PROGRAMME ( WCRP ) WMO / ICSU / IOC Wind , temperature and ice motion statistics in the Weddell Sea ( A compilation based on data from drifting buoys , vessels , and operational weather analyses )

The data from sea ice buoys, which were deployed during the Winter Weddell Sea Project 1986, the Winter Weddell Gyre Studies 1989 and 1992, the Ice Station Weddell in 1992, the Antarctic Zone Flux Experiment in 1994, and several ship cruises in Austral summers, are uniformly reanalyzed by the same objective methods. The buoys were capable of monitoring atmospheric pressure, air and ice temperatures, as well as position. The buoys were frequently arranged within groups of three to seven to allow calculation of reliable estimates of geostrophic winds and ice motion and under favorable conditions their spatial derivatives. Geostrophic winds for buoys operational regions are derived after matching of the buoy pressure data with the surf<i:ce pressure fields of the European Centre for Medium Range Weather Forecasts. Historical data from drifting ships are included into the temperature, air pressure and ice drift analyses. This report documents the mean structure as well as the variability of ice motion and spatial derivatives of ice motion, the statistics of surface pressure, geostrophic winds and air temperatures in the sea ice covered part of the Weddell Sea. To reflect the characteristics of ice dynamics on the basin-scale, all ice drift data are related to the geostrophic winds based on a complex linear model for daily averaged data. The composite patterns of mean ice motion, geostrophic winds, and geostrophic surface currents document cyclonic basinwide circulations. Geostrophic ocean currents are generally small in the Weddell Sea. Significant features are the coastal current near the southeastern coasts and the bands of larger velocities following the northward and eastward orientation of the continental shelf breaks in the western and northwestern Weddell Sea. In the southwestern Weddell Sea the mean ice drift speed is reduced to less than 0.5 % of the geostrophic wind speed and increases rather continuously to 1.5% in the northern, central, and eastern Weddell Sea. The linear model accounts for less than 50 % of the total variance of drift speeds in the southwestern Weddell Sea and up to 80 % in the northern and eastern Weddell Sea. The transmission of in-situ data from buoys to the Global Telecommunication System improves the quality of atmospheric analyses of the Numerical Meteorological Centers (NMC) significantly. This report presents atmospheric analyses of the European Centre for Medium Range Weather Forecasts for the periods with the largest number of buoys operational in the Weddell Sea. The analyses are used to obtain mean spatial distributions of 2-m temperatures and 10-m winds for different seasons. The variances of temperatures and winds with 6-hourly resolution are given for comparison with the variances derived from the buoy-measurements. ECMWF 2-m temperatures reflect only a very small fraction of the true variance until 1991. After a change in the ECMWF analysis scheme in 1992, the variance of the 2-m temperatures is only about 20 %smaller than the variance documented by the buoy measurements. The variance of ECMWF-10-winds is about 80 % of the variance of geostrophic wind fluctuations derived from the buoy data. 1 Objective of the report The role of sea ice within the polar climate system is determined by its motion in response to winds and ocean currents and by its thermodynamically induced formation and melting. The presence of sea ice significantly affects the energy and momentum exchanges between 2 Weddell Sea wind and ice motion the atmosphere and ocean. The dynamics and thermodynamics of sea ice are closely coupled by various processes. Among the quantities describing the development of the oceanic ice cover, the vector of sea ice motion is of special importance. The dominant time scale of ice motion variability is determined by the wind forcing, and thus reflects the close link in sea ice/atmosphere interaction. Spatial differences of sea ice motion are responsible for the opening and closing of leads between ice floes. These areas of open water contribute remarkably to the large scale heat flux from the ocean to the atmosphere particularly in wintertime. Sea ice motion and the salt rejection during ice formation generate turbulence in the oceanic mixed layer which may cause entrainment of warm intermediate water across its lower boundary. On a larger scale, the mean ice motion both contributes to the seasonal cycle of ice extent and transports fresh water from the regions of freezing to the regions of melting. The coupling of processe;; on different scales must he adequately represented in numerical models of the atmosphere, the sea ice and the ocean in order to obtain realistic results. For this purpose several parameters have to be determined from observations. Measurements are also needed to force numerical models and to test their quality. Automatic buoys on drifting sea ice provide an important and rather efficient means for a reasonable surface data acquisition. These systems are capable of measuring atmospheric pressure, air and ice temperatures to be communicated by satellites which also track the positions of the buoys. Systematic observations of this kind were started in 1986 and last up to the present time. Consequently large regions of the Antarctic sea ice belt were covered. The buoys were frequently arranged within groups of three to seven to allow calculation of reliable estimates of geostrophic winds and ice motion and, under favourable conditions, their spatial derivatives. Descriptions of the individual programs and related process studies have been published [Ackley and Holt, 1981, Allison, 1989, Crane and Bull, 1990, Hoeber, 1991, Kottmeier and Hartig, 1990, Ackley, 1981, Vihma and Launiainen, 1993, Wadhams et al., 1989, Kottmeier and Sellmann, 1996, Martinson and Wamser, 1990, Limbert et al., 1989]. The buoy data obtained can be also used as surface verification data for the satellite remote sensing of ice motion I Viehoff and Li, 1994, Drinkwater and Kottmeier, 1994]. On the basis of the above mentioned data of the Weddell Sea we will subsequently o decribe the mean structure and the variability of sea ice motion and of spatial derivatives of ice motion, o provide statistics of the surface pressure, the geostrophic wind and the air temperature for the sea ice region, o demonstrate the relationship between the ice motion, the surface wind field and ocean currents, o indicate improvements of the surface analyses of the European Centre for Medium Range Weather Forecasts due to the surface buoy measurements which are fed in near real time into the WMO Global Telecommunication System (GTS). Basically the report aims to characterize the full buoy data set from the Weddell Sea and to gi:'e an idea on the potential use for different research purposes. It is not the purpose of this report to present data records from certain regions or periods or to publish original 2 Data used for the analysis 3 scientific results. The basic data from individual buoys as well as the gridded fields shown in this report are available on request for scientific applications. We consider the data presented in this report as especially relevant for the verification of sea ice models. 2 Data used for the analysis 2.1 Data buoy measurements The first observations of about one year's period from the Weddell Sea in Austral winter were gathered during the involuntary drifts of the vessels Deutschland in 1911 and Endurance in 1915/1916, which got stuck in the pack ice fMeinardus, 1938]. During the.First Global Experiment of the Global Atmospheric Research Program {GARP) in 1979, meteorological buoys were deployed by parachute drop into the western Weddell Sea for the first time, providing meteorological data about every 1 or 2 days [Ackley and Holt, 1981]. Buoys were used more frequently in the years after 1986. During several wintertime ship operations such as the Winter Weddell Sea Project 1986 (WWSP86), the Winter Weddell Gyre Studies 1989 and 1992 (WWGS89, WWGS92), the Ice Station Weddell in 1992 (ISW), and the ANZFLUX study in 1994 buoys were launched, most of which survived 6 to 12 months (Augstein, 1987, Hempel, 1987, Augstein et al., 1991, Lemke, 1994). The Finnish meteorological buoys were deployed under FINNARP-expeditions in 1990 and 1992. The total number of drifting buoys and vessels in the Weddell Sea which provided a record length of at least one month, amounts to 80. Among these, 54 have been deployed in arrays of three to seven instruments. Data from summer ship operations are not included in this study. The buoys were produced by several manufacturers, who used quite different mechanical and electronic designs. The sensor equipment of the buoys also varied considerably. Most of the floatable buoys had an ice-strengthened nonmagnetic alloy or plastic hull. Air temperatures usually were measured by a thermistor within a radiation-shielded and selfaspirated housing about 1 m above the ice surface. The hull temperature was frequently measured by a sensor in thermal contact to the hull and the snow. These temperatures have been shown to differ considerably from the air temperature in summer and are not used here. Atmospheric pressure was measured by almost all buoys and, with only a few exceptions, by the very accurate and stable Paroscientific Digiquartz sensor (longterm accuracy of ~0.15 hPa). Various additional sensors were run successfully with drifting buoys but are not used in this analysis. These comprise wind, humidity, and snow fall sensors, thermistor profilers of about 2m length through the sea ice, thermistor chains of 250m length through the mixed layer and the pycnocline, and ocean currents meters and CTDs at certain depths. Data logging systems within the buoys collect and average the sensor data (usually over 10 ·minutes) and transmit them to two polar-orbiting satellites every 60 or 90 seconds. The buoy locations are determined from the Doppler shifts of the transmi