MODIS Cryospheric Products at the NSIDC DAAC
--Greg Scharfen (scharfen@kryos.colorado.edu), NSIDC DAAC, University of Colorado, Boulder, CO
The National Snow and Ice Data Center (NSIDC) Distributed Active Archive Center (DAAC) provides data and information on snow and ice processes, especially interactions among snow, ice, atmosphere, and ocean, in support of research on global change detection and model validation, and provides general data and information services to the cryospheric and polar processes research community. The NSIDC DAAC is an integral part of the multi-agency funded support for snow and ice data management services at NSIDC.
This report gives a brief overview of the planned cryospheric products from the Moderate Resolution Imaging Spectroradiometer (MODIS), how they relate to other products, and how they fit into EOS and NSIDC. These products are being developed by the EOS MODIS Science Team, and will be implemented at the NSIDC DAAC.
Currently, NSIDC produces cryospheric products from Defense Meteorological Satellite Program (DMSP) Special Sensor Microwave/Imager (SSM/I) data including polar brightness temperature grids, sea ice concentration, and a combined land snow extent (from National Oceanic and Atmospheric Administration/ National Environmental Satellite Data and Information Service [NOAA/NESDIS] weekly charts) and polar sea ice from SSM/I. Non-satellite data, such as meteorological fields, station data, and buoy measurements, are archived for comparison to satellite information and for input to sea ice and climate models. The NSIDC DAAC has supported the development of products to monitor ice surface temperature and ice motion by providing access to bi-polar subsets of Advanced Very High Resolution Radiometer (AVHRR) and TIROS Operational Vertical Sounder (TOVS) satellite data since 1992/93. Satellite altimetry data are being archived and distributed to support ice-sheet topography studies.
Table 1. V0 data sets currently held or under development at NSIDC
MODIS will be launched as part of the science payload on the first EOS platform (AM-1) in 1998. MODIS represents a technological improvement over the AVHRR sensors, which are the mainstay of the NOAA Polar Orbiting satellite program. MODIS will have a viewing swath width of 2300 km and will collect data in 36 spectral bands from 0.4-14 um, with a spatial resolution ranging from 250 to 1000 m. The AM-1 satellite will be in a sun-synchronous, near-polar orbit (at about 705 km altitude) with a descending node at about 10:30 a.m. local time. Follow-on satellites with MODIS-type instruments will be in similar orbits with either the same approximate schedule or a 1:30 p.m. local time ascending node. This configuration results in global coverage every one to two days by each satellite, although the polar regions will be covered more frequently because of the overlap between satellite passes at the poles.
MODIS data will be calibrated during normal inflight operations. Calibration will include radiometric checks and spectral band registration checks.
Snow cover is an important variable for global climate monitoring and change detection (Barry et al. 1995). Owing to its high albedo and large spatial variability, snow cover is a primary factor controlling the amount of solar energy absorbed at the surface. Changes in the extent of snow cover have a direct effect on the radiation budget. Snow cover also has significant effects on the seasonal temperature cycle due to its high latent heat of fusion. In many areas of the world, snow cover represents an important resource in terms of water supply and hydroelectric power. Links between snow cover extent and atmospheric circulation have been demonstrated (Walsh et al. 1985), and the extent of snow cover has been found to be inversely related to hemispheric surface air temperature (Robinson and Dewey 1990).
The MODIS snow cover products are being developed by the MODIS Science Team (Hall et al. 1995). This work follows on a history of monitoring large-scale Northern Hemisphere snow cover extent from visible and infrared satellite sensors, especially those aboard the NOAA polar orbiting satellites (Matson et al. 1986). Northern Hemisphere snow cover has been mapped by NOAA/NESDIS analysts from the Very High Resolution Radiometer (VHRR) and AVHRR sensors on these satellites since 1966. The record of data has been scrutinized and found to be most reliable from 1972 onwards (Robinson et al. 1993). Analyses of the record by Robinson and others show mean annual snow-covered area for the Northern Hemisphere to be 25.4 million square km. While this record is short, and includes much interannual and regional variability, a decrease in total snow-covered area after about the mid-1980s is readily apparent (Figure 1, Robinson, pers. comm., unpublished 1996). It is not known whether this is a short-term anomaly or part of a longer-term trend.
In addition to the spatial-coverage similarities of MODIS to AVHRR, it has some similar spectral characteristics to Landsat TM data. Much of the development work for MODIS algorithms has utilized the similar spectral bands of Landsat TM data. The MODIS Snow Cover Mapping algorithm (SNOMAP) is designed to utilize the reflectance characteristics in the visible and near-infrared regions of the electromagnetic spectrum. Wavelength-center locations and spatial resolution of MODIS bands that have corresponding TM coverage are given in Table 2 (Riggs et al. 1994). By utilizing the narrow spectral bands of these sensors, the SNOMAP algorithm is designed to identify snow cover routinely and to discriminate between snow and other features.
SNOMAP has two criteria tests that were developed with TM data and will be
applied to the MODIS data (Riggs et al. 1994). Digital numbers acquired
by the sensor are converted to reflectances using solar zenith angle
corrections. A key characteristic of snow is that its reflectivity is high in
the visible part of the spectrum and low in the near-infrared at about 1.6 um
(O'Brien and Munis 1975). The reflectivity of clouds remains high in both the
visible and near-infrared regions. These features are accounted for in the
calculation of the Normalized Difference Snow Index (NDSI), expressed for TM
data as: NDSI = (TM 2 [0.56 um] - TM 5 [1.65 um])/(TM 2 + TM 5) after Dozier
(1989). A NDSI value of 0.4 is used as the threshold for snow. This technique
can be used to classify pixels as snow vs. other bright features such as
clouds. A second threshold test is applied to the TM band 4 data to distinguish
between snow and water. A land-water mask and a cloud mask derived from MODIS
data are being integrated into SNOMAP. Use of the cloud mask is expected to
improve the ability to discriminate between thin clouds and snow. Errors with
this technique are minimal and are usually due to pixels containing cirrus
clouds or bright surface features misidentified as snow. The snow cover
algorithm will be updated with additional tests using data from the MODIS
Airborne Simulator (MAS) that has many of the same spectral characteristics as
the MODIS sensor. In addition, the algorithm may be modified after launch of
the AM-1 platform to incorporate other tests and possibly other MODIS bands
(Riggs et al. 1994).
Sea ice is an important variable affecting the energy balance of the polar
regions. Sea ice inhibits the exchange of heat and moisture between the
atmosphere and ocean, and dramatically increases the albedo of the polar
oceans. In the southern ocean the annual cycle of sea growth and decay is very
large, ranging from a maximum extent of 20 X 106 square kilometers
in September to its minimum area of approximately 4 X 106 square
kilometers in February (Zwally et al. 1983).
During freezeup, large quantities of salt are ejected, altering the density
structure of the water column.
Sea ice has been monitored by satellites using visible and infrared sensors
since the early 1970s. Data from the AVHRR sensor and its predecessors, along
with ship and aircraft observations, have been the basis of operational ice
mapping efforts at the National Ice Center in Suitland, Maryland, and the
Atmospheric Environment Service Ice Branch in Ottawa, Ontario, since the 1970s.
These data provide the basis for daily, weekly, and monthly charts showing ice
margins, large fractures and leads, and categories of ice concentration. They
are used extensively by ships operating in ice-infested waters. Limitations
include persistent cloud cover and darkness during much of the year.
In 1972, passive microwave satellite data became available for mapping sea ice
extent and concentration. Data from the ESMR, SMMR, and SSM/I passive microwave
satellite sensors have a coarser resolution than the visible and infrared data,
but have the advantages of being useful year round and with only minimal
interference by atmospheric moisture, including cloud cover. These data are
available on CD-ROM from NSIDC.
These two approaches for monitoring sea ice with remote sensing data are
complementary. The traditional visible and infrared analyses offer greater
detail, while the microwave data offer dependability under varying
environmental conditions. MODIS data for sea ice analyses will offer an
improvement over the traditional analyses in terms of spatial and spectral
resolution, but they will still be limited by cloud cover.
The MODIS ice algorithm uses the same tests as the SNOMAP algorithm (Riggs
et al. 1994). Significant reflectance differences between open water and
most ice types allow for detection of ice using the same algorithm. Tests with
TM data have shown the algorithm to be acceptable, except for detecting thin
ice. Investigation of additional tests to improve the analyses is continuing.
The sea ice product being developed by the MODIS Science Team is sea ice extent
on a daily basis, and composited to a weekly or ten-day maximum extent.
Additional products, including sea ice albedo, ice surface temperature, and ice
motion may be developed after the launch of the AM-1 satellite.
In September 1995, a workshop was held at NASA's Goddard Space Flight Center to
discuss the MODIS snow and ice products (Hall 1995). Invited participants gave
presentations on current snow and ice mapping systems using remote sensing
data. Participants were asked to evaluate the products and make recommendations
to improve their utility.
For the MODIS snow products, it was recommended that the resolution be improved
from 1000 m to 500 m. For the lake ice product, it was also recommended that
the resolution be improved to 500 m. For both the snow and ice products it was
felt that the compositing period should be specified by the user.
The participants recognized that there was a need for a sea ice product based
on optical wavelengths, but felt that it would be more useful if it included
other ice information, such as concentration, ice type, ice surface temperature
and albedo, if possible. The participants also recognized the utility of
current passive microwave products and the planned active microwave products
from RADARSAT.
The utility of gridding and projection schemes was also discussed. It was
recommended that the products be available in a polar grid as well as the
standard EOS grid (an adaptation of the International Satellite Cloud
Climatology Project grid).
Operational users expressed interest in receiving the data as soon as possible
(within 48 hours) for incorporation into near-real-time forecasts.
To the degree possible, these changes are being included in revisions to the
MODIS snow and ice product algorithms (D. Hall, pers. comm.). The Science Team
is continuing to investigate ways to provide other sea ice parameters.
Candidate algorithms are being investigated, and new techniques such as
spectral mixture modeling may be incorporated in the post-launch time- frame.
The EOS MODIS Science Team is developing snow and ice products to be
implemented at the time of the launch of the first EOS mission on the AM-1
platform in 1998. MODIS products will be distributed via EOSDIS from the GSFC,
EDC, and NSIDC DAACs. Present plans call for NSIDC to archive and distribute
the MODIS snow and ice products. These products will utilize the improved
spatial and spectral resolution of MODIS over current visible and infrared
sensors and offer an improvement over currently available similar products.
They will complement the related microwave and radar data to generate "all
weather" snow and ice products. Daily maps of global snow cover and sea ice and
lake ice extent will be produced and archived. In addition, the MODIS data
offer the potential for the development of other cryospheric products. More
information on these products may be found at the MODIS Home Page
(http://ltpwww.gsfc.nasa.gov/MODIS/MODIS.html), or the Algorithm Theoretical
Basis Document Home Page (http://spso.gsfc.nasa.gov/atbd/pg1.html).
Barry, R. G., J. M. Fallot, and R. L. Armstrong, 1995: Twentieth-century
variability in snow cover conditions and approaches to detecting and monitoring
changes: status and prospects. Progress in Phys. Geog. 19, 4,
520-532.
Dozier, J., 1989: Spectral signature of alpine snow cover from the Landsat
Thematic Mapper, Remote Sens. Environ., 28, 9-22.
Hall, D. K. (Ed.), 1995: First Moderate Resolution Imaging Spectroradiometer
(MODIS) Snow and Ice Workshop, NASA Conference Publication 3318,
GSFC, Greenbelt, MD, 133 pp.
Hall, D. K., G.A. Riggs, and V. V. Salomonson, 1995: Development of methods for
mapping global snow cover using Moderate Resolution Imaging Spectroradiometer
data, Remote Sens. Environ., 54, 127-140.
Matson, M., C. F. Ropelewski, and M. S. Varnadore, 1986: An Atlas of
Satellite-Derived Northern Hemisphere Snow Cover Frequency, National Weather
Service, Washington D.C., 75 pp.
O'Brien, H. W. and R. H. Munis, 1975: Red and Near-Infrared Spectral
Reflectance of Snow, Research Report 332, U.S. Army Cold Regions
Research and Engineering Laboratory, Hanover, NH, 18 pp.
Riggs, G. A., D. K. Hall, and V. V. Salomonson, 1994: A snow index for the
Landsat Thematic Mapper and Moderate Resolution Imaging Spectroradiometer,
IGARSS, 4, 1942-1944.
Robinson, D. A. and K. F. Dewey, 1990: Recent secular variations in the extent
of Northern Hemisphere snow cover, Geophys. Res. Lett.,
17, 1557-1560.
Robinson, D. A., K. F. Dewey, and R. R. Heim, Jr., 1993: Global snow cover
monitoring: an update, Bull. Amer. Meteor. Soc., 74, 689-696.
Walsh, J. E., W. H. Jasperson, and B. Ross, 1985: Influences of snow cover and
soil moisture on monthly air temperature, Mon. Wea. Rev., 113,
756-768.
Zwally, H. J., J. C. Comiso, C. L. Parkinson, W. J. Campbell, F. D. Carsey, and
P. Gloersen, 1983: Antarctic Sea Ice, 1973-1976: Satellite Passive Microwave
Observations, NASA SP-459, Washington, D.C., 206 pp.
Table 2. Corresponding MODIS an TM wavelengths (bands).
Band Spatial Resolution Center
Wavelength (um)Corresponding
TM Band1 250 0.645 3 2 250 0.858 4 3 500 0.469 1 4 500 0.555 2 6 500 1.640 5 7 500 2.130 7 13 1000 0.667 3 14 1000 0.678 3 16 1000 0.869 4 31 1000 11.030 6 Development of Methods for Mapping Sea Ice from MODIS
The MODIS Snow and Ice Workshop
Concluding Remarks
References
