BOOK SERIES: Antarctic Research Series Vol.70
DATE: 1996
SECTION: HABITAT REGION, pp.81-104
TITLE: SPATIAL AND TEMPORAL VARIABILITY IN WESTERN ANTARCTIC PENINSULA SEA ICE COVERAGE
AUTHORS: S.E.Stammerjohn and R.C.Smith
ABSTRACT
Spatial and temporal variability of sea ice coverage west of the Antarctic Peninsula, the Long Term Ecological Research (LTER) study area, is analyzed from October 1978 to August 1994 using surface sea ice concentrations derived from passive microwave satellite data. Ice coverage in the LTER region, nearby regions and the Southern Ocean are compared. Results show that various regions have distinct characteristics in seasonal and interannual variability in contrast to the Southern Ocean as a whole. For example, \fIseasonal\fP ice coverage in the LTER and Bellingshausen regions is distinct from other Southern Ocean regions in that the period of ice advance is relatively short in comparison to the period of ice retreat. In addition, the LTER and Bellingshausen regions are the only Southern Ocean regions which show long term persistence in monthly anomalous ice coverage, so that there is an oscillation between several consecutive high ice years followed by several consecutive low ice years. Cross spectral analysis of monthly anomalous ice coverage was performed to determine the longitudinal spatial extent of coherence among regions. There is no coherence between the eastern Antarctic Peninsula and LTER regions. There is low frequency coherence between Bellingshausen and Amundsen monthly anomalies, and the phase indicates that anomalous ice coverage in the Bellingshausen lags the Amundsen by approximately one year. Within the Bellingshausen region there is both high and low frequency coherence and a west-to-east propagation of anomalies. Comparisons of ice concentrations calculated from passive microwave and visible satellite data show that passive microwave derived ice concentrations overestimate surface ice concentrations between 0-40% (17.4% \(+-17.8%) and 40-75% (10.8% \(+-11.3%) but underestimate surface ice concentrations between 75-100% (9.2% \(+-6.9%). Ecological implications, as revealed by both the low and high resolution satellite data, are discussed. This LTER sea ice record provides a basis against which life-history parameters of primary producers and populations of key species from different trophic levels can be monitored and against which oceanographic and atmospheric variability in this region can be compared and modeled. 1. INTRODUCTION A dominant and distinguishing characteristic of Southern Ocean marine ecology is sea ice, which ranges from a minimum extent of 4 x 10\s-3\u6\d\s0 km\s-3\u2\d\s0 in February to a maximum extent of about 20 x 10\s-3\u6\d\s0 km\s-3\u2\d\s0 in September [. gordon taylor 1975, zwally comiso parkinson campbell carsey gloersen 1983 .], defining a seasonal sea ice zone which is more than the area of the Antarctic continent itself (13.2 x 10\s-3\u6\d\s0 km\s-3\u2\d\s0). Sea ice and the overlying snow cover influence ocean-atmosphere exchanges by modifying albedo, heat exhanges, momentum transfers and oceanic salt flux. The presence of an ice/snow cover changes the input of solar radiation to the water column, affecting marine biological processes and the heat balance of the upper ocean. The freezing and melting of sea ice contribute to the variability in salinity, and hence density gradients, of the upper ocean, thus influencing the vertical structure of phytoplankton distributions and abundances. From an ecological perspective, sea ice is a habitat, feeding site, refuge and breeding ground for organisms at all trophic levels. The marginal ice zone (MIZ), delimited by the influence of low density meltwater and scattered ice flows from the receding pack ice and by the penetration of ocean swell into the pack ice, is an ecosystem boundary where the flow of energy, the cycling of chemical elements and biological communities change dramatically. Often the MIZ is also an area of high biological productivity [. Nelson smith produced 1985, Wilson 1986, Nelson 1987, Dieckmann 1987, Sullivan 1988, Nelson smith muench gordon sullivan husby 1989, Nelson smith autumn 1990, Comiso maynard smith sullivan 1990, Comiso mcclain sullivan ryan leonard 1993 .]. Thus, sea ice is a key environmental variable of the Southern Ocean, and quantification of sea ice parameters, such as concentration, extent, thickness and timing of autumn advance and spring retreat, is essential to understanding linkages between sea ice and the Southern Ocean marine ecosystem. A central hypothesis of the Antarctic Marine Long Term Ecological Research (LTER) program [. Quetin Ross 13 1992, Ross Quetin 03 1992 .], based at Palmer Station on the southwest side of Anvers Island (Figure 1), is that biological processes in the Antarctic marine environment are strongly affected by physical factors, in particular the annual advance and retreat of pack ice. Interannual cycles and/or trends in annual pack ice are hypothesized to have significant effects on all levels of the food web, from total annual primary production to breeding success in seabirds [. ainley fraser sullivan torres smith science 1986, fraser ainley 1986 bioscience, vidal smith 1986, nelson smith importance 1986, walsh mcroy 1986, garrison buck fryxell 1987, ainley fraser daly 1988 .]. In addition to the direct ecological consequences, variability in sea ice coverage is a critical element in the complex atmosphere-hydrosphere-cryosphere global climate system. Ice-atmosphere [. Streten Pike 1980, Carleton 1981, Cavalieri Parkinson 1981, Budd 1982, Parkinson Cavalieri 1982, Zwally comiso parkinson campbell carsey gloersen 1983, Carleton Carpenter 1989 .] and ice-oceanic [. Gordon Taylor 1975, Gordon seasonality 1981, Hibler Ackley 1983 .] coupling have been demonstrated at various strengths and over a wide range of spatial and temporal scales. In addition, variability in sea ice may be a possible indicator of long-term climate change [. Cavalieri Zwally 1985, Mitchell greenhouse 1989, Manabe stouffer spelman bryan 1991, Manabe stouffer spelman 1992 .]. As a consequence of its central ecological importance, sea ice may play a key role in coupling physical and biological processes on shorter time scales to longer term processes. In the following we document the interannual and annual variability in western Antarctic Peninsula sea ice coverage from October 1978 to August 1994, with particular emphasis on the LTER study area. We make use of satellite data, both low resolution passive microwave and high resolution visible and infrared data, to investigate spatial and temporal variability of sea ice and the marine ecological significance of this variability. In addition, we directly compare the low and high resolution sea ice estimates and discuss the ecological implications of the low versus high resolution perspective. In Section 2 we provide a brief background for the remote sensing of sea ice, and in Section 3 we give a discussion and evaluation of sea ice concentration algorithms. In Section 4 the methodology for sea ice data analysis is described, and results are presented in Section 5. Discussion and summary are given in Section 6.