Abstract:
Ocean chlorophyll-a and sea surface temperature (SST) observations are critical to enable assessments of variability, long-term trends, and change in any ecosystem. SST, chlorophyll-a, and Net Primary Production were derived for nearly 13 years of medium resolution images from MODIS (2000-2012) and from SeaWiFS (1997-2010), and about 19 years of AVHRR observations (1993-2012) of improved 1-km “standard” ocean color and SST products. Additional complementary and nearly concurrent data presently continue to be collected at USF from the MODIS and AVHRR satellite sensors. SeaWiFS ceased operations in December 2010. Institute of Marine Remote Sensing (IMaRS) data link: http://imars.marine.usf.edu/.
Suggested Citation:
Muller-Karger, Frank; Roffer, Mitchell; Bostater, Charles. 2014. MODIS Moderate Resolution Imaging Spectroradiometer on AQUA satellite (VIS 1km bands 8-16, IR 1km bands 20-23,31-32) and SeaWiFS Sea-viewing Wide Field-of-view Sensor, Gulf of Mexico, September-1-1993 to December-31-2013. Distributed by: GRIIDC, Harte Research Institute, Texas A&M University–Corpus Christi. doi:10.7266/N73F4MJZ
Data Parameters and Units:
The University of South Florida focused on developing a series of analyses based on historical high- resolution Sea Surface Temperature and Ocean Color datasets collected by the USF Institute for Marine Remote Sensing. This includes data from the NOAA AVHRR polar orbiting series (e.g., NOAA 11-19), NASA Terra and Aqua MODIS, and data from the SeaStar/Orbview-2/SeaWiFS sensors. We focused on comparing these datasets with various other remote sensing ata, including AVIRIS. The results include a comprehensive analysis of ocean color patterns which have been previously (mis-)interpreted in the literature by other researchers as phytoplankton blooming as a result of the Deepwater Horizon oil spill. We also have continued to support an extensive analysis of larval fish surveys conducted by NOAA in the region (SeaMap) in the context of historical satellite imagery. All images can be found at http://imars.marine.usf.edu/ and http://www.roffs.com/research-environmental/deepwater-horizon-rig-oil-spill-monitoring/
Methods:
Ocean color observations can be used to estimate the spatial distribution of stocks and rate of change of various organic and inorganic materials in the ocean, including those associated with living and detrital particulate materials and with colored dissolved organic matter (CDOM). A useful index of bulk living matter, i.e. phytoplankton, in the deep ocean outside of coastal areas is chlorophyll-a concentration. The standard satellite-based estimates of chlorophyll-a concentration are based on the light-absorption properties of the chlorophyll pigments found in phytoplankton. Chlorophyll-a absorbs strongly in the blue and red parts of the visible spectrum, and tends to reflect light in the green (therefore the pigment is green). Therefore, the most widely-used remote sensing algorithms to estimate chlorophyll-a concentration use a ratio of the amount of blue light reflected relative to that of green light reflected by natural waters. These estimates are strongly affected by detrital matter and CDOM because these substances also absorb blue light. If these substances are present in any large quantity, in a manner that is not correlated with chlorophyll concentration and if they are not properly accounted for, the standard chlorophyll-a algorithms will overestimate phytoplankton biomass. This is common in coastal zones and areas affected by river plumes. Yet, these data are particularly useful for coastal applications to examine variations in patterns over space and time to make inferences about river discharge and resuspension of sediments from the bottom, and in general for assessing coastal water quality. The satellite data products most widely used are the chlorophyll-a concentration maps, but these can be complemented by estimates of light reflectance in various colors, or estimates of absorption and backscattering properties of the ocean. Our objective for this project was to generate a set of specific products within the initial focus on the Gulf of Mexico. Generating the products required continuous assessments of the calibration of the satellite data, a strategy for accurate atmospheric correction, and mature algorithms for the geophysical products of interest. The spatial and temporal variability observed in these parameters is an indicator of ecosystem change, and some of these factors may themselves induce ecosystem change. These products complement other baseline environmental, navigation, and socio-economic information required in preparation for and in response to emergency situations. Most of the work done on ocean color remote sensing applications has focused on waters away from the coast, in conditions referred to as Case I waters. Most of the Gulf of Mexico waters are Case I waters, including most shelf waters. Case II waters are most likely found near the coast, often in waters shallower than 20 m, in river plume areas, and within estuarine zones. MODIS and SeaWiFS Ocean Color Products Chlorophyll concentration fields were derived from images collected by the Sea-viewing Wide-Field-of-view Sensor (SeaWiFS) and the Moderate Resolution Imaging Spectrometer (MODIS) flown on NASA’s Aqua satellite. SeaWiFS data were collected using a High Resolution Picture Transmission(HRPT) antenna and MODIS data were obtained using an L-band antenna located at the University of South Florida, in St. Petersburg, FL. Data were also obtained from the NASA Goddard Space Flight Center Distribute Active Archive Center (GSFC DAAC). We used the "standard" atmospheric correction applied by NASA for ocean color products. Chlorophyll-a concentration was estimated using the standard NASA band ratio algorithms for SeaWiFS and MODIS (i.e. OC4 and OC3, respectively). Net Primary Production (NPP). The NPP product is based on the Vertically Generalized Production Model (VGPM). MODIS surface chlorophyll concentrations (Chl_a), MODIS sea surface temperature data (SST), and MODIS cloud-corrected incident daily photosynthetically active radiation (PAR) at 1km resolution were used as input data. The daily Chl_a, SST, and PAR were calculated first and then weekly and monthly means for each parameter were calculated from the daily products. These weekly and monthly Chl_a, SST, and PAR fields were then used to calculate the weekly and monthly NPP. Units for Chl_a, SST, and PAR are milligrams per cubic meter, degrees C, and Einsteins per day per square meter, respectively. The units for estimated NPP are milligrams Carbon per square meter per day. AVHRR and MODIS Sea Surface Temperature (SST) products Sea Surface Temperature (SST) was derived from infrared (IR) observations collected by the Advanced Very High Resolution Radiometer (AVHRR) sensors flown on the National Oceanic and Atmospheric Administration's (NOAA) Polar Orbiting Environmental Satellite (POES) series, and the Moderate Resolution Imaging Spectrometer (MODIS) flown on NASA’s Aqua satellite. AVHRR data were collected using a High Resolution Picture Transmission (HRPT) antenna, and MODIS data were obtained using an L-band antenna located at the University of South Florida, in St. Petersburg, FL. MODIS data were also obtained from the NASA Goddard Space Flight Center Distribute Active Archive Center (GSFC DAAC). Due to lack of onboard GPS, AVHRR imagery has frequent and serious navigation errors. Residual errors due to cloud contamination in the process to estimate SST also create a serious problem for time-series studies and anomaly detection (note that this error is most severe in MODIS night-time imagery processed by SeaDAS). Recently we completed reprocessing the entire IMaRS archive of AVHRR data accumulated since late 1993. Each of the >50,000 images was manually navigated to minimize error, and a temporal filter based on in situ and remote sensing data and statistical anomaly analyses, was developed to remove the cloud errors. The AVHRR SST products generated for this project include all images collected daily, both daytime and night time. All passes from all satellites were combined to build the time series. SST was computed using the multi-channel sea-surface temperature (MCSST) algorithm. The approximate root mean square (rms) error of the AVHRR SST retrievals, confirmed through comparisons with in situ data, is of the order of 0.5 C. SST fields were mapped to a Cylindrical Equidistant projection covering the Gulf of Mexico at a spatial resolution of about 1 Km2 per pixel. The MODIS products include only night time products derived with the 4 micron channels (NASA SST4 product). The SST4 algorithm is described at the NASA OBPG web site: http://oceancolor.gsfc.nasa.gov/DOCS/modis_sst/.
View Metadata