Jump to content Jump to footer
Install App

Install this app on your device for a better experience.

News detail view

Buoys, buoys – everywhere you look, buoys! Data from instruments moored in the ice characterise the sea ice, atmosphere and ocean during the MOSAiC expedition

Autonomous observations Sea ice drift Arctic

An extensive dataset of observations from the MOSAiC expedition shows that the Distributed Network of autonomous buoys can yield valuable insight into interactions between the sea ice, ocean and atmosphere.

Further, when combined with the results of sampling and manual measurements from the Central Observatory, the network can significantly improve our understanding of a wide range of processes. The data for periods in which the RV Polarstern couldn’t be on site is particularly valuable. Past and future analyses of MOSAiC’s buoy data will also help to improve predictive models.

Role of buoy observations during MOSAiC

The units of autonomous instruments deployed during the MOSAiC (Multidisciplinary drifting Observatory for the Study of Arctic Climate) expedition in 2019/20, or buoys for short, offer detailed information on how interactions between the sea ice, ocean and atmosphere are changing in the Arctic. An extensive study now shows the combined results from more than 200 buoys, unravelling the coupled system in all seasons of the year. They reflect both the considerable spatial variability in winter and the onset of melting in May.

This dataset was analysed by an international consortium of 50 researchers and the results have now been published. “This publication represents a further milestone in the assessment of the MOSAiC data. It addresses key scientific questions, while also offering a reference work for future analyses of buoy data. This joint accomplishment on the part of various countries was made possible by the extensive collaboration within MOSAiC,” explains Prof Benjamin Rabe, main author of the publication and part of the Physical Oceanography section at the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research in Bremerhaven. Rabe was also co-cruise leader for the winter 2020 segment of the expedition and has been co-coordinator of the OCEAN team throughout the MOSAiC.

Autonomous monitoring systems on the sea ice, like those deployed during the MOSAiC expedition, are important tools for determining sea-ice characteristics. They play a decisive part in understanding not just the development of sea ice, but also its interplay with complex processes in the ocean and atmosphere. They can be used throughout the year, providing important data in seasons and / or regions that are normally inaccessible (e.g. in winter). Accordingly, key data on the properties of the sea ice were gathered during the MOSAiC expedition, while the ice drifted through the Arctic and changed with the seasons. Particularly valuable: the buoy data for May and June 2020, a period in which the Research Vessel Polarstern had to leave the drifting Central Observatory for roughly a month; without the buoys, there would have been no data on the beginning of the melting phase. While the research icebreaker was away, the buoys in a network of autonomous stations with a radius of 80 to 100 km continued to gather data and transmit it to the mission’s home institutes via satellite. As such, buoys are indispensable tools for expanding our grasp of the coupled ocean-ice-atmosphere system and how it’s changing, along with the impacts of climate change on the polar regions.


Results of the data analysis

Perhaps the most obvious, but nevertheless highly remarkable finding from the expedition: the sea- ice drift was substantially faster than expected, not to mention faster than the drift speeds observed in the past 15 years, which were used as reference values for planning purposes. This had scientific consequences and a major influence on expedition logistics and structure. Figure 1 shows how the network of autonomous systems (DN – Distributed Network; see the section at the end of this article for further details) followed the route of the transpolar drift, crossed Fram Strait, and then drifted into the Greenland Sea. Figure 2 shows the positions of the DN sites at six selected points in time during the drift (marked accordingly in Figure 1). The images reflect the DN at the time of deployment, in late spring, and toward the end of the drift, as it neared Fram Strait. Observations of ice motion alone were used to analyse the ice dynamics – the feedback between the sea ice and the wind and ocean currents, which produces pack-ice hummocks and leads, and in turn, new and thin ice. The MOSAiC buoys revealed that the sea-ice mass balance is to a significant degree shaped by these dynamics. Thanks to the unprecedented number of buoys in the DN and correspondingly high number of nodes in the network, it was possible to describe and gain a better understanding of ice conditions at scales ranging from 10 to 100 km. In addition, the vertical processes on several floes along the drift route were simultaneously recorded by the buoy network. On the basis of this combined data, long-term forecasts can be made on the spatial heterogeneity of the coupled atmosphere, ice and ocean systems. The DN successfully monitored the transition to melting in the late spring / early summer, as well as the change between the regions affected by low-saline polar surface water and those affected by near-surface water masses originating in the Atlantic.

In the following, selected findings are presented for two closely observed periods: one in winter and one in early summer. The former represents an infrequently observed season in the Central Arctic, while the latter covers the timeframe in which the ship and all personnel had left the MOSAiC research area and only autonomous observations were possible.

Figure 3 shows a representative outcome of the winter case study. The sites were located approximately 50 km from one another, and these relative distances, with changes of between 1 and 2 km, remained fairly stable. During the 30-day case study, the buoy network covered a spatial gradient of absolute salinity in the upper mixed layer below the ocean’s surface, with generally higher values in the southwest and lower values in the northeast section of the area in question. The observed gradient is embedded in the large-scale gradient of near-surface salinity and freshwater concentration between the Eurasian and Amerasian Basins. Within the network, ocean salinity varied considerably, both weekly and even daily. At the same time, the spatial variation in surface temperature was less pronounced than the temporal variation (no more than 5° C). The median ice drift varied to nearly the same extent as the windspeed, though the drift wasn’t free, being affected by various forces, e.g. ice-internal forces. A semidiurnal fluctuation in the sea-ice drift was observed, which is indicative of forcing by tides and / or inertial motion in the upper layers of the ocean. These fluctuations demonstrate the coupling of the ocean and ice. During the case study, the atmosphere did not manifest these semidiurnal fluctuations. The observed variability in snow depth at different measuring points can be attributed to the redistribution of snow and to local effects.

The chief advantages of autonomous instruments are the spatial distribution of observations and their ability to close chronological gaps in the manual observations made at the Central Observatory. When said observations had to be suspended from 16 May to 18 June 2020 due to the Polarstern’s departure for a personnel rotation, thanks to the buoy network it was still possible to gather data from 83.4° N to roughly 82.4° N. Figure 4 shows a representative outcome of the summer case study, based on data gathered by mass-balance buoys. The figure reflects the data from three buoys (from the network’s main node). From 26 May, the surface temperature rose above the freezing point, which also led to a significant warming of the snow (above the dashed line) and the sea ice below. As can be seen in Figure 4a–c, the frigid core of the sea ice gradually warmed from May to June 2020 – initially in response to intensive warming from above, and gradually also from below. This also implies that the percentage of brine in the ice by volume slowly rose. Two of the three buoys already showed surface melting at this time (reduced snow-cover depth), while the ice thickness on the underside barely changed. Here, too, we can recognise the varying influence of the atmosphere and ocean on the sea ice.

The network of autonomous stations during the MOSAiC expedition

The MOSAiC Distributed Network (DN) operated during the drift from the eastern Eurasian Arctic to Fram Strait from late 2019 to the first half of 2020, and again, after the relocation of the Central Observatory to the Central Arctic, from late summer to early autumn 2020 (see Figure 5). During the MOSAiC expedition, this network of various autonomous, ice-bound buoys was intended to close the gaps in our understanding of temporal and spatial scales, particularly with regard to improving the representation of important processes in Earth system models. Characterising the variability of climate-relevant processes near a Central Observatory with the aid of local measurements allowed us to capture not only coupled system interactions at the interfaces between the ocean, ice and atmosphere, but also three-dimensional processes. “All in all, these observations not only improve our understanding of the constantly changing Arctic, but also provide unique opportunities to validate models. Ultimately, the analysis of the buoy data leads to new model parametrisations of processes in the coupled system of the ocean, sea ice and atmosphere,” says Benjamin Rabe, summarising the work involved in and outcomes of the comprehensive data analysis. The comprehensive nature of the instruments used, together with simultaneous icebreaker operations and a Central Observatory with supplementary measurements and manual sample gathering, are key aspects of the MOSAiC DN, representing an unprecedented combination.

The autonomous buoys deployed in the course of the year-long expedition varied in complexity. In this regard, each of the more than 30 types used is unique in terms of its technical specifications, measuring variables, and vertical and temporal resolution. They were installed at various sites, primarily within a 50-km radius from the Central Observatory. All of them recorded their data autonomously, transmitting the majority of it back to land via satellite; the remaining data was collected during maintenance checks or when the buoys were retrieved. The buoys were able to capture both vertical processes on an individual floe, and horizontally heterogeneous processes from 10 to 100 km and at smaller scales along the drift route, quasi-synoptically (almost simultaneously). The selected sites showed considerable variability in snow depth and ice thickness and permitted a high-resolution temporal and spatial characterisation of the ice motion and deformation.

The data and map archive at SEAICEPORTAL.de includes a freely accessible repository of buoy data and currently covers more than 600 buoys, including 475 in the Arctic, 214 of which were deployed during the MOSAiC expedition (see Figure 6). The buoy data is part of the International Arctic Buoy Programme (IABP) and the International Programme for Antarctic Buoys (IPAB). “Thanks to MOSAiC, the SEAICEPORTAL’s data portal has taken on a new importance and is now the largest portal for sea-ice-based buoy data. We have also established six new buoy types so as to offer more data, also from the atmosphere and ocean. As such, the MOSAiC data is the most extensive dataset at the sea-ice portal,” explains Dr Marcel Nicolaus (AWI), who was responsible for the sea-ice work during MOSAiC (see Figure 7).



The findings presented in the study underline the importance of carefully considering differences in the ice thickness and overall conditions at the buoys’ respective deployment sites; these factors can lead to substantial changes in mass and energy transfer between the atmosphere, sea ice and ocean.

Gathering measurements with a Distributed Network facilitated the analysis of processes, which would not have been possible using clusters of autonomous instruments, which are separated by greater distances than those between the DN sites; or using autonomous systems alone, without manual observations. Consequently, the network is of great scientific value in a number of ways: due to the range of temporal and spatial scales covered by the DN, it is particularly well-suited to understanding measurements and processes at scales smaller than the 10 to 100 km grid cells typically used in Earth system and climate models.

Lastly, the successful implementation demonstrates the feasibility of such networks. In addition, it provides important guidelines for future installations of autonomous observational networks, which should take place more frequently in a warming world if our goal is to better understand and assess the complex interactions in the climate system and how they affect sea-ice retreat.

“At the same time, the release of this publication marks the beginning of a new phase, one in which we will increasingly combine the buoy data with observations via satellite, from the air, on the ice, and beneath it (e.g. with the aid of Remotely Operated Vehicles). This will allow us to link temporal and seasonal changes in the data with spatial variability, which is necessary in order to gain a better understanding of the system at different scales,” says Marcel Nicolaus with regard to the next steps.This will allow us to link temporal and seasonal changes in the data with spatial variability, which is necessary in order to gain a better understanding of the system at different scales,” says Marcel Nicolaus with regard to the next steps.


Link to the publication:

Rabe, B., and 49 others (2024): The MOSAiC Distributed Network: Observing the coupled Arctic system with multidisciplinary, coordinated platforms , Elem Sci Anth, 12 (1).

doi: 10.1525/elementa.2023.00103


Prof Benjamin Rabe (AWI)

Dr Marcel Nicolaus (AWI)

Dr Renate Treffeisen (AWI)

Dr Klaus Grosfeld (AWI)


Write us an

email or use the contact form.