors: A
review. Remote Sens. Envion. 68:194-205.
Doney, S.C., N. Mahowald, I. Lima, R.A. Feely, F.T.M.J.-F. Lamarque and P.J. Rasch (2007).
"Impact of anthropogenic atmospheric nitrogen and sulfur deposition on ocean
acidification and the inorganic carbon system." Proceedings of the National Academy of
Sciences 104(37): 14580-14585.
Eplee, R.E. Jr., J.-Q. Sun, G. Meister, F.S. Patt, X. Xiong, and C.R. McClain (2010) Cross
calibration of SeaWiFS and MODIS using on-orbit observations of the Moon, Appl. Opt.,
in press.
Eplee, R.E. Jr., G. Meister, F.S. Patt, and C.R. McClain (submitted) The on-orbit calibration of
SeaWiFS, Appl. Opt., submitted.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Evans, R. H., & Gordon, H. R. (1994). Coastal Zone Color Scanner system calibration: a
retrospective examination. Journal of Geophysical Research, 99, 7293– 7307.
Franz, B.A., S.W. Bailey, P.J.Werdell, and C.R. McClain (2007) Sensor-independent approach
to the vicarious calibration of satellite ocean color radiometry, Appl. Opt. 46, 5068–5082.
Gordon, H. R. (1987). Calibration requirements and methodology for remote sensors viewing the
ocean in the visible. Remote Sens. Environ., 22:103-126.
Gordon, H. R., and M. Wang (1994), Retrieval of water-leaving radiance and aerosol optical
thickness over oceans with SeaWiFS: A preliminary algorithm, Applied Optics, 33, 443452.
Hand, J.L., N.M. Mahowald, Y. Chen, R.L. Siefert, C. Luo, A. Subramaniam and I. Fung (2004).
Estimates of atmospheric-processed soluble iron from observations and a global mineral
aerosol model: Biogeochemical implications, Journal of Geophysical Research
109(D17205): doi:10.1029/2004JD004574.
Hooker, S.B., McClain, C.R. and Mannino, A. (2007) NASA strategic planning document: a
comprehensive plan for the long-term calibration and validation of oceanic
biogeochemical satellite data, NASA Special Publication 2007-214152, NASA Goddard
Space Flight Center, Greenbelt, Maryland, 31 pp.
IOCCG (2006), Remote Sensing of Inherent Optical Properties: Fundamentals, Tests of
Algorithms, and Applications, in Reports of the International Ocean-Colour
Coordinating Group, No. 5, edited by Z.-P. Lee, p. 126, IOCCG, Dartmouth, Canada.
Johnson, B.C., E.A. Early, R.E. Eplee, Jr., R.A. Barnes, and R.T. Caffrey (1999). The 1997
Prelaunch Radiometric Calibration of SeaWiFS. NASA Tech. Memo. 1999-206892, Vol.
4, S.B. Hooker and E.R. Firestone, Eds., NASA Goddard Space Flight Center, Greenbelt,
Maryland, 51 pp.
Lee, Z., R. Arnone, C. Hu, P. J. Werdell, and B. Lubac (2010), Uncertainties of optical
parameters and their propagations in an analytical ocean color inversion algorithm,
Applied Optics, 49(3), 369-381.
Maritorena, S., O. H. F. d'Andon, A. Mangin, and D. A. Siegel (2010), Merged satellite ocean
color data products using a bio-optical model: Characteristics, benefits and issues,
Remote Sensing of Environment, 114 1791-1804.
McClain, C.R. (2009) A Decade of Satellite Ocean Color Observations, Annu. Rev. Mar. Sci, 1,
19–42
McClain, C.R., E.J. Ainsworth, R.A. Barnes, R.E. Eplee, Jr., F.S. Patt, W.D. Robinson, M.
Wang, and S.W. Bailey, 2000: SeaWiFS Postlaunch Calibration and Validation Analyses,
NASA Tech. Memo. 2000-206892, Vol. 9-11, S.B. Hooker and E.R. Firestone, Eds.,
NASA Goddard Space Flight Center.
McClain, C. R., R. A. Barnes, J. R.E. Eplee, B. A. Franz, N. C. Hsu, F. S. Patt, C. M. Pietras, W.
D. Robinson, B. D. Schieber, G. M. Schmidt, M. Wang, S. W. Bailey, and P. J. Werdell
(Eds.) (2000), SeaWiFS Postlaunch Calibration and Validation Analyses, Part 2., 57 pp.,
NASA Goddard Space Flight Center, Greenbelt, Maryland.
Moore, T. S., J. W. Campbell, and M. D. Dowell (2009), A class-based approach to
characterizing and mapping the uncertainty of the MODIS ocean chlorophyll product,
Remote Sensing of Environment, 113 2424-2430.
NASA, 2006. Earth’s Living Ocean: The Unseen World. An advanced plan for NASA’s Ocean
Biology and Biogeochemistry Research, DRAFT, 2006.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
NRC, 2007. Earth Science and Applications from Space: National Imperatives for the Next
Decade and Beyond. The National Academies Press, Washington, D.C.
O'Reilly, J., S. Maritorena, B. G. Mitchell, D. Siegel, K. L. Carder, S. Garver, M. Kahru, and C.
McClain (1998), Ocean color chlorophyll algorithms for SeaWiFS, J. Geophys. Res., 103,
24937-24953.
Salama, M. S., and A. Stein (2009), Error decomposition and estimation of inherent optical
properties, Applied Optics, 48(26), 4947-4962.
Stone, T.C., H.H. Kieffer and I.F. Grant (2005) Potential for calibration of geostationary
meteorological satellite imagers using the Moon, Proceedings of SPIE, 5882, doi:
10.1117/12.620097.
Wang, M., and H. R. Gordon (2002). Calibration of ocean color scanners: how much error is
acceptable in the near infrared? Remote Sens. Environ. 82:497-504.
Wolfe, R. E., M. Nishihama, A. J. Fleig, J. A. Kuyper, D. P. Roy, J. C. Storey, and F. S. Patt
(2002). Achieving sub-pixel geolocation accuracy in support of MODIS land science.
Remote Sens. Environ. 83:31-49.
APPENDIX 1
Oil Spill Monitoring from GEO-CAPE
Every year, huge quantities of oil and petroleum products enter the sea, land, and
groundwater (NAS, 2003). Monitoring of oil spill at sea is critical in assessing of spill’s
characteristics, fate, and environment impact. Satellite instruments for spill monitoring include
optical, microwave, and radar (e.g., synthetic aperture radar, SAR) sensors, each having its own
advantages and disadvantages (Fingas and Brown, 1997 and 2000; Brekke and Solberg, 2005).
Although SAR is perhaps the most often used, it suffers from high cost, lack of coverage, and
difficulty to differentiate oil from other suspicious features (Alpers and Espedal, 2004). Most
importantly, the only SAR signal is the dampened surface backscattering due to modulation of
the oil slick/film to surface wave, which is difficult to use for thickness estimates. Optical
instruments provide alternative means that can potentially overcome these difficulties. In
particular, ocean color instruments on geostationary platforms may present unprecedented
opportunities to monitor oil spill and other oil pollution events. The DeepWater Horizon (DWH)
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
oil spill event in the Gulf of Mexico in spring and summer 2010 presents an example of why a
geo-stationary, well-designed ocean color sensor is required.
The use of optical remote sensing to detect oil spill started decades ago (e.g., McDonald
et al., 1993). Hu et al. (2003) first demonstrated the advantage of using the quasi-operational
satellite instrument MODIS for spill monitoring in a turbid lake, where 2 images per week were
obtained after cloud screening, with no cost. The detection was possible because the high water
turbidity provided a “bright” background where the highly light-absorbing oil films could be
visualized. In the oligotrophic ocean, the water background is also dark, making oil detection
difficult. However, Hu et al. (2009) showed that when MODIS imagery contained sun glint (i.e.,
specular reflection of the solar beam), high contrast was found between oil slicks and the
background water. The contrast is not due to the difference between optical properties of the oil
film and the water (as evidenced by the lack of contrast in other glint-free images), but due to the
oil-modulation of the surface capillary waves – the same principle for SAR measurements (Chust
and Sagarminaga, 2007). However, quantification of the relationship between sun glint
magnitude, reflectance contrast, oil film thickness, and oil type still remains problematic, mainly
due to lack of continuous observations and in situ groundtruthing measurements.
Fig. 6 shows several MODIS examples of the DWH oil spill, which started after the
tragedy explosion of the DWH oil rig (28.74oN, 88.37oW) on 21 April 2010. Figs. 6a and 6b
show the suspicious slick feature among the scattered clouds. The feature shows spectral shape
that is different from that of nearby clouds, with relatively lower reflectance in the blue
wavelengths due to enhanced Rayleigh scattering along the sun glint beam. MODIS images from
subsequent days show the spatial evolution of the slick, confirming that this is surface oil. Figs.
6c and 6d show two other examples from the same day from MODIS/Terra and MODIS/Aqua
observations, respectively. Due to changes in the solar/viewing geometry, the same oil slicks
show positive contrast in the morning but negative contrast in the afternoon. Further, the spatial
contrast in some of the slicks in Fig. 1d diminished due to combined effect of both optical
properties and surface Fresnel reflection of the oil film.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
The quasi-operational MODIS (and supplemental MERIS) 250-m observations provided
the first images of oil spill at the beginning of the event, when SAR coverage was limited. As the
spill developed from late April to July, coordinated efforts among the international remote
sensing communities led to enhanced SAR coverage by using multiple SAR instruments, tilted to
observe the spill whenever possible. However, the thickness of the oil film is impossible to
obtain from SAR observations. In contrast, there are some preliminary evidence from laboratory
measurements that oil film thickness may be obtained from surface reflectance measurements in
the visible and shortwave-infrared (Svejkovsky and Muskat, 2006; Król et al., 2006; JPL
unpublished data). The limited observations have not considered the influence of the surface bidirectional Fresnel reflectance that strongly modulates the reflectance signal, not to mention of
the optical properties related to oil types (e.g., new or emulsified). As a result, one of the
controversies in estimating the daily spill volume of the DWH event comes from the unknown
thickness of the oil slick.
In summary, several applications of MODIS (and other) multi-spectral imagery in spill
monitoring have shown the following advantages:
1. Wide and daily coverage (> 2000 km swath)
2. Medium resolution (250 – 300 m)
3. Multi-spectral from 412 to 2130 nm
However, existing ocean color instruments also suffer from several weaknesses, including
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Cloud contamination
Lack of ability to detect small-scale (10s of meters) slicks
Lack of continuous observation (the multi-sensor approach suffers from inter-calibration
problems)
4. Unknown relationship between optical contrast (from visible to shortwave-IR) and oil
thickness
5. Unknown relationship between spectral shapes and oil thickness/type
6. Unknown ability to differentiate oil and other features, for example Sargassum slicks,
Trichodesmium patches, phytoplankton or fish induced surface surfactant, shoals, and
coastal freshwater jets.
1.
2.
3.
GEO-CAPE will provide continuous observations during the same day, which can
improve spatial coverage (assuming non-persistent clouds) and temporal coverage (#1 and #3).
More importantly, the multi-observations from the same instrument at hyperspectral wavelengths
will provide potential capabilities to derive information on oil thickness and type (#4 and #5) as
well as to differentiate oil from other features (#6). Combined with the cloudfree (#1) and
higher-resolution (#2) SAR observations as well as targeted groundtruthing measurements,
GEO-CAPE may provide completely novel information on oil slicks and therefore significantly
enhance our capability in spill monitoring. The research community, on the other hand, needs to
put more effort in studying the optical properties of oil, both in the laboratory and in the field.
References and further readings
Alpers, W., and H. A. Espedal (2004). Oils and surfactants. In: C. R. Jackson and J. R. Apel (eds)
Synthetic Aperture Radar Marine User’s Manual. U.S. Department of Commerce, Washington,
DC, September 2004. pp263-275.
Brekke, C, and A. H. S. Solberg (2005). Oil spill detection by satellite remote sensing. Remote Sens.
Environ. 95:1-13.
Chust, G., and Y. Sagarminaga (2007). The multi-angle view of MISR detects oil slicks under sun
glitter conditions. Remote Sens. Environ. 107:232-239.
Fingas, M., and C. Brown (1997). Remote sensing of oil spills. Sea Technology, 38:37-46.
Fingas, M., and C. Brown (2000), Oil-spill remote sensing – An update. Sea Technology, 41:21-26.
Hu, C., et al. (2003). MODIS detects oil spills in Lake Maracaibo, Venezuela. EOS, Transactions,
AGU, 84(33):313,319.
Hu, C., X. Li, W. G. Pichel, and F. E. Muller-Karger (2009). Detection of natural oil slicks in the
NW Gulf of Mexico using MODIS imagery. Geophys. Res. Lett. Vol. 36, L01604,
doi:10.1029/2008GL036119.
Król, T., A. Stelmaszewski, and W. Freda (2006). Variability in the optical properties of a crude oil
– seawater emulsion. Oceanologia, 48:203-211.
MacDonald, I. R., et al. (1993). Natural oil slicks in the Gulf of Mexico visible from space. J.
Geophys. Res. 98(C9);16,351-16364.
NAS. 2003. Oil in the sea III. Committee on Oil in the Sea: Inputs, Fates, and Effects, Ocean Studies
Board and Marine Board, Divisions of Earth and Life Studies and Transportation Research
Board, National Research Council. The National Academies Press.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010
DRAFT
November 17, 2010 – V1.6
Svejkovsky, J., and J. Muskat (2006). Real-time Detection of Oil Slick Thickness Patterns with a
Portable Multispectral Sensor. Final Report submitted to the U.S. Department of the Interior
Minerals Management Service, Herndon, VA, July 31, 2006, Contract No. 0105CT39144.
1
GEO Oceans white paper draft V1_6.docx – Nov. 17, 2010