The project consists of two parts: the generation of a data set of sea ice extents and areas, and associated scientific analyses. The objective of the first part is to produce a 30-year, research quality sea ice data set for climate change studies. The data set will build on an existing 18-year data set derived from satellite passive-microwave observations and currently archived at the National Snow and Ice Data Center in Boulder, CO. We will extend this data set by using historical data from the 1970's from the National Ice Center and new data from DMSP Special Sensor Microwave Imagers and the upcoming EOS-PM Advanced Microwave Scanning Radiometer. These data sets will be cross-calibrated to ensure a consistent 30-year data set following methods developed earlier and based on matching the geophysical parameters during periods of sensor overlap. The principal products will be Arctic and Antarctic sea ice extents and areas, derived from sea ice concentration maps. The second part of the proposal will center on the analysis and use of the 30-year data set. The science objectives are (1) to define and explain the hemispheric, regional, seasonal, and interannual variabilities and trends of the Arctic and Antarctic sea ice covers and (2) to understand any observed hemispheric asymmetries in global sea ice changes. Hemispheric sea ice cover asymmetries have been found in the existing 18-year record and have also been suggested from some model experiments simulating future conditions assuming a gradual increase in atmospheric CO2. We will examine the proposed 30-year record to determine the degree and nature of the hemispheric asymmetry in it and to place the sea ice observations in the context of other climate variables through comparisons with simulations from the NOAA Geophysical Fluid Dynamics Laboratory and Hadley Centre climate models.
Media studied: Arctic seaice
Procedures: A description of the procedure to be followed in cross-calibrating the input data sets has recently been published (Cavalieri et al., 1999). The procedure centers on matching the geophysical parameters rather than trying to match sensor radiances, a philosophy earlier advocated by Zabel and Jezek (1994). Matching of the time series was achieved by adjusting algorithm constants relative to the SMMR data (Cavalieri et al., 1999).
Both sea ice extent and area are derived from daily sea ice concentration maps. Thus the primary products to be quality controlled are the ice concentrations. Steps followed to ensure quality sea ice extents and areas are: - Remove bad brightness temperature data (e.g., bad orbits, scans, etc.). - Fill in isolated missing-data pixels on the brightness temperature maps by spatial interpolation. - Generate and quality control the sea ice concentration maps, removing spurious sea ice (from weather effects and land contaimination of pixels near the coast) over known ice-free ocean areas. - Temporally interpolate for missing ice concentration areas and days. - Generate time series of sea ice extents and areas and examine them for spikes or gaps, then correct. This procedure has been described in detail by Cavalieri et al. (1999).