1996-2002 Proposal for Palmer LTER PALMER LONG-TERM ECOLOGICAL RESEARCH PROJECT R. C. Smith, PI, K. S. Baker, W.R. Fraser, E.E. Hofmann, D.M. Karl, J.M. Klinck, L.B. Quetin, R.M. Ross, W.Z. Trivelpiece, M. Vernet, Co-PIs 1. Summary & Results of Prior Support 1.1 Summary: "Long-Term Ecological Research on the Antarctic Ma- rine Ecosystem: An Ice-Dominated Environment" (DPP-9011927; 10/1/90 to 9/31/96) The present proposal requests funds to continue, for a second six year period, the Palmer Long-Term Ecological Research (PAL) pro- gram which focuses on the marine ecosystem in the western Antarc- tic Peninsula (WAP) region. The addition of the PAL to the LTER Network in October 1990 extended the geographical & ecological range significantly & provided the opportunity to link ecological processes between hemispheres as well as across terrestrial & ma- rine biomes. A central tenet of the PAL is that the annual advance & retreat of sea ice is a major physical determinant of spatial & temporal changes in the structure & function of the Antarctic marine ecosystem, from total annual primary production to breeding suc- cess in seabirds. We are currently evaluating a number of te- stable hypotheses linking sea ice to: - the timing & magnitude of seasonal primary production, - the dynamics of the microbial loop & particle sedimentation, - krill abundance, distribution, & recruitment, & - the breeding success & survival of apex predators. The overall objectives of the PAL are to: (1) document the interannual variability of annual sea ice & the corresponding physics, chemistry, optics, primary production & the life-history parameters of secondary producers & apex preda- tors within the PAL area, (2) create a legacy of critical data for understanding ecological phenomena & processes within the Antarctic marine ecosystem, (3) identify the processes that cause variation in physical forc- ing & the subsequent biological response among the representative trophic levels, (4) construct models that link ecosystem processes to environmen- tal variables, which simulate spatial/temporal ecosystem rela- tionships, & employ such models to predict & validate ice- ecosystem dynamics. Since 1991 the PAL program has included spatial sampling during annual & seasonal cruises in portions of our regional grid in the WAP region (Fig. 1a) & temporal sampling from spring through fall (October to March) in the area adjacent to Palmer Station (Fig. 1b). Our program was designed to sample at multiple spatial scales within one regional scale grid, permitting repeated sam- pling on both seasonal & annual time scales, thus addressing both short & long-term ecological phenomena, as well as providing a basis for specific mechanistic studies. To date, there have been seven regional PAL cruises & two additional cruises emphasizing microbial dynamics (Tables 1 & 2). Core & other variables rou- tinely sampled &/or occasionally monitored from shipboard during our annual cruises are listed in Tables 3 & 4, respectively. Tables 5 & 6 list those obtained from Palmer Station throughout each field season. Documentation & data storage are organized through an electronic hub at the Institute for Computational Earth System Science (ICESS) at the University of California at Santa Barbara which also serves as a data archive as needed. There are online definitions of core data, datasets & metadata, organized to facilitate rapid information exchange & online data documentation (Fig. 2). Core data currently available electroni- cally to PAL investigators are listed in Tables 7 & 8. The PAL program is multidisciplinary & currently involves ten Principle Investigators (PIs). PAL program milestones in the con- text of the LTER network as a whole are summarized in Table 9. In 1992 the component of microbial ecology, including measurements of carbon exchange between air-sea & water-sediment interfaces, was added to the program by the addition of D. Karl. Until this renewal, D. Karl was funded separately from the PAL grant, although his research was fully integrated & he has had full PI status with the PAL group. In this renewal his funding will be integrated with the PAL grant. In 1993 K. Baker was made a PI in recognition of the importance of data management to the program. In 1994 M. Vernet replaced B. Prezelin for the phytoplankton com- ponent. Work to date, summarized below & described in detail in publica- tions (Sect. 1.3), indicates that the PAL is ideally sited in a climatically sensitive region where ecosystem studies have the potential for detecting, against a background of natural varia- bility, long-term trends &/or human disturbance to the Antarctic ecosystem. It is an area where "natural" experiments can be con- ducted to investigate mechanisms linking physical forcing & ecosystem response under vastly different year-to-year climatic conditions & where perturbations such as global warming, ozone- related UV-B increases & associated anthropogenic impacts on eco- logical processes can be studied in an otherwise remote & rela- tively pristine environment. The PAL location makes an ideal na- tural laboratory in which to study long-term trends & cycles. An overview & synthesis of various aspects & processes associated with the Antarctic marine ecosystem in the WAP & in the vicinity of Palmer Station, as well as analyses of historical data & early PAL results, will be published in early 1996 in a book entitled, Foundations for Ecological Research West of the Antarctic Penin- sula. Ten chapters were authored or co-authored by PAL PIs. During the first several years of the PAL program, considerable effort was devoted to establishing a sampling & research routine for long-term research in this area. This early effort is now beginning to pay dividends in a significant number of results (Table 10 & Sect. 1.2) & publications (Sect. 1.3). The upward ramp in productivity by the PAL (shown in Table 10) is in agree- ment with the exponential intellectual growth expected for long- term programs as described by Likens (1985). 1.2 Results of Prior Support R.C. Smith (remote sensing & bio-optics component) & B.B. Prezelin & M. Vernet (primary production component) The combined efforts of these two components have focused on (1) a sea ice & surface air temperature climatology, (2) the spatial & temporal variability of phytoplankton biomass & primary produc- tion, & (3) factors controlling phytoplankton biomass, production & community structure. The approach included analyzing (1) sa- tellite data to derive ice coverage & phytoplankton biomass dis- tributions, (2) surface air temperature data taken by research & automatic weather stations in the WAP region, & (3) bio-optical field measurements of photosynthetically available radiant energy (PAR), pigment biomass & plant pigments (HPLC) & C-3 14 measure- ments for phytoplankton growth & physiological parameters. There has been full participation to date in all field cruises & nearshore time series sampling by both of these components. Results from the sea ice & surface air temperature climatologies in the PAL region show (1) an annual sea ice cycle involving a relatively short period of ice retreat (about 5 months) followed by a longer period of ice advance (about 7 months), which con- trasts with other regional annual cycles where ice retreat is longer than ice advance, (2) a confirmation that while the South- ern Ocean as a whole shows relatively little interannual varia- bility (< few %), deviations up to 66% & 130% from mean maximum & minimum ice coverage, respectively, have been observed for the PAL region, (3) a confirmation of a significant warming trend (+4-5(degreesC) in mid-winter surface air temperatures in the WAP region over the past half-century (1941-1991), (4) a significant anti-correlation between surface air temperature & sea ice ex- tent, & increased variability in fall & winter observed in both variables, & (5) a long-term persistence in monthly sea ice & surface air temperature anomalies (Fig. 3), wherein two to four high ice /low temperature years are followed by one to three low ice /high temperature years, a pattern which is coherent with the Southern Oscillation Index (SOI), suggesting teleconnections between the WAP & lower latitudes. Significant findings from phytoplankton studies include: (1) a persistent on/offshore gra- dient modulated alongshore by latitudinal variability which fol- lows the annual advance & retreat of sea ice, (2) a strong sea- sonal & interannual variability in timing & intensity of phyto- plankton blooms, strength in on/offshore & alongshore gradients & overall annual biomass accumulation, (3) a consistency between the regional gradients in phytoplankton biomass & the temporal observations from Palmer Station, suggesting that the dynamics in chl-a accumulation in the vicinity of Palmer Station are representative of those within the shelf waters & that the sea- sonal nearshore variability is related to the PAL regional varia- bility in chl-a biomass, (4) a tight coupling between productivi- ty & daylength, with more than 90% of the biomass & primary pro- duction occurring between November & April (not including a pos- sible & currently unknown contribution during winter), (5) a sea- sonal structure that, generally, includes two major pulses in primary production in late spring (Dec/Jan) & in summer (Feb/Mar), where the first pulse is generally larger, although in 'high' production years (i.e., 1994-95) three large pulses were observed between mid-Dec & late Feb. D.M. Karl (microbial dynamics & carbon flux; OPP- 9118439) To date, the research supported by this component has focused on three related topics: (1) dissolved inorganic carbon (DIC) & or- ganic carbon (DOC) pool dynamics, (2) controls on the distribu- tions, abundances & growth rates of bacteria, & (3) particle sed- imentation & export production. The approach involved making re- peated measurements of carbon pools (total DIC & DOC, bacterial biomass) & carbon fluxes (bacterial production, dissolved organic matter turnover rates, particle sedimentation) in the PAL region during the annual cruises, & conducting field experiments to test various ecological predictions of the general PAL hypothesis re- garding the role of ice in the control of ecosystem processes & carbon/energy flux. By the end of this current project period, we will have participated in six PAL cruises in addition to two spe- cial focus microbial dynamics cruises. Several significant results include: (1) documentation of carbon dioxide partial pressure (pCO2) by continuous underway measurement which reveals significant reductions in pCO2 associated with the spring phyto- plankton bloom, with values as low as 260 ppm compared to atmos- pheric values of 340-350 ppm, suggesting that the entire PAL re- gion may be a moderate sink for atmospheric CO2 during summer, (2) confirmation of our previous observations supporting a tem- poral decoupling between primary production by phytoplankton & secondary production by bacterioplankton (reasons for this enig- matic decoupling are not known but have important implications for parameterization of secondary production & nutrient regenera- tion in polar waters & in our modeling of PAL region primary pro- duction), (3) quantification of the downward flux of particulate matter which displays annual variations in excess of 1000-fold from values of 2 g m-2 d-2 in summer to <1 mg m-2 d-2 in winter & which displays interannual variability, with highest export in heavier than average ice years, & (4) in collaboration with Ed DeLong (UCSB), continued investigation of the relatively high abundances of archaebacteria in the water column near Palmer Sta- tion (up to 30% of the total rRNA), an observation that derived from B. Prezelin's PAL research. L.B. Quetin & R.M. Ross (prey component) Research by this component focused on three topics: (1) zooplank- ton abundance & community structure, with particular focus on Eu- phausia superba (Antarctic krill), (2) krill distribution, number & size of aggregations, & (3) krill/sea ice interactions. Our general approach involved sampling the water column with oblique- ly towed 1-m or 2-m nets while simultaneously obtaining acoustic backscatter signals from a transducer (BioSonics, 128 kHz). Specimens were analyzed immediately, preserved or used for phy- siological experiments. To date, the research for this component has resulted in several significant findings. (1) Estimates of short-term grazing pressure & flux of fecal pellet carbon to the sediments, calculated in conjunction with PAL data on phytoplank- ton biomass & production, vary with the temporal & spatial abun- dance & community composition of larger zooplankton (Table 11). (2) Krill aggregations, distribution, number, size & composition vary on several temporal scales. (3) Integrated krill biomass levels & length frequency distributions vary with season. During summer, there is a strong on/offshore gradient, with small krill abundant & dominant inshore, whereas the larger reproducing krill are primarily located offshore on the outer continental shelf & at the shelf break. This on/offshore gradient in krill size parallels that found for phytoplankton biomass for all cruises. During fall & winter all size classes shift towards inshore re- gions, & large aggregations (which contain a disproportionate amount of the biomass) become most abundant inshore. (4) There was a sharp decrease in the spatially averaged krill biomass dur- ing the transition from summer to fall in 1993, which suggests that large scale shifts in biomass distribution are possible in short time periods. This has implications for estimates of total krill biomass, as well as for local penguin populations, causing changes in foraging behavior, both within a season & interannual- ly. (5) Quantitative diver censuses of the under-ice population (August 93) show that the majority of the krill which are closely coupled to the ice in winter are the young-of-the-year or AC0 (Age Class 0, krill in the first year of existence) population, whereas adults are rare. The differences in AC0 & adult krill distributions in winter are attributed to 'risk balancing'. (6) Interannual variability in recruitment success for Antarctic krill, as shown by patterns in length frequency distributions from the mesoscale grid for the four spring/summer cruises (Fig- ure 4), is extremely high & follows the pattern predicted from initial PAL hypotheses concerning the effect of sea ice on winter-over survival of AC0s (ie., strongest year classes are as- sociated with years in which winter ice is both above average in extent & retreats late in spring). W.R. Fraser & W.Z. Trivelpiece (seabird component) Seabird research at Palmer Station, which has been supported both by the National Science Foundation & the National Oceanic & At- mospheric Administration through the National Marine Fisheries Service, encompassed basic & applied research, in which monitor- ing of key parameters related to seabird ecology played a criti- cal role. An important assumption guiding our general research approach is that the persistence of any seabird population over time ultimately reflects the coincident availability of suitable nesting & foraging habitats. We concluded the first funding phase of the PAL program with the development of a conceptual model that addressed the role that sea ice has in affecting the availa- bility of these habitats. Our approach embraced a step-wise series of analyses involving new & historical data on Pygoscelis adeliae (Adelie penguin) demography, foraging ecology, behavior & biogeography. Three key results were particularly influential in structuring model development. (1) Population changes of many Southern Ocean, upper-trophic level predators during the last five decades are better explained by changing winter sea ice con- ditions resulting from environmental warming than by the long- accepted tenet that depletion of baleen whale stocks led to an increase in krill availability & competitive release. (2) Varia- bility in the terrestrial breeding habitat of Adelie penguins, a result of changing precipitation patterns because of increasingly warmer winters with less sea ice, is an important, but only re- cently recognized factor affecting long-term reproductive perfor- mance & recruitment. Because the disappearance of breeding groups can be fairly rapid, it is suggested that very short term changes in the environment can result in rapid shifts of the po- pulations in an area. (3) There is a direct causal relationship between variability in sea ice cover, krill recruitment, krill availability & predator foraging efficiency. Based on these results, the aforementioned model describes a "habitat optimum" for Adelie penguins that varies in space & time in accordance with changing atmospheric & oceanic processes & the consequent effects on sea ice formation, which ultimately mediates the avai- lability of breeding & foraging habitats. The model & the ana- lyses on which it is based have multiple implications for South- ern Ocean ecosystem studies & illustrate how community structure & function may respond to climate change. E.E. Hofmann & J.M. Klinck (modeling & physical oceanography com- ponent) The ODU component of the PAL took responsibility for (1) process- ing & analysis of hydrographic data, (2) describing the hydrogra- phy & circulation in the WAP region & over Antarctic continental shelves in general, & (3) developing circulation & coupled physical-biological models. The approach consisted of partici- pating in the collection of & all post-processing of hydrographic measurements from the first four PAL cruises. Results from his- toric and PAL hydrographic data show that the Antarctic shelf re- gions are greatly influenced by a warm, salty water mass, the Circumpolar Deep Water (CDW). The circulation pattern in the WAP region shows that the large-scale geostrophic flow may be com- posed of one or more clockwise gyres. This mesoscale variability is likely the result of the rugged bottom topography & has impli- cations for the transport & retention of physical & biological properties. Surface drifters indicate that the circulation in Bransfield Strait is clockwise & may be continuous with the cir- culation in the WAP above 500 m. Also, the westward flowing Po- lar Slope Current, which has been observed north of the South Shetland Islands, does not appear to extend beyond Smith Island. The Princeton Ocean General Circulation Model (POGCM) has been configured to use geography, bathymetry, & hydrographic climatol- ogy for the PAL region, & simulations are now being done with this model to investigate processes underlying the circulation in the WAP region. In support of ecosystem synthesis studies, we have also analyzed some of the multidisciplinary data sets col- lected during the first four PAL cruises in an effort to under- stand krill distributional patterns in relation to other habitat characteristics. This work showed the need for a mathematical model on the growth & development of krill larvae which has been developed & is now operational. K.S. Baker (data management component) Data management is discussed in detail in Section 5, which also provides a discussion of the PAL data policy & a guide to PAL on- line data. 2. Project Description 2.1 Antarctic Marine Ecosystem - Introduction The Antarctic marine ecosystem, the assemblage of plants, an- imals, microbes, ocean, sea ice & island components south of the Antarctic Convergence, is among the largest readily defined biomes on Earth (36 x 10^6 km2) (Hedgpeth 1977; Petit et al. 1991). This environment is composed of an interconnected system of functionally distinct hydrographic & biogeochemical sub- divisions (Treguer & Jacques 1992) & includes open ocean, frontal regions, shelf-slope waters, sea ice & marginal ice zones. Oce- anic, atmospheric, & biogeochemical processes within this system are thought to be globally significant, have been infrequently studied & are poorly understood relative to more accessible ma- rine ecosystems (Harris & Stonehouse 1991; Johannessen et al. 1994). The PAL area west of the Antarctic Peninsula (Fig. 1a) is a complex combination of a coastal/continental shelf zone (CCSZ) & a seasonal sea ice zone (SIZ), as this area is swept by the yearly advance & retreat of sea ice. Polar regions are unique in that sea ice, a dominant & distin- guishing characteristic of Southern Ocean marine ecology, forms a range of habitats for animals as well as extensive & varied sur- faces for algal & microbial populations. The influence of sea ice on nearshore marine ecosystems is pervasive & strongly af- fects the vertical zonation of sublittoral habitats (Clarke 1996). In the pelagic realm, sea ice not only provides a surface for algae but is also thought to provide a refuge & an important wintertime grazing area for juvenile krill at the ice/water in- terface (Smetacek et al. 1990; Ross & Quetin 1991). In addition, different springtime seabird habitats are associated with varying sea ice coverage, which therefore alter trophic level interac- tions, foraging efficiency &, ultimately, breeding success (Hunt 1991; Ainley et al. 1994). These habitats include: (1) open leads & polynyas, through which seabirds can gain access to the water column & underside of sea ice, (2) the ice edge, which is a major ecological boundary and can either be compact or diffuse, & (3) the outer marginal ice zone (MIZ), where meltwater contri- butes to stabilization of the water column & provides the poten- tial for enhanced phytoplankton growth. The MIZ is an area bounded on the open ocean side by the stabilizing influence of meltwater & on the pack ice side by the penetration of ocean swell. The physical action of ocean swell imparts distinctive structure to Antarctic sea ice (Ackley et al. 1979) & creates a range of ice-related habitats which support the development of diverse biological sea ice communities (Legendre et al. 1992; Ackley & Sulllivan 1994). The MIZ can be up to 250 km in width & is often an area of high productivity (Smith & Nelson 1985) rela- tive to the open ocean. It is therefore an ecosystem boundary where the flow of energy, the cycling of nutrients & the struc- ture of biological communities change dramatically, both tem- porally & spatially. In addition, the areal extent of sea ice cover & the associated timing of the MIZ in relation to specific geographic areas (e.g., seabird rookeries) have high interannual variability. The range for variable ice cover is shown schemati- cally in Fig. 1a. Sea ice extent for low (1989), average (1993) & high (1987) ice years are indicated (Stammerjohn & Smith 1996). Mesoscale oceanic circulation patterns in the WAP region are not presently well-known (Hofmann et al. 1996). The barotrophic cir- culation is also not known, & there are no direct current meas- urements anywhere in the region. Even tidal characteristics are poorly known (Amos 1993). There are some suggestions of the direction of flow on the shelf from dynamic calculations & other indirect means. The Antarctic Circumpolar Current (ACC) clearly stands out in all calculations & flows northeasterly off the shelf. There are a number of indications of southward flow on the inner part of the shelf. This is shown schematically in Fig. 5. Some studies suggest the existence of two gyres (Stein 1992), but it is not certain whether permanent eddies are part of the coastal circulation, or if the coastal flow is merely convoluted & not sampled sufficiently to resolve the features. The most prominent water mass in the WAP is Circumpolar Deep Water (CDW) is characterized by temperatures above 1.5degC, salinities of 34.6 to 34.8 & oxygen values below 4.5 ml l^-1. It is found between 200 & 700 m & present in all seasons throughout the region exam- ined. Below 200 m this water mass floods the WAP continental shelf. CDW is also found in Bransfield Strait, but with distri- bution limited to the northern side of the Strait near the South Shetland Islands (Hofmann et al. 1996). The CDW brings macronu- trients & dissolved inorganic carbon, in addition to warm salty water, onto the shelf. The presence of this water mass near the bottom of the mixed layer has considerable implications for heat, salt & carbon budgets in this region. Also, the circulation pat- terns & the presence of CDW may impact the timing & coverage of sea ice, as well as the transport &/or retention of physical & biological properties in the PAL area. Smith et al. (1996a) have recently reviewed historical phyto- plankton biomass & productivity data for the PAL region & con- clude that the average productivity of the region is on the order of 100 to 200 g C m-2 y-1 which is roughly about a factor of 5 lower than other productive coastal areas of the world's oceans (Chavez & Barber 1987). Factors that regulate primary produc- tivity within this environment include those that control cell growth (light, temperature, & nutrients) & those that control the accumulation rate of cells & hence population growth (water column stability, grazing, sinking & advection). It is important to note that the process of primary production is a complex in- tegration of these independent factors & may vary independently from total biomass. For example, the accumulation of chl-a dur- ing a bloom may result either from a moderate rate of primary production in the absence of grazing or from a high rate of pro- duction in the presence of grazing. Sea ice mediates several of these factors & often, but not always, conditions the water column for an initial bloom which is characterized by a pulse of production restricted in both space & time. The relationship of the relative contribution of bloom versus non-bloom primary pro- ductivity to the timing & balance of the resulting food web dynamics is currently unknown. Also, we only roughly know the relative contribution of microalgal & bacterial production of biogenic carbon in the MIZ, under pack ice & within sea ice (Legendre et al. 1992; Smith et al. 1995b) & we identify this as a critical area for our further research. Finally, there appears to be a metabolic uncoupling, in time, between phytoplankton & bacterioplankton activities in Southern Ocean habitats contrary to studies conducted in temperate marine ecosystems (Karl 1993), & the timing & fate of biogeochemical fluxes of carbon associated with these various pelagic & sea ice-related communities is presently unknown. Like most marine food webs, the trophic relationships in An- tarctica are complex. However, the links between primary produc- ers, grazers & apex predators (seabirds, seals & whales) are often short & may involve fewer than three or four key species (Fig. 6). Predators tend to concentrate on a core group of prey species, for example, the abundant euphausiids & fish close to the base of the food web. The general sampling approach in our study capitalizes on the close coupling between trophic levels, the limited number of species involved, & the fact that one of the dominant predators is seabirds that nest on land & are thus easily accessible during the spring & summer breeding season. An important apex predator is the Adelie penguin, which dominates the seabird assemblage near Palmer Station in terms of abundance & biomass. During the summer breeding season, these seabirds depend on resources found within their foraging range, which is thought to be within 100 km of the breeding sites (Fraser & Triv- elpiece 1996a). Finding sufficient prey within these foraging ranges is critical to the reproductive success of these birds. Preferred prey for Adelie penguins are krill, a keystone species with a circumpolar distribution. However, the timing of prey availability & abundance is highly variable from year to year. 2.2 Conceptual Model of Physical & Biological Linkages Solar radiation, atmospheric & oceanographic circulation as well as sea ice coverage are the physical forcing mechanisms driving variability in biological processes at all trophic levels in An- tarctica. Extreme seasonality & the relatively large interannual variability (both in magnitude & timing) of physical forcing may be compared & contrasted with conditions for biological growth, development & survival of key species from each trophic level, providing a conceptual model for the discussion of trophic link- ages (Fig. 7). As discussed below, some of the components in Fig. 7 are known with relative certainty, while others are sug- gested according to our current knowledge & related hypotheses. A key point is to first identify, then understand & ultimately model these temporal linkages. Long-term, systematic time-series data are essential for this effort. Daylength, precipitation, air temperature, & sea ice cover have a strong seasonal cycle at high latitudes. Weather, while mild by comparison to the interior of Antarctica, is punctuated by ep- isodic storm events, is modified by a changing seasonal balance between relatively warm & moist maritime & relatively cold & dry continental influences, & shows more year-to-year variability in winter (Smith et al. 1996b). In spite of this large interannual variability, statistically significant warming trends have been detected in the relatively short WAP station records (Smith et al. 1996b & references therein). However, until 1994 weather data for the WAP region were available only from stations located on the coast or on islands near the coast. At the request of the PAL, automatic weather station (AWS) units (Bromwich & Stearns 1993) have now been installed at the end of Bonaparte Point near Palmer Station & in the Hugo archipelago, a small group of low lying islands, approximately 90 km west southwest of Palmer Sta- tion (Fig. 1b). Data from coastal (AWS Bonaparte) & oceanic (AWS Hugo) stations help to characterize the different weather regimes & relative influences of a coastal versus a more oceanic environ- ment. With the AWS on Racer Rock in the Gerlache Strait, there is now a suite of three AWS installations within the PAL study area, providing more continuous weather data, in addition to the weather records from Palmer, Faraday & Rothera Stations (Baker & Stammerjohn 1995). There are also seasonal variations in the hydrography of the PAL. For example, in winter, temperature & salinity are uniform down to approximately 100 m (Hofmann et al. 1996). Beginning in spring & during summer, the ocean warms from the surface, creat- ing one or more layers of warmer water that are a few tens of me- ters thick. Episodic storms mix these layers so that the verti- cal temperature structure near the surface can be quite variable. The freshening of the surface layer during summer due to sea ice & glacier melt also can play a key role in stabilizing the water column & hence increase rates of primary production. Although variations in salinity are small, the density & stability of the upper water column are more a function of salinity than tempera- ture. It appears that the CDW floods the shelf throughout the year, characterizing the deep waters in the PAL study area. The annual heat flux associated with the CDW (12 W m-2) is sufficient to be an important factor in determining the extent & thickness of sea ice in the WAP region. It is therefore critical to under- stand the factors controlling the variability of the annual heat flux associated with the CDW. The timing & extent of the annual advance & retreat of sea ice in the vicinity of Palmer Station (i.e., the Adelie penguin foraging area) are highly variable & have been discussed in detail (Stam- merjohn 1993; Stammerjohn & Smith 1996). The satellite time series is still too short to statistically resolve any persistent low frequency periodicities, but there is evidence that high in- terannual variability in magnitude, timing of ice advance & re- treat, duration of near maximum & minimum ice coverage, & ap- parent clumping of high or low ice years have significant impacts on the survival rates, distributions &/or life histories of key indicator species (Ross & Quetin 1991a; Quetin & Ross 1992; Quetin et al. 1994; Siegel & Loeb 1995; Fraser et al. 1992b; Fraser & Trivelpiece 1996a). Further, we are beginning to define the physical forcings (e.g., thermodynamics versus advective processes) controlling variability in sea ice coverage & the resulting ecological consequences. Our current models of the trophic organization of Antarctic ma- rine ecosystems have evolved considerably during the past decade. Prior to 1980, energy flow in Southern Ocean habitats was thought to be dominated by relatively short & therefore efficient transfers from large (>20 Mm) phytoplankton cells to krill & sub- sequently to apex predators. More recently, our concept of the marine food web expanded to reflect the potential roles of heterotrophic microorganisms including bacteria, protozoans & small (<150 Mm) non-krill crustaceans. Heterotrophic microorganism-based food webs, also referred to as microbial loops (Azam et al. 1983), are present in all aquatic environments including the Antarctic marine ecosystem. These detritus driven systems are fueled by non-respiratory community carbon losses, including dissolved & particulate organic matter released by ex- cretion, predation & mortality. Because microbial loops require several trophic levels to transfer carbon & energy to apex preda- tors, most detritus based food webs are inherently inefficient & sometimes constitute major energy sinks. It is important to em- phasize that comprehensive, quantitative ecosystem studies of en- ergy & carbon flow in Antarctic food webs do not exist. At best, there are order of magnitude estimates for a few selected re- gions. A major, unexpected result of the Antarctic field studies conducted to date is the apparent short-term uncoupling of algal & bacterial metabolic processes (Cota et al. 1990; Karl et al. 1991; Karl & Bird 1993). Reasons for this uncoupling are not well understood at present. Consequently, we must view the mi- crobial loop models as hypotheses that deserve a thorough, quan- titative field evaluation. Sea ice provides one of the major habitats for microorganisms in Antarctic marine ecosystems, & it is possible that some microbial food webs may be entirely ice-associated (Palmisano & Garrison 1993). In the PAL program, we have just begun to systematically investigate these unique habitats. Preliminary results suggest that the bacterial community goes through several, as yet poorly defined, stages of succession as the annual pack ice recedes (Christian & Karl 1995a). The water is initially seeded with bacteria, as well as substantial inputs of organic matter that may be quite different in chemical composition than the organic matter in the water column. Phytoplankton blooms at the receding ice edge provide additional sources of organic matter, through excretion, lysis, & grazing, that are likely to be present only for a short time. Furthermore, photochemical alteration of dis- solved organic matter may be accelerated in the sea ice habitat, because it is immobilized in a region of high UV & visible light flux. These processes are likely to play important roles in the adaptation of marine bacteria to this seasonally variable en- vironment, & the biochemical characteristics of the bacteria may change rapidly with time. Future studies will focus on the quan- titative role that variations in ice cover may have on the pres- ence & intensity of microbial processes. Associated with the increase in daylength & the melting of sea ice, both contributing to water column stratification, phyto- plankton biomass (chl-a) in the PAL area starts to accumulate near the end of November in an average ice year. In non-average ice years the subsequent timing shifts accordingly. Mean chl-a in the top 30 m can increase from <0.5 mg m-3 in a pre-bloom period, to higher than 15 mg m-3 during a spring bloom, with average values between 1 & 3 mg m-3 (Smith et al. 1996a). Blooms are mostly dominated by cells >20 Mm, which are typically large or chain-forming diatoms, although cells <20 Mm also grow during a bloom (i.e. cryptomonads & prasinophytes). On occasion, the bloom is dominated by the colonial prymnesiophyte Phaeocystis pouchetii. Smaller diatoms & flagellates dominate during non- bloom & post-bloom periods (Holm-Hansen et al. 1989). The termi- nation could be caused by cell advection, sinking &/or zooplank- ton grazing which leads to recycling of materials in the euphotic zone & reduces export particle flux. Large fluxes of organic matter (> 1-2 gC m-2 d-1) have been observed in sediment traps, sometimes after a period of intense erosion of the mixed layer, as might occur during a storm. In contrast, the development of a massive coastal bloom, even for a short period of time (days or weeks), can significantly decrease inorganic nutrients which can also lead to bloom termination. Surface nitrate concentrations in coastal waters of the WAP region can decrease from 25 MM, when non-bloom chl-a concentrations are typically 0.5 mg m-3, to near depleted levels (<0.1 MM), when bloom chl-a concentrations can be >35 mg m-3 (Kocmur et al. 1990). The coastal bloom, in terms of continued new/export production, crashes unless nutrients are then replenished by mixing & erosion of the mixed layer or by ad- vection of richer offshore waters. However, primary production can continue without allochthonous nutrients, if nutrients are locally regenerated, but new production ceases. During summer, in both open ocean & coastal areas, chl-a concentrations remain intermediate (1-2 mg m-3), although periodic blooms may still oc- cur nearshore. Fall phytoplankton blooms may appear in late February & March. The relative contribution of MIZ related production to overall production continues to be debated. Smith & Nelson (1986) & Smith et al. (1987b) show evidence that the MIZ, especially in spring, supports high phytoplankton biomass &/or high production rates. On the other hand, recent observations (Lancelot et al. 1993; Boyd et al. 1995) suggest that specific meteorological con- ditions influence whether blooms do or do not occur in the MIZ. Nonetheless, total annual productivity is thought to be dominated by the high production rates associated with spring blooms, whose development may be timed & paced by ice-driven water column sta- bility &/or favorable meteorological conditions. The timing of this burst of productivity & consequent food availability for prey (krill) & subsequently, predators (penguins), as well as the habitat considerations associated with these environmental condi- tions, creates a complex trophic matrix & associated temporal linkages (Figs. 6 & 7). These couplings are subject not only to the progression of the seasons, but also to episodic events that disrupt &/or reset the cycle of water column stability, phyto- plankton productivity & subsequent linkages. Prey/predator trophic interactions (ie., krill/Adelie penguins) are strongly mediated by critical periods during reproduction of both prey & predator. Our chosen prey/predator pair is composed of relatively long-lived species. Antarctic krill live for 5 to 7 years, reproducing as early as their third summer. Ovarian development begins in austral spring, & the rate of ovarian development is dependent on food availability. A prolonged spawning season runs from late December to early March. Both the proportion of the population reproducing & the length of the spawning season are dependent on food availability during spring & summer. Release of ice algae from melting ice & ice-edge blooms, occurring prior to open-water blooms, are thought to be one source of food essential to high reproductive output throughout the summer (Quetin et al. 1994). After spawning the embryos sink, hatch at depth, & the early non-feeding AC0s ascend through the water column. The first critical period occurs when early AC0s arrive at the surface, about three weeks after release & need food. Winter is the second critical period, because un- like adults, larvae lack energy stores. The six-month fall & winter period of low food availability in the water column can create starvation conditions for the AC0s (Ross & Quetin 1991a). The essential winter grazing ground for AC0s is the under-ice ha- bitat where they feed on ice algae. Thus, recruitment & AC0 sur- vival & growth are hypothesized to be enhanced by the presence & duration of winter ice. Adelie penguins have a circumpolar distribution & a breeding sea- son that passes through a series of stages. The season begins with a three week courtship period, during which time both members of the pair remain ashore fasting. This is followed by egg-laying & a month long incubation period from late November through December. The incubation duties begin with the male on the nest which requires him to fast for an additional two weeks, while the female goes to sea to feed. During this first critical period the female is believed to return to the ice edge in search of a predictable source of krill in order to restore her body condition (Trivelpiece & Fraser 1996). If the female fails to replenish her supply of fat & return to the nest within two weeks, the male abandons the nest to forage, & the eggs are lost. If the female is successful in finding food, she relieves her mate at the nest, & he spends the following two weeks at sea re- covering from his five weeks of fasting. Upon the return of the male, the pair alternate between attending the eggs & foraging at sea on progressively shorter time intervals, until they are switching duties daily by the time the eggs (usually two) hatch. Following hatching, the pair continue alternating between guard- ing the chicks at the nest & foraging for food for their young until the chicks reach approximately three weeks of age. A second critical period affecting breeding success occurs in mid to late January, when the chicks are between three & seven weeks old. During this "creche stage", the food/energy demands of the chicks are at their highest. The parents must find adequate sup- plies of prey (typically krill) within a foraging area of about 100 km, or preferably much closer, otherwise breeding success may be significantly reduced. 2.3 Working Hypotheses The PAL program remains focused on understanding the ecological role of sea ice with the primary object being to gain a general understanding of the physical & climatic controls on interannual sea ice variability, the effects of this variability on trophic interactions, & the biogeochemical consequences thereof (PAL Group 1996; Smith et al. 1995c). Our observational & experimental programs reflect this primary research goal. In the following, we restate the central null hypothesis & present several alternate hypotheses that together comprise our integrated, transdisci- plinary research prospectus. We then discuss in both general & specific terms the various ecological predictions that derive from the alternate hypotheses through the use of conceptual models that illustrate the hypothesized coupling between sea ice & biological processes at various trophic levels. H0: Neither the presence nor the extent of annual sea ice in the PAL study area influences ecosystem structure & dynamics. HA1: Interannual variations in ice dynamics are a quasi-predictable manifestation of ocean & atmospheric circulation processes which influence the extent of CDW upwelling onto the continental shelf. The presence of this warm, salt water mass affects the heat & salt budgets of the region & hence controls ice dynamics. HA2: Primary & secondary production are enhanced during high ice years, causing a general intensification of biogeochemical cy- cling rates & particle export processes. HA3: Increased food production & under-ice refuge during sequential above normal sea ice years promotes optimal recruitment & growth of krill &, in subsequent years (1-2 yr lag), a greater breeding success & survival of apex predators (e.g., Adelie penguins). HA4: The exchange of carbon dioxide between the atmosphere & the sur- face ocean is influenced by ice dynamics, CDW upwelling, primary & secondary production rates & organic particle sedimentation. The WAP region may be an important, albeit temporally-variable, source/sink term in global carbon budgets. Ackley & Sullivan (1994) have described the seasonal cycle of pack ice development, formation, & the entrainment, accumulation & growth of microalgae within various pack ice micro-environments (Ackley & Sullivan, Fig. 1). Several other studies (Smetacek et al. 1990; Ross & Quetin 1991a; Quetin et al. 1996) have presented comprehensive summaries of the life cycle of Antarctic krill, in- cluding vertical distribution & timing of early life history stages in relation to seasonal cycles of daylight, ice cover & phytoplankton. Figures 8a&b present summary conceptual diagrams of these processes & the hypothesized linkages for the PAL re- gion. Key points of these figures are that the development of the sea ice community (1) is a dynamic process, (2) is a strong function of the distinct environmental conditions during sea ice forma- tion, development, deformation & subsequent melt, & (3) has a varied history which leads to discrete physical, optical, chemi- cal & biological features within distinct ice-related habitats. Consequently, the timing & extent of sea ice, from the formation to the subsequent deformation & eventual melting, plays a complex & not fully understood role in the ecology of sea ice biota. In- terior sea ice communities were first described by Ackley et al. (1978), & the physical & biological controls of Antarctic sea ice communities have been addressed by Ackley & Sullivan (1994) & Garrison & Mathot (1996), while Maykut (1986) & Lange et al. (1989) have described the formation of sea ice in detail. Some of these processes are conceptualized in Figs. 8a&b which show (1) the incorporation of biological material during fall frazil ice formation, (2) the in situ growth of material within the ice dur- ing the lifetime of the ice cover & subsequent deformation, & (3) the ice melt & release of biological material to the water column in spring. Microalgal sea ice communities include those scavenged from the water column during the formation of frazil ice & those deposited by flooding & rafting (Ackley & Sullivan 1994). There are few quantitative surface sea ice observations in the WAP region but qualitative shipboard observations suggest a region with little, if any, multi-year ice & significant (but not yet quantified) deformation & ridging prior to melt. A critical issue is the relative contribution of sea ice algae to the total overall primary production within this ecosystem. The estimates of Smith et al. (1995b) & Legendre et al. (1992) for the Southern Ocean are reasonably consistent (0.46 & 0.29 GtonC y-1, respectively) when considering only the CCSZ & SIZ (Table 12), however other historical estimates vary by more than an ord- er of magnitude. With caution, oceanic regional production esti- mates permit consideration of the relative contribution of dis- tinct hydrographic/biogeochemical sub-divisions of the Antarctic marine ecosystem & facilitate comparison among geographic re- gions. These estimates suggest that the contribution of sea ice algae to the base of the food web is quantitatively important, in addition to any contribution via conditioning of water column production. Also, a high proportion of sea ice production may be new, rather than recycled, production so that the impact may be particularly significant with respect to export flux (see discus- sion below). Further, these estimates, & associated assumptions & errors, emphasize the need for better characterization of the areas comprising the distinct biogeochemical zones (CCSZ, MIZ, SIZ) & the periods over which sea ice & water column blooms take place within these zones (e.g., "Sampling Strategy", Sect. 2.5). Figures 8a&b also conceptualize hypothesized linkages between the seasonal cycle of sea ice & krill. Note that a distinction is made between these linkages for AC0 & adult krill. AC0s are hy- pothesized to be obligate inhabitants of the underside of sea ice during winter (Ross & Quetin 1991a). Biological material en- trained during ice formation & the subsequent growth of microal- gae within the fall/winter sea ice can be utilized by AC0s. Over-rafted sea ice provides refuge for AC0s as well as increased surface areas, often with light-trapping properties, for enhanced ice algae populations. During the spring ice melt, ice algae released into the water column not only provides a possible ino- culum for a spring bloom, but a source of food for both AC0s & adult krill that are developing into reproductive condition. Figure 9 summarizes these hypothesized relationships between sea ice, microalgae & krill which will be discussed in greater detail below. Sea ice & associated phyto & zooplankton also play a key role in both the timing & magnitude of carbon flux exported from the eu- photic zone. In Figs. 10a&b, which show idealized examples of ecosystem structure & function, we hypothesize that there are two separate end-member biological controls on particle export: phy- toplankton control & zooplankton (krill) control under both low & high ice conditions. When krill population densities & herbivore grazing pressures are low (i.e., phytoplankton control of export, Fig. 10a), we hypothesize that large accumulations of phytoplank- ton (i.e., "blooms") will occur with attendant decreases in ni- trate & total inorganic carbon. In the PAL study area, these spring-summer blooms are common. In the absence of efficient grazing by macrozooplankton, the growth of photoautotrophs is eventually limited by nutrients (nitrate, or for diatoms, sili- cate) &/or by light. Under these environmental conditions, cell aggregation & sinking of whole, ungrazed phytoplankton cells dom- inates the total particulate matter export from the euphotic zone. However, the presence of ice may influence these processes by at least two independent mechanisms. (1) When the ice melts in early spring, the release of ice algae serves as both an agent for immediate particle export as well as for subsequent in situ growth & export processes. This seed population can influence both the timing of the bloom & the relative abundance of poten- tial phytoplankton species. (2) AC0 & adult krill associated with sea ice in spring could impose a grazing pressure on the phytoplankton population, suppressing the accumulation of chl-a &, therefore, the presence of a bloom in the early spring-early summer (Nov-Jan) period. However, another important consequence of grazing activity is that it could regenerate nutrients (espe- cially the conversion of particulate organic nitrogen, PON, to NH4+) which in turn could sustain the growing season, thus triggering phytoplankton species succession & a secondary bloom in the late summer to early fall period. A coupled particle ex- port pulse would accompany this fall bloom. When zooplankton are relatively abundant (Fig. 10b), their effi- cient grazing activities prevent the accumulation of chl-a & cause a fundamental shift in the nature of the exported materials from primarily intact phytoplankton cells to mostly fecal pellets (i.e., zooplankton control of export). An important biogeochemi- cal consequence of this shift is that the C:N & C:P ratios of fe- cal pellets are generally much greater than those of the phyto- plankton cells. This results in a stoichiometric uncoupling of export from production & favors a net removal of C, relative to N or P, from the surface layers of the ocean. This net sequestra- tion of particulate carbon in the subeuphotic zone & depletion of dissolved inorganic carbon in the surface layers establishes con- ditions necessary for the transport of atmospheric carbon into the ocean. Our direct measurements of the air to sea gradient of partial pressure of CO2 in the PAL study area (>100 ppm, with the sea water being lower) are consistent with this hypothesized role of the Antarctic coastal regions serving as a large but transient sink for atmospheric CO2. The regeneration of N by krill, pri- marily in the form of NH4+, also has important implications for new vs. regenerated production & for phytoplankton species suc- cession, as mentioned above. We hypothesize that an above aver- age ice year increases the probability & intensity of the "zoo- plankton control" of exporting particulate matter. Therefore, we suggest that phytoplankton control dominates in below average ice years & that zooplankton control dominates in above average ice years, primarily as a result of the role of sea ice in the krill life cycle (Figs. 8 & 9). 2.4 Motivation & Future Direction Phytoplankton Production & Export Processes Photoautotrophic mi- croplankton provide the base of the food web (Fig. 6) & therefore influences the temporal & spatial variability of higher trophic levels, while a fraction of this production is transported to depth providing a potential sink for CO2. Our observations show that the seasonal & interannual variability of phytoplankton biomass in the shelf-slope system of the WAP area has fundamen- tally different characteristics compared to pelagic areas of the Antarctic marine ecosystem (Smith et al. 1995a,1996a). Regional coverage from the annual PAL cruises shows that the on/offshore gradient in pigment biomass is an enduring characteristic (El- Sayed 1968), with nearshore areas roughly four times higher than offshore areas. Further, there is some evidence for an alongshore gradient, perhaps associated with the timing of the seasonal alongshore retreat of sea ice &/or latitudinal atmos- pheric & oceanic influences. Estimates of total primary production for the Southern Ocean & the PAL area have been computed using both a light-chlorophyll production (LCP) (Smith et al. 1987; Bidigare et al. 1992; Morel & Berthon 1989; Byers et al., submitted) & an ice edge production model (Smith & Nelson 1986; Wilson et al. 1986; Smith et al. 1988). Input parameters (chlorophyll concentrations, total pho- tosynthetically available radiation (PAR) & sea ice concentra- tions) for these models were derived from satellite data. Chlorophyll concentrations & PAR, in addition to hydrodynamic conditions, have also been determined from shipboard & Palmer Station observations during the PAL field seasons. Comparisons of the spatial & temporal variability of model results with both historical shipboard & our field observations (Smith et al. 1995b,1996a) are consistent. A sensitivity analysis of the PAL area shows that the annual retreat of the MIZ is likely to con- tribute between 5 & 40% to the productivity of this CCSZ/SIZ area. However, comparison of field & model results, in addition to direct measurement of photosynthesis-irradiance parameters, suggests that considerable phytoplankton productivity within the Southern Ocean takes place under light-saturated conditions. Under these conditions LCP models for estimating productivity will need to be revised &/or replaced with models that more accu- rately reflect the physiological conditions of phytoplankton within Antarctic waters. In order to document the temporal & spatial variability in phyto- plankton abundance & distribution within the PAL area & to under- stand the underlying controlling processes, we will concentrate on (1) processes associated with ice edge & shelf-related blooms (phytoplankton accumulation) & their regulation, (2) on/offshore processes which are responsible for the observed patterns on the continental shelf, (3) the balance between phytoplankton growth vs loss rates (respiration, sedimentation & grazing) on the ob- served patterns of phytoplankton distribution,& (4) the relative contribution of sea ice algae production to overall production within the PAL area. Time series data of primary production are virtually nonexistent for the Southern Ocean, so we expect to use our field data, as well as laboratory experimental manipulation, to test our general hypotheses (Figs. 8, 9 & 10) linking sea ice, microalgae & krill. The role of the ocean as a reservoir in the global carbon cycle is dependent largely upon the export flux of planktonic primary production from the euphotic zone (Eppley & Peterson 1979; Willi- ams & von Bodungen 1989). This supply of reduced carbon & energy to intermediate ocean depths occurs by downward advection & dif- fusion of dissolved organic matter (Toggweiler 1989), gravita- tional settling of particulate matter (McCave 1975) & by the vertical migrations of pelagic animals (Longhurst & Harrison 1989) & phytoplankton (Villareal et al. 1993). Each of these in- dividual processes, collectively termed the "biological pump" (Volk and Hoffert 1985) is controlled by a distinct set of en- vironmental factors, & therefore, the relative contribution of each process may be expected to vary with changes in habitat, wa- ter depth &, in polar regions, sea ice (Figs. 8&10). Results from broad-scale, cross-ecosystem analyses suggest that the export of living & non-living materials from the euphotic zone is a positive, non-linear function of total integrated pri- mary production (Suess 1980; Pace et al. 1987; Martin et al. 1987; Wassman 1990), with values ranging from less than 10% in oligotrophic waters to greater than 50% in productive coastal re- gions. It should be emphasized, however, that the field data from which the existing export production models were derived are limited & that open ocean & Antarctic habitats are both under- represented. A majority of the Southern Ocean is characterized by high surface nutrient concentrations but low rates of primary production & export production (Holm-Hansen et al. 1977; Honjo 1990). Broecker (1982) suggested that if all the surface nu- trients in the Southern Ocean were efficiently used by the phyto- plankton, the biological pump activity could transfer a signifi- cant amount of atmospheric CO2 to the ocean's interior. However, at the present time the biological pump appears to be functioning at less than full capacity in polar environments (Knox & McElroy 1984), & a resolution of this enigma is of great importance in the study of oceanic carbon cycles (e.g. Broecker 1991; Longhurst 1991). The intensive austral spring/summer phytoplankton bloom in the Antarctic Peninsula region can result in phytoplankton standing stocks in excess of 20 mg chl a m-3 & sustained production rates of 2-5 g C m-2 d-1 (Holm-Hansen & Mitchell 1991; Vernet et al. 1995). Though the phytoplankton bloom productivity is well- documented, we know less about the fate of this seasonally- accumulated organic matter. During the past decade, numerous field experiments using particle interceptor traps (i.e., sedi- ment traps) were conducted to determine the amount & nature of the particulate matter exported from representative marine and freshwater ecosystems (Honjo 1990). A fairly extensive particle flux data base exists for the Antarctic Peninsula region (Karl et al. 1994b,1996), & sediment-trap derived particle flux estimates in the WAP region are among both the highest (>1 g C m-2 d-1 dur- ing & immediately following the spring bloom) & the lowest (<0.05 mg C m-2 d-1 during austral winter) values reported for the world ocean. Furthermore, during the initial stages of the spring- summer phytoplankton bloom, particle export may be decoupled from contemporaneous new production for periods of several weeks or more. In addition, the abrupt termination of the spring bloom may be controlled, at least in part, by storm events which have the net effect of dissipating the phytoplankton crop & reducing net photosynthesis & growth. If the water column is allowed to re-stratify, a second spring bloom is possible. In order to model carbon & energy flows in the Antarctic coastal ecosystem, we require specific additional information on the residence time of particulate organic matter in the euphotic zone & on the processes responsible for controlling the rates of export. The low to undetectable fluxes of total mass & biogenic matter consistently observed during austral winter periods are enigmat- ic. Although the Antarctic winter is known for harsh environmen- tal conditions of deep-mixed layers & low solar radiation, most oceanic regions in the Antarctic Peninsula area support measur- able net primary production year round. Cochlan et al. (1993a) measured chl-a concentrations of 4 mg m-2 (0-75 m) & photosyn- thetic rates of 35-50 mg C m-2 d-1 in Gerlache Strait during July 1992. Furthermore, the reported f-ratio of the wintertime phyto- plankton assemblage (e.g., [NO3-] uptake rate divided by the sum of [NO3- & NH4+ & urea] uptake rates) was 0.87 suggesting that the majority of primary production was supported by "new" nitro- gen (Cochlan et al. 1993b). In spite of these substantial new & total production rates, the measured wintertime export fluxes in Antarctic coastal waters are <0.5 mg C m-2 d-1. Obviously the time scale for integration & reliable comparison of new & export production rates in high latitude, seasonally-phased environments such as the PAL region must be longer than those used previously in temperate habitats that are sometimes assumed to be in steady-state biogeochemical balance. Given the large interannual variability of ice dynamics & the consequent well-documented ef- fects on primary production & particle export processes, the proper time scale may need to be extended to several years, or perhaps longer, to achieve an acceptable balance. Microbial Processes Microorganisms, including unicellular algae, bacteria, viruses, protozoans & small metazoans, are vital com- ponents of Southern Ocean ecosystems (Karl 1993a). They are largely responsible for the production & decomposition of organic matter, for the primary uptake & regeneration of inorganic nu- trients & for export of carbon & energy to intermediate ocean depths. Furthermore, microbial growth & metabolism can have a profound effect on seawater pH & redox state & can influence the distribution, speciation & availability of certain elements & compounds. Consequently, field data on individual groups of mi- croorganisms & on the complex interactions among them are neces- sary for complete assessment of the role of marine microorganisms in both local & global environments. The microbial loop plays an important role in marine & freshwater plankton ecosystems in all climatic zones (Hobbie 1994). Howev- er, models of microbial ecosystem dynamics continue to elude a general scientific understanding. The marine bacterioplankton are a mixed community, both physiologically & taxonomically diverse, & substantially uncharacterized (DeLong et al. 1994). Yet there are characteristics that appear ubiquitous. Uptake of organic monomers like glucose or leucine, & the activities of en- zymes such as leucine aminopeptidase & B-glucosidase, are common metabolic features. Abundance (biomass) & growth rate (produc- tion) are among the most widely measured parameters in aquatic microbial ecology & the collective data sets of measurements have provided the necessary guidance for the design of the next gen- eration of field experiments. A major obstacle, however, is our inability to characterize much of the dissolved organic matter (DOM) pool in the ocean & to determine the rates at which its components are utilized by bacteria. This will be a major goal of our future field research effort. DOM concentrations in the Antarctic Peninsula region include some of the lowest as well as the highest concentrations ever measured (Karl et al. 1996). There has been a tendency to focus on biomass & production so there is much less information on the kinds of substrates that sustain natural populations in the ocean. Our inability to characterize much of the DOM in the ocean has en- couraged a "stoichiometric" as opposed to a "biochemical" ap- proach. Organisms are viewed in terms of C, N, & P rather than protein, carbohydrate, etc., although different compounds con- taining similar amounts of each element are unlikely to be equivalent to the organism. The biochemical compounds utilized by the bacterioplankton may differ substantially among ocean basins & regions. Identifying oceanographic biogeochemical provinces within which consistent patterns of substrate utilization occur is an important step towards incorporating the bacteria into eco- logical models & defining appropriate experimental protocols for routine use in the field. Growth of the heterotrophic bacterioplankton community is far more difficult to parameterize in ecological models than than that of phytoplankton or microzooplankton. A functional response based on the Monod (1942) equation has been widely used to simu- late nutrient limitation in plankton populations. Although there is uncertainty about the exact form of the equation & the values of the coefficients, the independent variable is far more easily identified for phytoplankton (nitrate &/or ammonium) or protozoo- plankton (prey abundance) than for bacteria (Fasham et al. 1990; Fasham 1993). A more comprehensive description of DOM, including identification of sources & sinks as well as chemical composi- tion, will provide critical data for understanding the role of microbial food webs in the PAL region. Macrozooplankton Ecology Antarctic krill is a keystone species within the Antarctic marine ecosystem & a critical link between primary producers & many top predators. While widely studied, the hypothesized linkages (Fig. 8) have yet to be verified. Processes that control recruitment & distribution as well as in- teractions with food (phytoplankton) & predators (Adelie penguins) need to be better understood. Our conceptual view of these processes has improved significantly but has become more complex in light of the results from the last several years (Quetin et al. 1994, 1996). Adult krill survive winter by a combined strategy of lowered metabolic rates & stored food reserves (Quetin & Ross 1989), but AC0s spawned in summer have not yet accumulated the lipid reserves necessary to survive the austral winter without eating. Winter-over survival of AC0s depends on finding a food source other than open water phytoplankton. Ice algae, either in the sea ice or in the water column after winter storms have broken up the ice & released the cells, is a logical source (Figs. 8a&b). The presence of fall water column blooms will also impact survival by delaying the onset & thus the duration of near starvation condi- tions in the water column. Physiological condition in AC0s de- creases progressively with starvation (i.e., lower condition fac- tor, lower lipid & zero or negative growth rates) & has been do- cumented to be lower during winters of low versus high ice extent (Ross & Quetin 1991c), supporting the premise that AC0s benefit from ice biota. Condition factor for AC0s emerging from winter will reflect the sum of nutritional conditions throughout the winter & early spring. Poor condition factor & poor recruitment are probably correlated, as has been shown for fish, thus condi- tion factor can be an index of recruitment, which in turn can be related to seasonal sea ice dynamics. However, recruitment success is a function of both AC0 survival & reproductive output of adult krill, combined with estimates of spawning stock to estimate population fecundity. Results of several indices (total egg number & volume, average egg size, & timing & frequency of spawning) show that reproductive output of individual Antarctic krill is high after an above mean spring ice season (Stammerjohn, pers. comm.) combined with high phytoplank- ton biomass & production in the subsequent January (Smith et al. 1995a,b), & is low after a below mean spring ice season, which also had low phytoplankton biomass & production in the subsequent January. These preliminary results support the original hy- pothesis that reproductive output in krill depends on food avai- lability, with ice edge blooms postulated as an important food source in spring, which emphasizes the importance of the timing of ice retreat, but also suggests that the role of summer food availability cannot be ignored. Another factor in reproductive output is spawning stock. Siegel & Loeb (1995) suggest that recruitment success is not correlated to previous summer krill stocks. This lack of correlation may be because either (1) variation in stock size of larger reproducing krill is lower than for individual year classes, (2) the relevant parameter is the proportion reproducing not the absolute abun- dance (Fig. 11), or (3) interannual variability exists in the alongshore distribution of spawning populations within the PAL grid. Spawning aggregations are consistently found in the south- ern region off Adelaide I. but not near Anvers I., suggesting that the predictable supply of larvae is from females in those regions that are always covered with sea ice in the early spring (Stammerjohn 1993; Stammerjohn & Smith 1996). Interannual variability in the composition & abundance of the zooplankton community inhabiting the PAL study area, which in- cludes the foraging range of the Adelie penguins nesting near Palmer Station, is driven primarily by variations in oceanic cir- culation & seasonal sea ice dynamics that alter distribution of the zooplanktonic assemblages with respect to fixed geographical locations. Diets of Adelie penguins from Palmer Station are gen- erally dominated by krill > 35 mm in total length (Fraser & oth- ers in 1992-94 AMLR Field Season Reports), which are age class 3 or older, although other prey items are also taken in times of low krill abundance. For example, Thysanoessa sp., a small eu- phausiid, was found in diets (Fraser et al. in Rosenberg 1995 AMLR Field Season Reports) when krill were scarce (Quetin et al. 1995a). In addition, the size distribution of krill within the foraging range of Adelie penguins is a function of both recruit- ment variability, which affects the length frequency pattern for the entire krill population, & the on/offshore gradient, where smaller krill are onshore & larger krill are offshore. This gra- dient is an annual summer phenomenon, but may shift on/offshore with respect to specific geographical features, altering whether the size distribution of krill within the presumed foraging area matches that of the entire population (Fig. 12). The grazing impact of krill is proportional to phytoplankton pro- duction & grazer abundance & composition (Ross et al. 1995c). However, high grazing rates also mean an increase in ammonium ex- cretion rates (Ikeda & Dixon 1984) which may affect nutrient up- take & the coupling between the grazer & primary production com- munities. In a detailed study on the interaction between phyto- plankton community composition & grazing, Haberman (Haberman et al. 1993; Haberman pers. comm.) found that grazing rates of krill depend on the physiological state & morphology of the phyto- plankter. These differences will affect both the impact of the grazers on the phytoplankton community, & also the usefulness of that community as a food source to secondary producers. However, analysis of a seasonal series of instantaneous growth rate (IGR) experiments with AC0s collected from the water column in 1991-92 suggests that growth rates increase with total phytoplankton biomass (integrated over the top 30 m) up to about 55 mg chl-a m-2 (Fig. 13). Growth rates inshore will be higher than those offshore because of the higher phytoplankton biomass inshore. Seabird Ecology A key assumption guiding the PAL seabird research is that the persistence of any seabird population over time re- flects the coincident availability of suitable nesting & foraging habitats (Fraser & Trivelpiece 1996a). Implied, is that interac- tions between physical & biological processes ultimately drive both the magnitude & direction of change in seabird populations. Although considerable progress has been made during the last de- cade towards identifying some of the physical & biological vari- ables that may be important determinants of change in Southern Ocean seabird populations (Croxall et al. 1988; Fraser et al. 1992; Fraser & Patterson 1996), interactions between these vari- ables are just becoming apparent (Trivelpiece & Trivelpiece, sub- mitted; Trivelpiece et al., submitted). A particularly crucial area of research is understanding how the physical environment influences the abundance & distribution of prey on which these predators depend (Croxall 1992). Understand- ing these processes may be viewed as one of the fundamental ob- jectives of the PAL because it requires an integrated link between components & the basis for designing & testing related hypotheses. Seabirds are long-lived upper-trophic level predators which can integrate the effects of variability of the physical & biological environment over large spatial & temporal scales. The expression of this variability can, for example, be measured an- nually as changes in breeding success (Croxall et al. 1988; Triv- elpiece et al. 1990), or over the course of decades & centuries as changes in populations & biogeography (Fraser et al. 1992; Emslie 1995; Fraser & Patterson 1996; Fraser & Trivelpiece 1996a; Trivelpiece & Fraser, 1996). The factors that affect seabird po- pulations at smaller scales can thus provide the basis for under- standing ecological processes in the marine environment at larger scales. Adelie penguins, the predator selected as representative species of the PAL (Smith et al. 1995), exhibit high, sometimes extraor- dinary interannual variability in breeding success, overwinter survival & other aspects of biology (Fraser et al. AMLR Reports 1988-1994; Trivelpiece et al. 1990a,b; Fraser et al. 1992a). Although some of this variability is due to factors that are in- dependent of the marine environment (Fraser & Patterson 1996), it is clear that the dominant signal is mediated by the marine sys- tem, with sea ice playing a pivotal role because of its signifi- cance in the life history strategies of prey &/or their predators (Daly, 1990; Ross & Quetin 1991a; Fraser et al. 1992a; Fraser & Trivelpiece 1995b,1996a; Trivelpiece & Fraser 1996). Our research during the last six years has thus focused on understanding how variability in annual seal ice conditions (extent & duration) af- fects the prey field, & is in turn manifested by apex predators. To address this focus, recent & long-term data (20 years) on the diets of Adelie penguins were combined to test whether (1) krill recruitment is linked to variability in sea ice cover (Table 13), (2) the recruitment signal is apparent as a change in the size- frequency distributions of krill obtained annually by Adelie penguins &, (3) foraging trip durations, a sensitive indicator of krill availability (CCAMLR 1992), increase or decrease in accor- dance with successive years of low & high krill recruitment, respectively. This synthesis was motivated by a recent model (Priddle et al. 1988) which suggests that low recruitment years skew krill size-frequency distributions towards the larger size classes while high recruitment years have the opposite effect. In the PAL study region, maxima in sea ice cover (i.e. the overwintering habitat of larval krill, Ross & Quetin 1991a) occur every 5-7 years (Fig. 3; Fraser et al. 1992a; Stammerjohn 1993; Stammerjohn & Smith 1996). This periodicity could determine the structure & abundance of krill populations on a regional scale. This would suggest that the parameters that ultimately determine predator population responses, such as the combined effects of breeding success & survival on recruitment, may be similarly punctuated by quasi-predictable highs & lows that follow, after an appropriate time lag, the ice-induced krill recruitment signal (Fraser & Patterson 1996c; Fraser & Trivelpiece 1995d; Trivel- piece & Trivelpiece, submitted; Trivelpiece et al., submitted). As shown in Table 13, ice cycles start & end with populations of large krill, the spawning stock (Ice Events 1 & 5); recruitment follows winters of ice maxima, creating a super cohort (Ice Event 2); this cohort maintains its identity because of missing age classes that result from sea ice minima & failed or poor recruit- ment (Ice Events 3,4,5). This pattern also implies that within a cycle, the krill abundance minimum & maximum should occur during Ice Events 1 & 2, respectively (i.e minimum & maximum representa- tion of age classes), as suggested by the Priddle et al. (1988) model. A 17-year record of krill stock estimates (cf. Siegel & Loeb 1995) agrees with this prediction, showing that krill abun- dance extremes within a cycle are specifically linked to Ice Events 1 & 2 (Fig. 14). Equally important in terms of understand- ing some of the mechanisms by which signal transfer occurs between variability in ice cover & the response of upper-trophic level predators, Adelie penguin foraging trip durations exhibit an inverse relationship with krill abundance (Fig. 14, the 1989- 1993 series of ice events). These results provide the first evidence of a direct, causal re- lationship between variability in sea ice cover, krill recruit- ment, prey availability & predator foraging behavior in Southern Ocean ecosystem studies. Given the potential consequences that interannual variability in foraging success may have on predator breeding success & survival (cf. Croxall et al. 1988), they also support the earlier expressed idea that predator recruitment pat- terns may indeed be punctuated by highs & lows that are ultimate- ly mediated by the periodicity, duration & extent of sea ice cov- er. This hypothesis underpins the conceptual framework of a re- cently proposed model that describes how climate change may af- fect apex predator populations over multiple time & space scales, thus ultimately affecting biogeography & community structure (Fraser & Trivelpiece 1996a). To better address the implications of this model, the focus of future research includes: (1) the ef- fects of local nearshore processes such as the timing & duration of phytoplankton blooms on the foraging ecology of Adelie penguins (foraging trip durations & the species composition & characteristics of the diet), (2) the effects of regional charac- teristics in the duration, extent & periodicity of sea ice cover on Adelie penguin survival & breeding success (recruitment), & (3) the effects of within & between-year differences in precipi- tation on the availability of Adelie penguin nesting habitat. Hydrography/Circulation & Biological Modeling Studies To under- stand the processes responsible for the observed water mass & circulation distributions in the WAP region, Princeton Ocean Gen- eral Circulation Model (POGCM) was adapted. The POGCM is a publi- cally available, three-dimensional, time dependent, stratified fluid, hydrostatic, finite difference, primitive equation ocean circulation model (Blumberg & Mellor 1987). The model dependent variables are three components of fluid velocity, temperature & salinity which are known on a vertically & horizontally staggered three dimensional grid. The density & pressure are calculated from depth, & temperature & salinity are determined using a non- linear equation of state (Mellor 1991) & the hydrostatic assump- tion. The vertical dimension is scaled by the local water depth (sigma transformation) which allows bottom topography to be correctly represented, & grid points are distributed in the vert- ical to resolve surface & bottom boundary layers. The horizontal grid is specified as a general curvilinear coordinate system (e.g., spherical coordinates), & variables are distributed ac- cording to the Arakawa C scheme (Mesinger & Arakawa 1976). A second moment turbulence closure submodel is used to estimate vertical mixing coefficients (Mellor & Yamada 1982) & can thereby represent time variable surface mixed layers. Surface forcing is specified as fluxes of heat, salt & momentum. Heat flux can be directly specified or calculated as part of a complete surface radiation budget. Both bottom drag & lateral diffusion provide dissipation in the model. For efficiency, time integration is ac- complished by a splitting scheme where the vertically integrated (shallow water) dynamics proceed with a shorter time step while internal dynamics & mass redistribution occur with a longer step. Means are taken to avoid time step constraints associated with small vertical grid spacing & large diffusion coefficients. The circulation simulations done with POGCM have focused on the time-dependence & thermodynamics of the mixed layer & the larger scale flow over the continental shelf. These latter simulations have considered the role of surface forcing versus offshore forc- ing due to the Antarctic Circumpolar Current (ACC) in structuring the flow over the shelf. The mixed layer simulations have indi- cated that double diffusive processes are likely important in determining the amount of exchange between CDW & the upper water column. The biological modeling aspect of the PAL program has focused on developing a time-dependent size-structured model of the growth & dynamics of Antarctic krill. This model represents the life his- tory of krill, between 10 & 60 mm in length, in terms of 300 energy-based (i.e., calories) size classes. Growth (or shrinkage) processes, which result from positive (negative) net production, result in the transfer of individuals between size classes. Net production is determined from the difference between assimilation & respiration. Environmental conditions (e.g., temperature) & food availability (phytoplankton composition) are input to the model as prescribed idealized time series or as time series con- structed from measurements. Preliminary results with the size- based krill model indicate that the size composition of the available food (i.e., the phytoplankton population structure) is important in the survival of krill larvae, especially in the winter. Moreover, seasonal variations appear to exist in the physiological processes that contribute to krill growth. Our efforts in the next phase of this project will consist of data analysis & modeling components. We will continue analysis of hydrographic data collected on PAL annual & process cruises. In particular, we will investigate interannual variations in the distribution of CDW. Mixing of CDW results in reduction of the oxygen content of the overlying waters by 25 to 45%, which sug- gests an average annual entrainment rate for the WAP of 0.7 to 1.43 x 10-6 m s-1. The freshwater input needed to balance the sa- linity input from CDW is on the order of 0.63 m y-1, which can be supplied by local precipitation & advection of ice into the re- gion from the Bellingshausen Sea, which then melts. The annual heat flux associated with CDW is 12 W m-2, which is sufficient to melt this amount of ice. Hence, CDW is an important factor in determining the amount & extent of sea ice in the WAP region, & understanding of how this water mass varies is critical to under- standing variations in ice cover. We will also continue our analysis of hydrographic data in terms of attempting to determine interactions/correlations between the physical environment & bio- logical structures. Our modeling efforts will be directed towards continued develop- ment of the circulation model & in particular we will develop sea ice & mixed layer dynamics models that can be interfaced with the circulation model. We currently have several different mixed layer models operational & are evaluating these for use with the POGCM & for application to the WAP continental shelf region. Similar preliminary calculations are underway with sea ice models. Simulations with this combined sea ice-mixed layer- circulation model will be done to investigate processes control- ling circulation & sea ice distributions in the WAP region. In terms of interdisciplinary modeling, the size-structured krill model will be embedded in the circulation model. This combined model will be used to investigate residence times, exchanges & general transport pathways for krill in the WAP region. The results of these simulations will be used to interpret/explain variations in penguin recruitment and abundance, as well as the large-scale krill distributions. Our final modeling effort will be to embed a primary production model in the mixed layer & cir- culation model. This combined model will be used to address questions related to the role of physical and biological processes in controlling the observed distribution of phytoplank- ton in the WAP continental shelf area. We have already made some progress in determining the correspondence between physical vari- ables, such as mixed layer depth & water column stability, & phy- toplankton biomass. 2.5 Science Plan & Methods Sampling Strategy The harsh & logistically difficult working environment of the Southern Ocean, in addition to the extreme variability of physi- cal, chemical, optical & biological processes, dictates a varied & flexible sampling strategy (Smith et al. 1987a & Fig. 15). Ships, the classical oceanographic sampling platform, can provide relatively accurate point surface measurements of a wide variety of desired variables, in addition to vertical measurements of the water column. Ships can also be used to obtain 'along-track' samples & observations. The disadvantage of ships is their lim- ited spatial coverage & working constraints during foul weather. Moored buoys are even more limited in spatial coverage but are extremely valuable in providing long time series at selected lo- cations. Satellite remote sensing is often the most effective way (& in some cases, the only way) to study large-scale surface phy- sical & biological processes in polar regions (Comiso 1995). >From the perspective of long-term observations, satellite sam- pling is spatially & temporally comprehensive, & the measurements operationally consistent. Remote sensing by aircraft fills a gap between conventional surface measurements & those made by satel- lite sensors. In the PAL region, with high level clouds typical- ly 90% of the time, aircraft also provide a unique opportunity (by flying under the generally high cloud cover) to utilize ocean color methods (visible portion of the spectrum) for estimating near surface pigment biomass under conditions which would limit satellite ocean color observations. To date, the PAL have used data from satellites, ships, small boats (zodiacs), moorings both at sea & on land (AWS units), & land-based observations at the seabird nesting sites. The proposed collaboration with the Brit- ish Antarctic Survey (see Sect. 2.7) is expected to add remote sensing using aircraft during our next funding period. We are also following the technological progress of autonomous vehicles, both airborn & underwater, & small seabird transmitters so as to incorporate these advances in our sampling strategy when this technology becomes economically cost effective. Our science plan addresses both long-term & short-term processes & encompasses sampling flexibility while maintaining long-term consistency with our first funding period. The general plan maintains our regional scale summer sampling program of hydro- graphic, optical & biological sampling within the 200 to 600 line region of the PAL grid (Fig. 1a) & the high density sampling of phytoplankton, krill, seabird observations within the Adelie foraging area (Fig. 1b). While maintaining this unique & impor- tant long-term data set, we anticipate increased flexibility by less intensive sampling within some areas & periods, in order to increase attention to sea ice related linkages. Table 14 presents our proposed ship schedule. Austral summer cruises are timed to match the critical period of Adelie penguin breeding (Fig. 7) in order to investigate trophic level linkages. PAL January cruises typically have two sampling modes: (1) large scale cardinal grid sampling including stations "inside" the is- lands, & (2) finer scale sampling within the Adelie foraging area. The first mode determines key environmental variables sam- pled on a fixed grid which facilitates separating long-term sys- tematic trends from interannual variability. The second mode is aimed at observations linking water column variables with fine scale (few km) krill & seabird observations. Tables 1 & 2 list annual cruises to date, & brief cruise reports are summarized in Antarctic Journal of United States (AJUS) articles (references in Sect. 1.3.2). Samples & experiments conducted during the mesos- cale cruises help to address several of our general hypotheses, including variations in: (1) oceanic & atmospheric processes which influence the extent of CDW upwelling onto the continental shelf (HA1), (2) primary production & physical processes in- fluencing the abundance & distribution of phytoplankton (HA2,3,4), (3) recruitment, growth rates, & reproductive output of krill, in addition to other zooplankton assemblages present, as well as the distribution, size, density, & depth of krill aggregations & length frequency distributions (HA3,4). The finer scale grid during these cruises allows us to investigate these same hypotheses with greater resolution within the penguin forag- ing area during the critical period of chick rearing, & provides prey population information which can be compared to data col- lected from the rookeries on foraging duration, diet samples, & reproductive success of penguins. Table 14 also shows that in two of the next six field seasons, sea ice process cruises are planned (dependent upon predicted sea ice conditions) during spring retreat (Sep/Oct/Nov97) & fall/winter growth (Apr/May/June00) periods. In addition, it is anticipated that increased flexibility during summer cruises will allow some time to sample in the southern end of the PAL grid at the ice edge. The ship time at the ice edge & in the ice will be designed to address short-term mechanistic processes & hypotheses linking sea ice, microalgae, krill, penguin & export processes (Figs. 8,9 & 10) within & outside of areas occupied by popula- tions of juvenile Adelie penguins dwelling on the pack ice. For example, Figs. 8 are conceptualized diagrams showing the annual cycle of hypothesized linkages between sea ice, ice algae, water column phytoplankton, krill & the export of carbon from surface layers, which we hope to elucidate from the two ice process cruises planned in fall/winter & spring (Table 14). Fall & early winter are the periods when pancake ice forms & aggregates into a continuous, consolidated ice cover (Lange et al. 1989). Biologi- cal material is incorporated into pancake ice when wind, wave ac- tion & frazil ice formatin give rise to the accumulation of suspended biological material into circular discs of ice. This is also a transition time for AC0s as they move from a strong diel vertical migration through the water column to a close asso- ciation with under ice surfaces as found later in the winter. The timing of fall blooms, early ice advance & the linkage with AC0s has been little studied in the WAP area &, as hypothesized, this period may be of vital significance in AC0 survival under certain conditions. Key objectives of this fall ice process cruise will be the investigation & understanding of sea ice growth processes & its related biota during that period. In ad- dition, an effort will be made to determine where Adelie penguins are located during this time. Sea ice conditions in spring & early summer are hypothesized to be characterized by the melting cycle where there is a return of particulate organic material (POM) back into the water column ei- ther as a seed for phytoplankton bloom or enhanced sedimentation (Figs. 8). The retreating pack ice is expected to have primarily AC0s under the ice, with progressively older stages of krill as- sociated with the MIZ &/or ice edge bloom (Quetin et al. 1992). Investigations during this spring cruise would include: (1) the degree of coupling of AC0 & adult krill with ice & the associated physical characteristics, (2) linkages between processes associ- ated with krill, ice algae, nutrients, gases, bacteria & particle flux, (3) quantification of the relative contribution of ice- related production compared with production driven by non-ice processes, & (4) relationship to MIZ & the first critical period for breeding Adelies (Fig. 7). These ice process cruises would work from the open sea, to within the MIZ & then penetrate into the pack ice to a distance sufficient to be beyond significant surface wave influence (or as far as the ship can safely & use- fully operate). Another significant element of our sampling strategy is directed toward time series data taken in the vicinity of the Adelie nest- ing sites near Palmer Station. These data document seasonal pro- gression & allow both the regional cruise & shorebased seabird data to be placed into a more comprehensive seasonal & interannu- al perspective. The field season at Palmer Station also permits longer time scale mechanistic studies, which are impractical & costly to conduct on ship. Finally, the Palmer Station data pro- vide surface validation for both satellite & aircraft remotely sensed observations. Continued Palmer field work will be con- sistent with our previous time series observations but with in- creased emphasis on process oriented studies in the local area. We anticipate, among all PAL components, 10 personnel on station throughout each field season (Oct-Mar). An important contribution to our time series is the data from the sediment traps. One trap will be maintained at the Hugo I. site which already has 3 years of continuous data. A second mooring is located near a recently moored site (Domack, personal communi- cation & see Sect. 2.7) in Palmer Basin (64 degrees 50.11'S 64 degrees 08.36'W), & one or two more sites will be added in this Basin along the "hypothesized" penguin foraging axis. It is also planned to add both temperature sensors & current meters on these deep mooring so as to better understand the seasonal changes in local water masses & current movements in this area. Thus, the entire suite of data from cruise, mooring & Palmer Station field observations helps elucidate some of the mechanisms underlying the space & time variability of trophic interactions. Methods Space limitations do not permit a detailed discussion of methods which are outlined within the tables listing PAL data (Tables 3,4,5,6) under the heading "Analytic Standard & QA/QC". Quality Assurance / Quality Control is discussed in detail in the follow- ing section (2.6). Typically our methods follow recognized stan- dard procedures often using state-of-the-art equipment. For ex- ample, the Bio-Optical Profiling System (BOPS-II) is a recently modified instrument that allows real time readout of water column properties (listed bio-optics, 'BO', in Table 3) as well as pro- viding a 12-bottle rosette for discrete water sampling of the wa- ter column based upon real time readout. This is a robust sea- going instrument that over the first six years of the program has been used for nearly 2000 successful casts. Krill methods in- clude both acoustical techniques & nets of several sizes to docu- ment temporal & spatial variability in secondary producer abun- dance & distribution. Each have their intrinsic errors, but are equally valid (Everson 1988) & yield different information. On cruises, oblique net tows with simultaneous acoustic transects about 2 km in length are centered at each station. This approach allows local & regional variability in acoustic biomass & krill aggregation characteristics & distribution to be correlated with other habitat characteristics. Identical or very similar methods are used for sampling & analysis of seasonal data which is col- lected from zodiacs, & for the fine scale grids to study the re- lationships between krill & Adelie penguins within the foraging area. Penguin observations follow the CEMP standard methods (the 'CEMP protocols', CCAMLR 1992; CCAMLR - Commission for the Con- servation of Antarctic Marine Living Resources, CEMP - CCAMLR Ecosystem Monitoring Program). Seabird data collection at sea follows methods summarized in Spear et al. 1992). 2.6 Core Measurements & Quality Assurance / Quality Control Pro- cedures There are both scientific & logistical considerations involved with the establishment of any long-term, time-series measurement program. Foremost among these are site selection, choice of vari- ables to be measured & general sampling design, including sam- pling frequency. Equally important design considerations are those dealing with the choice of analytical methods for a given candidate variable, especially an assessment of the desired accu- racy & precision, availability of suitable reference materials, & the hierarchy of sampling replication & mesoscale horizontal variability. All LTER programs include a core suite of environ- mental variables to track both physical & biological processes in the habitat of interest. For the PAL program we selected parame- ters that might be expected to display detectable change on time scales of days to a few decades. The PAL sampling design includes (1) continuous & discrete meas- urement of key environmental parameters, (2) collection of zoo- plankton & fish using nets/trawls, (3) measurements/observations of seabird distribution, abundance & biomass, & (4) satellite- based remote sensing observations (Tables 3,4,5,6). During each annual cruise water samples are routinely collected generally from the surface to ~200 m for the measurement of a variety of chemical & biological variables. At selected stations, samples are collected over the full depth of the water column. To the ex- tent possible, we collect samples for complementary biogeochemi- cal measurements from the same or from contiguous casts to minim- ize aliasing caused by time-dependent changes in the density field. This is especially important for samples collected in the upper 200 m of the water column. Water samples for salinity determinations are collected from selected water bottles to iden- tify sampling (bottle trip) errors. Approximately 10-20% of the water samples are collected & analyzed in triplicate to assess & track our analytical precision in sample analysis. Our primary study area is characterized by cold (<1degC) surface waters with high nitrate concentrations (>30 MM), seasonally variable surface mixed-layers (10-200 m), & variable standing stocks of living organisms (1-1000 Mg C l-1). Ideally, the suite of measurement parameters should provide a data base to validate existing biogeochemical models & to develop improved ones. Our list of core measurements has evolved since the inception of the program in 1990, & now includes both continuous & discrete physi- cal, biological & chemical ship-based measurements, in situ bio- logical rate experiments, & observations & sample collections from bottom-moored instruments (Tables 3&5). Continuity in the measurement parameters & maintenance of quality, rather than the methods employed, are of greatest interest. Detailed analytical methods are expected to change over time through technical im- provements. The precision & accuracy of each determination is of utmost importance, especially if the program objectives are to assess environmental change. For the PAL program, the precision of most measurements is determined by the collection & analysis of replicate samples on a routine basis. This information is then used as a measure of analytical information which in turn is used as a measure of analytical reproducibility for each of the candi- date variables. The question of measurement precision is more problematic because some of the parameters that we measure on a routine basis (e.g., primary production, bacterial production, bacterial cell number) have no commercially-available reference standards. However the measurement accuracy for most of the ecosystem variables (e.g., oxygen, salinity, nutrients, carbon dioxide, alkalinity, particulate carbon, nitrogen, phosphorus & mass) can be determined using certified (e.g., NBS or NIST) reference materials. The environmental sensors used for the con- tinuous measurement of pressure, temperature & conductivity are routinely calibrated at the Northwest Regional Calibration Facil- ity in Seattle. Our optical sensors are periodically calibrated using both recalibration by the original manufacturer & more fre- quently (pre & post-cruise) at our own optical calibration facil- ity at UCSB using optical standards traceable to NIST. In addi- tion, SeaWiFS related optical instruments have & are cross cali- brated against SeaWiFS instruments of other investigators. Fi- nally, we have recently initiated an interlaboratory exchange program for the measurement of selected variables, & we antici- pate an expansion of these activities during the next phase of our field work. 2.7 Regionalization & Cross-site Efforts British Antarctic Survey (BAS) Research on the nearshore marine ecosystem will be undertaken by the BAS at Rothera Station on Adelaide I. (Fig. 1a) in the 1996/97 season. This will comprise a long-term program (at least ten years) of year-round oceanographic monitoring, together with a series of individual autecological or process studies. It is also intended to fly ocean color sensors from BAS DHC-6 (twin otter) aircraft fitted for remote sensing to provide detailed spatial coverage of Marguerite Bay & the adjacent PAL grid area. This work will be in collaboration with the bio-optics component of the PAL who will be collecting periodic surface bio-optical data in the vicinity of Palmer for validation & algorithm development studies. It is anticipated that aircraft coverage of the PAL grid will provide complementary spatial & temporal cover- age not otherwise available to the PAL. Emphasis on the nearshore (benthic) marine ecosystem will complement the PAL, permit a latitudinal (along the peninsula) comparison with the PAL work, & complement (with aircraft) the PAL ship & station sampling strategies. Antarctic Marine Living Resources (AMLR) The U.S. AMLR program, supported by NOAA, is based at the north- ern tip of the Antarctic Peninsula several hundred kilometers north of the PAL grid in the vicinity of Elephant I. (Fig. 1a). The objective of this long-term study, initiated in the mid 1980's, is to describe the functional relationships between krill, their predators, & key environmental variables. Emphasis is on an ecosystem approach to resource management within the An- tarctic, with particular focus on fisheries impacts on krill & their dependent predator populations. The program includes moni- toring the impacts of the krill fishery in the area of King George & Elephant Islands & has also included collaborative krill-penguin research at Admiralty Bay with W. Trivelpiece. AMLR also funds research of Adelie penguins at Palmer (B. Fraser), which serves as a nonfished control site. The AMLR study pro- vides complementary information for comparison with PAL data in an area with different oceanographic & sea ice regimes. Canadian Space Agency (CSA) Radarsat program Based on results of our PAL sea ice work, R. Smith is now funded under the sponsorship of NASA, to participate in the internation- al research of the CSA Radarsat program & will be obtaining high resolution (100 m) Synthetic Aperture Radar (SAR) data for the WAP area. Our objectives for the SAR data, within the context of the PAL are to: (1) more accurately demarcate sea ice-related ha- bitats of the marine ecosystem of the Southern Ocean, (2) deter- mine the seasonal & interannual space/time variability of these habitats in relationship to atmospheric & oceanographic forcing, (3) relate sea ice variability to subsequent impacts on keystone species within this marine ecosystem, & (4) evaluate parameters derived from SAR data with parameters derived from other satel- lite data (e.g, SSM/I, AVHRR, SeaWiFS) & surface data so as to validate & improve satellite-derived products from the Radarsat data. LTER Cross-site comparison In Jan 1996, D. Karl received funding from the NSF-DEB to conduct an LTER cross site comparison. This project, "Microbial loop dynamics and regulation of bacterial physiology in subtropical and polar marine habitats" will evaluate the structure & function of the microbial food web in two contrasting habitats: the An- tarctic coastal/shelf/oceanic ecosystem & a subtropical oceanic ecosystem in the central North Pacific Ocean. Both habitats are isolated from significant inputs of terrestrially-derived organic matter thereby providing the basis for a comprehensive, compara- tive study of bacterial metabolism of autochthonous organics of marine phytoplankton origin. Strong contrasts in plankton commun- ity parameters (high vs. low inorganic nutrients, large vs. small phytoplankton, metazoan vs. protozoan grazers) make these two sites ideal for this cross-site investigation. The project will be embedded within ongoing programs at the two sites, the U.S. JGOFS Hawaii Ocean Time-series (HOT) in the North Pacific & the PAL in Antarctica, & will build upon the core datasets that are collected. Southern Ocean Ocean Color R. Smith is funded by NASA, "Bio-Optics, Photoecology & Remote Sensing Using SeaWiFS", to develop bio-optical models specifical- ly for the Southern Ocean. These models provide a quantitative representation of the link between in-water biogenic material & ocean optical properties, thus allowing data from optical sensors to be used for remote &/or proxy estimates of important biologi- cal parameters in the ocean, & permitting sampling over a range of space/time scales that otherwise would not be possible (Fig. 15). We anticipate use of optical sensors on moorings, the BAS aircraft & satellite (SeaWiFS, MODIS) which will significantly complement our research objectives throughout the duration of the PAL program. Sediment Core Study In 1995, scientists in the PAL initiated a collaborative program with E. Domack (Hamilton College) & A. Leventer (Univ. Minnesota) to investigate the 200-300 year productivity cycles in the PAL region that have been revealed through a comprehensive analysis of sediment cores collected in the Palmer Basin. Regional trends in climate, including but not limited to warming, ice shelf melt- ing, sea ice dynamics & predator-prey cycles all affect particle composition & fluxes, as well as the long term sediment accumula- tion rates. Since 1992 several gravity cores have been recovered & analyzed by Leventer, Domack & colleagues. Measurements include 14C-chronology, sedimentology & geochemistry, magnetic suscepti- bility & a quantitative description of diatom & foraminifera as- semblages. In Dec 1995, on a PAL cruise, three additional cores were collected, & a permanent sediment trap site was established. We expect the PAL data sets on contemporaneous ecosystem processes to be a great help in resolving the paleoclimate histo- ry of this region. 3. Literature Cited (Note: PAL citations are given in Section 1.3) 4. Site Management The Protocol on Environmental Protection to the Antarctic Treaty, signed in 1991, designates Antarctica as a natural reserve & sets forth requirements for all activities in Antarctica. Subsequently, the Antarctic Treaty Consultative Meeting approved a protected area management plan for "Multiple- Use Planning Area: SW Anvers Island & Vicinity" which includes much of the PAL study area. Treaty nations are to voluntarily follow the guideline for the protection of fauna & flora, while the plan is being rewritten to conform with the new format & guidelines established by the 1991 Madrid Protocols, which in- creased protection of the Antarctic environment. An overview of the Antarctic Treaty System is provided in workshop proceedings (Polar Research Board 1985), & Naveen (1996) has provided a re- cent review with emphasis on the potential adverse effects of hu- man disturbance on the Antarctic environment within the context of the Antarctic Treaty. Of immediate concern to PAL is the ability to guarantee that the site will remain undisturbed by uncontrolled human influences. Naveen (1996) also discusses the Antarctic Site Inventory pro- ject, initiated as a pilot study in 1994 by the U.S. NSF, to determine if periodic site inventories of biological & physical features in areas commonly visited by tourists would provide a way to monitor potential environmental impacts. Within the PAL area, visitors are not permitted to land on most islands with nesting seabirds during the breeding season, & the few sites where visitors are permitted are under study for possible adverse impacts. Also of concern is how far from pristine the environment of the WAP area has become (Palmer LTER Group 1996). In connection with shore-based scientific research stations & the growing tourist industry, Kennicutt & McDonald (1996) summarize & discuss inven- tories of contaminants, contaminant sources, transport & deposi- tional processes, & potential biological impacts in the WAP area. Although there is evidence that organisms have been exposed to contaminants, most events are local (100s of meters) & are con- fined to regions of human activity. Fossil fuel spills from ship traffic pose the greatest risk of future contamination, although the nature & volume of the potential spills would indicate that long-term damage would be minimized (Kennicutt & McDonald 1996). Overall, these authors conclude that the WAP is still relatively pristine. Site management of Palmer Station is carried out by the Antarctic Support Associates (ASA), under contract to the NSF. The Executive Committee consisting of the ten PIs is the primary governing body of the PAL. The committee provides general scien- tific direction & budget guidelines. Issues are decided in the Executive Committee by majority vote. Formal communication is maintained between PIs with a periodic agenda sent by email & two annual meetings. The lead PI (Ray Smith) is the direct adminis- trative contact of the PAL to the LTER Network & the NSF. Karen Baker has been assigned responsibility for data management issues with respect to the LTER Network. Charleen Johnson has been as- signed the responsibility of Single Point of Contact (POC) for PI interaction in coordinating logistic matters with the ASA. The Marine Science Institute at UCSB handles & coordinates contract issues for the PAL. The Institute for Computational Earth System Science at UCSB is the data hub for the PAL. The lead PI coordi- nates the monthly agenda, Executive & Steering Committee meetings & the overall field season, & performs other general administra- tive functions. The research, modeling & data management activi- ties of the PAL are divided into several components with each ad- ministered by one to two PIs. The PIs of each component plan the detailed logistics for field season research & are responsible for collection & publication of specific data sets & entry of data & results into the PAL data base. Field work at Palmer Sta- tion is often the responsibility of technicians or graduate stu- dents in the absence of the PI. Undergraduate student volunteers comprise an important element of the field teams & are of great importance to our success. 5. Data Management PAL data management is based on several distinct concepts: (1) acceptance of a diversity of computer platforms & tools among the various PIs, (2) establishment of a distributed system of commun- ication with effective connectivity among PIs, & (3) development of an electronically available central database which offers con- tinuity, accessibility & extensibility for evolving long-term core data. We follow a decentralized model of data management, where each PI is responsible for a subset of core & non-core data. Documentation & data storage are organized through an elec- tronic hub at the Institute for Computational Earth System Sci- ence (ICESS) at the University of California at Santa Barbara which also serves as a data archive as needed (http://www.icess.ucsb.edu/lter). The data archive structure has been organized to facilitate rapid information exchange & online data documentation while supporting platform independence & low maintenance overhead. As the eighteenth LTER site, we benefit from the collective experience of the other LTER sites (Michener et al. 1994). Information System This section of the database provides general information on the PAL project as well as documentation describing the data taken for each field study (metadata). Information is maintained in flat ascii files which are easily available to all investigators. Quality control for both metadata & field data is the responsi- bility of the individual investigators at their respective insti- tutions. The central data archive is a backup of each individual investigator's dataset and, in addition, is backed up on a regu- lar schedule. In order to organize the metadata, a common vocabulary was developed & documented. For example, 'study' consists of either a field cruise or a season at Palmer Station. Within each study, data sets exist either as part of the pre-defined core data sets or as part of the non-core opportunistic & mechanistic study data sets. The study types & data set definition list is maintained online (Coreform Definition Table). Several documents for each study are standard: (1) an overview of the study, (2) site maps, (3) a participant list describing who was on site for the study, & (4) an event log listing chronologically the type & location of measurements made during the study. The event log provides an in- itial cross index of all component participation for the duration of each study. Efforts to streamline documentation are continual- ly under development (e.g., recent provision of consistently prepared data forms for at sea operations & the encouragement of online procedure manual development for all components). A group calendar, meeting schedule, field documents & the PAL an- notated bibliography are maintained online in addition to all pertinent documents such as articles, abstracts, proposals, & meeting notes. Background material & several summary talks were installed on the web in order to make information more readily available. Communication links with outside groups are also available at our web site, such as information on weather station locations. Several of the cross-site activities including the all-site bibliography, the climate committee reports & the biodiversity committee efforts are co-ordinated by the PAL da- tamanager. For example, a site species list was developed by building upon the National Oceanic Data Center's comprehensive taxonomic list. The list was acquired on CDROM, the PIs were pro- vided with pertinent lists which they updated for the PAL region, & the results were posted at our web site (http://www.icess.ucsb.edu/lter/biodiversity). Investigation of Linnaeus Protist Taxonomic Software is ongoing. Historical & long-term weather data have been compiled by the da- tamanager & are available publically. Data sets from past Antarc- tic projects within the PAL area, such as BIOMASS & RACER, have been made available or referenced through the data base. Data sets such as coastline & bottom topography for the WAP area have been acquired & maintained. In addition, the datamanager is an active participant in obtaining & archiving weather data, as well as attending yearly Automatic Weather Station (AWS) meetings. Last year the AWS meeting was hosted at a PAL home institution. Data from two AWS units, which were installed for the PAL in cooperation with the University of Wisconsin, are monitored daily so that station failures can be rapidly corrected, as was the case in March of 1995 when a battery failure was corrected. Further, the historical Palmer Station weather record has been obtained, quality control of current Palmer station weather data initiated, & historic data validated in part by comparison with other historical records (Baker & Stammerjohn 1995; Smith et al. 1995). Personnel Structure & Relationship to Projects The PAL datamanager works with the PIs to create on-line documen- tation & to manage & archive project data. The datamanager par- ticipates in PAL PI meetings & attends workshops & field planning exercises. In recognition of the significant role data management must play in the development of an LTER project, the PAL da- tamanager was made a member of the PAL Executive Committee (com- posed of all PAL PIs) in 1993 & has a separate component budget in this proposal. Currently, the data management position is funded to develop & maintain the central data structure in coor- dination with each of the individual PIs at their respective in- stitutions. PAL data management is designed to take advantage of the existing strengths of each team member's home group expertise & foster an integration of component data & manage core data. Since each component has unique, independent hardware, a software organization specifically optimized for their own research agenda & each with an independent history of development, local computer development remains the responsibility of individual components. However, our datamanager provides as powerful a connectivity as possible within this diversified computer environment. Thus, there is a standardization of vocabulary & information structure, while the variety of platforms creates an enrichment of options in terms of data analysis & display. Computer system analysts at ICESS have provided input on network- ing, software, hardware & database planning, while making avail- able state-of-the-art computer technology & ideas to address our scientific pursuits. Routine system functions such as data back- ups, gopher & web servers, color products & peripheral interfac- ing are in place. Storage of large datasets such as satellite data is available at ICESS in addition to the disks serving as the initial local data hub. Because the PAL participants reside at different home institu- tions across the country or are conducting research in the An- tarctic either on station or aboard ship, the development of the internet & the recent increase in reliable network software has played a critical role in the flow of data between individuals. Net tools have been adopted quickly & used effectively to make information readily available across all platforms. Internet con- nections for each component & location have been accomplished in a variety of ways, including campus broadband connections, gator- box tunneling across the internet, & dial-up modems. Browsing tools, such as www & gopher, & file transfer tools, such as ftp, are being used effectively. Project discussions are carried on using the internet. Networking all the project participants also facilitates the transfer of both computer knowledge & computer resources which makes a significant contribution toward integrat- ing group ideas & data. The PAL datamanager has played an active role in the LTER data management community, attending the annual datamanager meetings since 1990, as well as other LTER organized meetings such as the all scientist meetings. Recognizing the value of cross-site coor- dination among LTER datamanagers, the PAL datamanager helped plan & coordinate the first extended datamanager meeting in 1994 when speakers & datamanagers from other projects were invited to share information & techniques & also coordinated cross-site informa- tion such as the "Site Capabilities' lists (http://lternet.edu:70/00/doc/SiteCapabilities), which facilitate sharing hardware & software information across all sites. Subse- quently, the PAL datamanager became a member of the LTER DataTask Committee, whose function is to co-ordinate meetings & da- tamanager communications. The PAL datamanager is also a member of the McMurdo Users Advisory Group & the Antarctic Communica- tions & Computers Working Group which provides input on field needs and logistics. Further, consistent co-ordination with ASA for both ship & station equipment & computers is ongoing. Data Availability & Data Policy The intent is to have all data online as rapidly as possible so that investigators can have other component data available while evaluating their own. All PAL investigators have accounts on the ICESS server & are therefore networked to all data & documenta- tion in the PAL central archive. The PAL group developed a data policy in order to define the rules of data sharing & to assure that each contributing scientist has first priority in the use & publication of their own data, while making the data available as rapidly as possible in order to promote rapid data assimilation between groups. Data policies at other data sites were considered (Porter 1993; Porter & Callahan 1993) in developing the PAL data policy (gopher://gopher.icess.ucsb.edu:70/11/datainfo/datamanagement/datapolicy). The PAL data policy states that data be put online by 1.5 years after collection, be available for the next 2 years for internal use with the data collector as sole proprietor if desired, then be available for the 2 years following for collaborative research within the PAL, & then after this period be publically available. All documentation is public & can be requested through the responsible PI, upon being entered in the database either through gopher or the web: gopher://gopher.icess.ucsb.edu http://www.icess.ucsb.edu/lter The data is first available online to the PAL group through their ICESS accounts and then as it is released, the data appear with the documentation on the public web. It is the responsibility of the individual investigator to provide their data & standard da- taset documentation in digital ascii format in the central data- base as soon as it is processed. The database is structured so that these files are immediately visible online but do not have to be overseen by a librarian fluent in html so that the PAL ar- chive is automatically updated whenever an investigator adds to or updates their metadata &/or data files. Future Future work will include continued efforts in facilitating con- nectivity & the submission of dataset forms & datasets. The go- pher connection has proven useful but will be converted to a more transparent web structure. We have been primarily concerned dur- ing our first funding increment with the establishment of data taking, organization, storage, & availability, but it is recog- nized that as the amount of data increases, we need to consider the possibility of a relational database as well as the current cross site analysis tools. The database now is organized in such a way as to facilitate conversion to other structures. The prior- ity will remain to produce a robust and powerful system while re- quiring as little maintenance & support as possible. An emphasis on publically available software will also continue with an eye toward maintaining data convertibility. 6. Outreach Human Resources. Volunteers, undergraduates & graduate students have played an im- portant role in our field & laboratory work throughout our first six years. The quality of our volunteers has been exceptionally high. Because we provide a unique opportunity to participate in Antarctic research, a significant number of our volunteers are mid-career adults who have enriched our program with their own scientific & technical abilities while helping in our field research. This mix of mid-career adults with students has been especially valuable in providing our younger volunteers with a broad spectrum of both research & 'real life' interactions. All volunteers return with a deep appreciation for the importance of scientific research & for Antarctica as a unique environment. Typically each field season includes 6 to 8 volunteers, whose only cost to the program is travel to & from the Antarctic. Undergraduates have been involved both as volunteers & as parti- cipants in NSF's Research Experiences for Undergraduates (REU) program. During three of the five field seasons, PAL PIs served as advisors for REU students who joined research teams in An- tarctica both onboard ship & at Palmer Station & stateside in la- boratories. In 1991-92 7 REU students joined 3 research teams, in 1992-93 11 REU students joined 4 research teams, & in 1994-95 1 REU student joined the PAL field team. The research areas for these REUs included hydrography, ocean optics & remote sensing, primary production & phytoplankton physiology, & secondary pro- duction (pelagic zooplankton & fish). The overall objective of the REU program is to provide an educational experience to ac- quaint students with all aspects of the research process & to en- courage them to continue their education in science. These ex- periences for PAL REU students include: (1) a seminar series, (2) pre-season training in the advisor's laboratory, (3) 10 weeks in Antarctica as an essential member of a research team, & (4) in- dependent research projects involving data analysis & preparation of publications in the home laboratory. PAL REUs at UCSB receive academic credit for independent studies &/or field work in oceanography for their participation. One of the benefits to the student participants is the integrative aspects of the PAL as an interdisciplinary research program. Students evaluate the program at the end of their award, & the majority believe that their ex- perience was beneficial in helping them to find a focus within aquatic science or to decide whether to continue in science as either technicians or graduate students. Graduate students. The following graduate students have received or are receiving full or partial funding from the PAL program. UCSB: - Sharon Stammerjohn, Jan92-Dec93, M.A.: "Spatial & temporal variability in Southern Ocean sea ice coverage" - Tom Frazer, Sep90-Nov95, Ph.D.: "On the ecology of larval krill, Euphausia superba, during winter: krill - sea ice interactions" - Karen Haberman, Sep91-present, Ph.D.: "Grazing by the Antarctic krill, Euphausia superba: Effects of phytoplankton type & food quality on ingestion, assimilation & growth of krill" - Caroline Shaw, Sep95-present, M.A.: "Interannual variations in the ovarian cycle of Euphausia superba west of the Antarctic Peninsula" - Mark Moline, Sep91-present, Ph.D.: "Variability and regulation of coastal Antarctic phytoplankton dynamics on interannual, seasonal and subseasonal time scales (1991-1994)" - Heidi Dierssen, Sep93-present, Ph.D.: "Remote sensing of pigment concentrations and primary productivity in the West Antarctic Peninsula region" - Phil Handley, Sep94-present, M.S.: "Annual and seasonal variability in hydrography near Palmer Station, Antarctica" U of Hawaii: - James Christian,Sep90-Dec95, Ph.D.: "Biochemical mechanisms of bacterial utilization of dissolved & particulate organic matter in the upper ocean" - John Dore, Sep89-May95, Ph.D. - Ricardo Letelier,Sep88-May94, Ph.D. ODU: - Cathy Lascara, Sep91-present, Ph.D.: "Seasonal and mesoscale distribution of Antarctic krill, Euphausia superba, in relation to environmental variability" - David Smith, Sep92-present, Ph.D.: "Hydrography and circulation in the West Antarctic Peninsula region" MSU: - John Carlson, Jun95-present, Ph.D.: "Long-term trends in Adelie Penguin populations in the vicinity of Palmer Station, Antarctic Peninsula: The effects of variability in the breeding habitat" - Donna Patterson, Jan94-present, M.S.: "The effects of human activity on the biology of the Adelie penguin on Torgersen Island, Antarctic Peninsula " - Nina Karnovsky, Jun94-present, Ph.D.: "The fish component of Pygoscelid penguin diets: Implications for Resource Partitioning & ecosystem monitoring" - Tracey Mader, Jun95-present, M.S.: "The impacts of Leopard Seal predation on Pygoscelid penguins" Public Education There have been a number of magazine articles & a video made about the PAL program including: "Life on a melting continent" & "The secret lives of krill" (Stevens 1995a&b, Newton 1992). Also, all PIs have been involved in presenting lectures & slide shows in local schools & community groups. Currently, Bill Fraser, a PAL PI, is teaching a class on "The Ecology of Antarctica" to several hundred high school students world-wide via internet from Palmer as part of a project called "Blue Ice". International Interactions The Scientific Committee on Antarctic Research (SCAR) & the Convention for the Conservation of Antarctic Marine Living Resources (CCAMLR) are international governing bodies that receive recommendations on a host of issues related to the Southern Ocean through a diverse network of specialists & working groups. Two PAL PIs, William R. Fraser & Wayne Z. Trivelpiece, are, respectively, United States representatives to SCAR & CCAMLR, & address these bodies through participation in the Bird Biology Subcommittee of the Scientific Committee on Antarctic Research (WRF) & the CCAMLR Ecosystem Monitoring program (WRF & WZT). The most direct interactions with the PAL occur primarily through the CCAMLR Ecosystem Monitoring Program (CEMP), which seeks annual data on the ecology of Adelie penguins as part of its efforts to develop long-term monitoring programs. These data are delivered to CEMP through the Antarctic Marine Living Resources Program (AMLR), which provides U.S. funding for collection, analysis & preparation of annual reports (see reports by Fraser et al., 1988-1993).