No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Antarctic sea ice: A self-organizing map-based perspective
Abstract
Self-organizing maps (SOMs) provide a powerful, non-linear technique to optimally summarize a complex geophysical dataset using a user-selected number of 'icons' or SOM states, allowing rapid identification of preferred patterns, predictability of transitions, rates of transitions, and hysteresis in cycles. The use of SOMs is demonstrated here through application to a 24 year dataset (1973-96) of monthly Antarctic sea-ice edge positions. Variability in sea-ice extent, concentration and other physical characteristics is an important component of the Earth's dynamic climate system, particularly in the Southern Hemisphere where annual changes in sea-ice extent (temporarily) double the size of the Antarctic cryosphere. SOM-based patterns concisely capture the spatial and temporal variability in these data, including the annual progression of expansion and retreat, a general eastward propagation of anomalies during the winter, and sub-annual variability in the rate of change in extent at different times of the year (e.g. retreat in January is faster than in November). There is also often a general seasonal hysteresis, i.e. monthly anomalies during cooling follow a different spatial path than during warming.
... For instance, Tymvios et al. (2010) tested several different array sizes, and ultimately chose a 36-node array of 500-mb geopotential heights for analysis after showing that this size array best delineated among the precipitation events that were the focus of their research. Reusch et al. (2007) has suggested the array not be symmetric for the stability of the calculations, although symmetric arrays have been presented in some articles (e.g. Tozuka et al., 2008). ...
... Leloup et al. (2007) base a 10Â10 SOM on multiple metrics of sea-surface temperature – clustered using a HAC into 12 subgroups – to assess the role of phase transitions of ENSO events. Reusch et al. (2007) study the North Atlantic Oscillation (NAO), using wintertime monthly SLP anomalies, to identify the causes of the trends in the teleconnection, and visualize the intermediate patterns between the opposite phases. Johnson et al. (2008) utilize the continuum provided by SOM nodes in analyzing the spatial and temporal trends in the NAO to assess the differences in circulation regimes between the pre-1978 and post- 1978 periods. ...
... To date, few studies have directly compared the efficacy of SOMs to other methods. Among those few studies, however, Reusch et al. (2005a, 2007) directly compare PCA patterns to SOMs, and discover that both methods have unique advantages, though the SOMs show improved visualization of data, are less likely to mix patterns than PCA, and thus more clearly depict 'blended' patterns when a SOM is expanded to include additional nodes. Michaelides et al. (2001) show that SOMs characterize precipitation variability better than traditional clustering methods. ...
Self-organizing maps (SOMs) are a relative newcomer to synoptic climatology; the method itself has only been utilized in the field for around a decade. In this article, we review the major developments and climatological applications of SOMs in the literature. The SOM can be used in synoptic climatological analysis in a manner similar to most other clustering methods. However, as the results from a SOM are generally represented by a two-dimensional array of cluster types that 'self-organize', the synoptic categories in the array effectively represent a continuum of synoptic categorizations, compared with discrete realizations produced through most traditional methods. Thus, a larger number of patterns can be more readily understood, and patterns, as well as transitional nodes between patterns, can be discerned. Perhaps the most intriguing development with SOMs has been the new avenues of visualization; the resultant spatial patterns of any variable can be more readily understood when displayed in a SOM. This improved visualization has led to SOMs becoming an increasingly popular tool in various research with climatological applications from other disciplines as well.
... Despite the amount of literature presented using SOMs for climate related studies, little work has been done with the method for classifying land cover types or sea ice relationships. Reusch and Alley (2007) investigated monthly Antarctic sea-ice edge positions, using SOMs to cluster the data and give an annual average ice cycle (Reusch and Alley 2007), but a similar study has not been conducted for the Arctic using SOMs. To demonstrate the application of SOMs in surface cover transition studies, here specifically cryocover, this analysis will look at cryo-cover patterns for three key transitional months-March, June, and October. ...
... Despite the amount of literature presented using SOMs for climate related studies, little work has been done with the method for classifying land cover types or sea ice relationships. Reusch and Alley (2007) investigated monthly Antarctic sea-ice edge positions, using SOMs to cluster the data and give an annual average ice cycle (Reusch and Alley 2007), but a similar study has not been conducted for the Arctic using SOMs. To demonstrate the application of SOMs in surface cover transition studies, here specifically cryocover, this analysis will look at cryo-cover patterns for three key transitional months-March, June, and October. ...
Northern Hemisphere cryo-cover (here defined as both snow and sea ice covers) is dynamic and has displayed distinct spatial and temporal variability, as well as shifts in stability (predominantly becoming less stable) in recent decades. To better understand these transitions, this manuscript focuses on three transitional months of cryo-cover transformation, March, June and October, and implements Self-Organizing Maps (SOMs) on a MeaSuREs derived (temporal coverage of 1979–2012) cryo-cover concentration dataset. Cryo-cover transformation was analyzed on a regional and hemispheric basis with trends calculated for all areas identified by the SOMs. Results suggest cryo-cover decreases (in area and stability) during March and June as pixels shift from SOM identified stable patterns to more transient patterns. October results suggest increases in transient cryo-cover, a result of declining stability in sea ice but also increasing snow cover predominantly centered around 50 °N latitude. All months experience regions of increasing and decreasing cryo-cover trends. Decreasing trends are predominant in March and June, although some areas of increase are apparent even during these months. By contrast, October has significant area experiencing increasing trends in cryo-cover, as well as large decreasing trends in the ice domain. Given accelerated rising temperatures in the Northern Hemisphere high latitudes, transformation of cryo-cover will continue, and monitoring these changes spatially (using methodologies like SOMs) will be necessary to better understand the radiative and environmental impacts.
... The SOM has been used in a similar way (i.e. for clustering and classifying data contained in 2-dimensional maps or images), in many applications of environmental science, climatology, geology, and oceanography. These include analyzing sea surface temperature46474849, plankton [50, 51], ocean current patterns [43, 52], estuary and basin dynamics [53], sediment structure [54], atmospheric pressure [55, 56], wind patterns [39], storm systems [41], the El Niño weather conditions [42], clouds [57], ice [53, 58, 59], rainfall [44, 60, 61], oil spills [45], the influence of ocean conditions in droughts [62], and the relationship between sardine abundance and upwelling phenomena [40]. Data concerning fisheries was analyzed in different perspectives using a SOM in [63]. ...
... The SOM has been used in a similar way (i.e. for clustering and classifying data contained in 2-dimensional maps or images), in many applications of environmental science, climatology, geology, and oceanography. These include analyzing sea surface temperature46474849, plankton [50, 51], ocean current patterns [43, 52], estuary and basin dynamics [53], sediment structure [54], atmospheric pressure [55, 56], wind patterns [39], storm systems [41], the El Niño weather conditions [42], clouds [57], ice [53, 58, 59], rainfall [44, 60, 61], oil spills [45], the influence of ocean conditions in droughts [62], and the relationship between sardine abundance and upwelling phenomena [40]. Data concerning fisheries was analyzed in different perspectives using a SOM in [63]. ...
Self-Organizing Maps (SOMs), or Kohonen networks, are widely used neural network architecture. This paper starts with a brief
overview of how SOMs can be used in different types of problems. A simple and intuitive explanation of how a SOM is trained
is provided, together with a formal explanation of the algorithm, and some of the more important parameters are discussed.
Finally, an overview of different applications of SOMs in maritime problems is presented.
... Johnson & Feldstein (2010) presented an SOM analysis that illustrated coupled variability between the North Pacific sea level pressure field and outgoing longwave radiation in the tropical Indo-Pacific region so as to shed light on the relationship between the North Pacific continuum and tropical convection. Reusch et al. (2007) used the SOM to analyze the monthly mean sea level pressure for North Atlantic climate variability. A review of SOM classifications of atmospheric circulation patterns within synoptic climatology is provided in Huth et al. (2008), and an overview in remote sensing applications is seen in Filippi et al. (2010). ...
... Arctic and adjacent regions Cassano et al. (2006, 2007, 2010), Higgins & Cassano (2009), Finnis et al. (2009a, b), Skific et al. (2009a, b) North Atlantic, Greenland Reusch et al. (2007), Schuenemann et al. (2009), Schuenemann & Cassano (2010a, b) Europe, Middle East & Northern Africa Michaelides et al. (2007) Northern hemisphere, North Pacific Johnson et al. (2008, 2010), Johnson & Feldstein (2010). Australia Hope et al. (2006), Verdon-Kidd & Kiem (2008), Alexander et al. (2010) Synoptic climatology or synoptic patterns of atmospheric circulation based on SOM analysis of sea level pressure, geopotential height at different pressure levels ...
... The SOM has been used in a similar way (i.e. for clustering and classifying data contained in 2-dimensional maps or images), in many applications of environmental science, climatology, geology, and oceanography. These include analyzing sea surface temperature [46][47][48][49], plankton [50,51], ocean current patterns [43,52], estuary and basin dynamics [53], sediment structure [54], atmospheric pressure [55,56], wind patterns [39], storm systems [41], the El Niño weather conditions [42], clouds [57], ice [53,58,59], rainfall [44,60,61], oil spills [45], the influence of ocean conditions in droughts [62], and the relationship between sardine abundance and upwelling phenomena [40]. ...
... The SOM has been used in a similar way (i.e. for clustering and classifying data contained in 2-dimensional maps or images), in many applications of environmental science, climatology, geology, and oceanography. These include analyzing sea surface temperature [46][47][48][49], plankton [50,51], ocean current patterns [43,52], estuary and basin dynamics [53], sediment structure [54], atmospheric pressure [55,56], wind patterns [39], storm systems [41], the El Niño weather conditions [42], clouds [57], ice [53,58,59], rainfall [44,60,61], oil spills [45], the influence of ocean conditions in droughts [62], and the relationship between sardine abundance and upwelling phenomena [40]. ...
... The classification has been created using the self-organizing map (SOM) algorithm (Kohonen, 2001), which has already been used in Antarctica for assessment of atmospheric circulation development during the 20th and 21st centuries (Lynch et al., 2006), near-surface wind field climatology over the Ross Ice Shelf (e.g., Seefeldt and Cassano, 2012;Nigro and Cassano, 2014), for forecast performance evaluation (Nigro et al., 2011;Nigro et al., 2017), synoptic conditions reconstruction (Reusch et al., 2005b) and sea ice characteristic analysis (Reusch and Alley, 2005). The nodes on the SOM are referenced by their row and column number. ...
The influence of synoptic‐scale circulation on air temperature variation in the ice‐free and glaciated areas on the eastern side of the Antarctic Peninsula (AP) has been analysed. For this purpose, a new classification of atmospheric circulation with 15 synoptic patterns in the AP region was developed using the self‐organizing maps technique. The synoptic patterns were compared with air temperature observations from coastal and glacial sites on James Ross Island, northeastern AP, in the period 2005–2015. The most frequent synoptic pattern with a frequency of 13.7% was dominated by a low‐pressure centre in the northwestern Bellingshausen Sea, which extended over the AP to the Weddell Sea. On the other hand, the largest inter‐annual variability was observed for a synoptic pattern with a low‐pressure centre in the southern Bellingshausen Sea. This synoptic pattern also had the highest air temperature anomalies at both investigated sites year‐round. Air temperature anomalies at the lower lying site (Mendel station) were the lowest during a high‐pressure ridge dominating the AP region due to a combination of local and synoptic‐scale processes. At Davies Dome, the glacial site, southerly barrier winds advecting cold air from the ice‐covered Weddell Sea during a strong low‐pressure system in the Weddell Sea ensured the coldest air temperature anomalies.
... Geographical information system (GIS) and ANN are commonly used tools in the analysis of geospatial data. ANNs have been used in cryospheric studies to derive glacier ice thickness estimation (Clarke, Berthier, Schoof, & Jarosch, 2009;Haq, Jain, & Menon, 2014;Haq & Azam, 2017), to understand the glacier length variations (Steiner, Walter, & Zumbuhl, 2005;Zumbühl, Steiner, & Nussbaumer, 2008), to predict snow cover in the mountain ranges (Mishra, Tripathi, & Babel, 2014;Sauter, Schneider, Kilian, & Moritz, 2009;Sauter, Weitzenkamp, & Schneider, 2010) to simulate melt water runoff from glaciers (Caiping & Yongjian, 2009), to classify glacier and snow cover (Bishop, Shroder, & Hickman, 1999;Czyzowska-Wisniewski, van Hendrikx & Hreinsson, 2012) to classify patterns of the extent and concentration of sea ice (Reusch & Alley, 2007), and to predict snow cover in mountain ranges (Sauter et al., 2009;Sauter et al., 2010). ...
This study attempts to model snow wetness and snow density of Himalayan snow cover using a combination of Hyperspectral image processing and Artificial Neural Network (ANN). Initially, a total of 300 spectral signature measurements, synchronized with snow wetness and snow density, were collected in the field. The spectral reflectance of snow was then modeled as a function of snow properties using ANN. Four snow wetness and three snow density models were developed. A strong correlation was observed in near‐infrared and shortwave‐infrared region. The correlation analysis of ANN modeled snow density and snow wetness showed a strong linear relationship with field‐based data values ranging from 0.87–0.90 and 0.88–0.91, respectively. Our results indicate that an Artificial Intelligence (AI) approach, using a combination of Hyperspectral image processing and ANN, can be efficiently used to predict snow properties (wetness and density) in the Himalayan region.
Recommendations for resource managers
• Snow properties, such as snow wetness and snow density are mainly investigated through field‐based survey but rugged terrains, difficult weather conditions, and logistics management issues establish remote sensing as an efficient alternative to monitor snow properties, especially in the mountain environment.
• Although Hyperspectral remote sensing is a powerful tool to conduct the quantitative analysis of the physical properties of snow, only a few studies have used hyperspectral data for the estimation of snow density and wetness in the Himalayan region. This could be because of the lack of synchronized snow properties data with field‐based spectral acquisitions.
• In combination with Hyperspectral image processing, Artificial Neural Network (ANN) can be a useful tool for effective snow modeling because of its ability to capture and represent complex input‐output relationships.
• Further research into understanding the applicability of neural networks to determine snow properties is required to obtain results from large snow cover areas of the Himalayan region.
... Each node in the map has the same dimensionality as the original data. The SOM method has been successfully applied to answer other questions about sea ice in the past, including characterizing variability of Antarctic sea ice [Reusch and Alley, 2007] and assessing the dependence of Arctic sea ice variability on synoptic systems [Mills and Walsh, 2014]. Although SOMs are usually applied to spatial fields in the geosciences [e.g., Johnson et al., 2008], there is nothing that restricts the method to spatial data. ...
Using Arctic sea-ice concentration derived from passive microwave satellite observations in autumn and early winter over the 1979–2014 period, the Arctic region was objectively classified into several smaller regions based on the interannual sea-ice variability through self-organizing map analyses. The trend in regional sea-ice extent (RSIE) in each region was removed using an adaptive, nonlinear, and non-stationary method called Ensemble Empirical Mode Decomposition, which captures well the accelerating decline of Arctic RSIEs in recent decades. Although the linear trend in RSIE is negative in all regions in both seasons, there are marked differences in RSIE trends and variability between regions, with the largest negative trends found during autumn in the Beaufort Sea, the Barents–Kara Seas, and the Laptev–East Siberian Seas. Winter weather patterns associated with the nonlinearly detrended RSIEs show distinct features for different regions and tend to be better correlated with the autumn than early winter RSIE anomalies. Sea-ice losses in the Beaufort Sea and the Barents–Kara Seas are both associated with a cooling of Eurasia, but in the former case the circulation anomaly is reminiscent of a Rossby wave train, whereas in the latter case the pattern projects onto the negative phase of the Arctic Oscillation. These results highlight the non-uniform changes in Arctic sea ice and suggest that regional sea-ice variations may play a crucial role for the winter weather patterns.
Sea ice variability in the opposite polar regions is examined holistically by applying the self-organizing map (SOM) method to global monthly sea ice concentration data over two periods. The results show that the variability modes of sea ice decrease in the Arctic correspond to an overall sea ice increase in the Antarctic, and vice versa. In particular, the monthly sea ice anomaly patterns are dominated by in-phase variability across the Arctic that is stronger in the marginal seas particularly the Barents Sea than the central Arctic Ocean. The corresponding Antarctic sea ice variability is characterized by a zonal wavenumber-3 structure or a dipole pattern of out-of-phase variability between the Bellingshausen/Amundsen Seas and the rest of the Southern Ocean. The frequency of occurrence of these dominant patterns exhibits pronounced seasonal as well as decadal variability and the latter is closely related to the Pacific decadal oscillation and Atlantic multidecadal oscillation. Other less frequent patterns seem to be associated with the central Pacific El Niño and spatially heterogeneous interannual variability of sea surface temperature (SST) in the Indian and the Atlantic Oceans. The dominant modes explain 57% of the four-decade domain-averaged trends in the annual polar sea ice concentration, with more explained in the eastern than western Arctic Ocean and in the Weddell Sea and the Amundsen Sea in the Antarctic. The spatial patterns of the leading modes can be largely explained by the dynamic (sea ice drift) and thermodynamic (sea ice melt) effects of the anomalous atmospheric circulations associated with SST and sea level pressure anomalies.
Significance Statement
The purpose of this study is to extract the main modes of monthly global sea ice concentration variability in the past four decades, explain the mechanisms behind the occurrences of these modes, and examine the contributions of these modes to the trend in annual global sea ice concentration. Sea ice extent in the past four decades has shown a significant declining trend in the Arctic and a slight, but significant increasing trend in the Antarctic. By jointly analyzing the sea ice variability and trends in the two polar regions, the results here provide a reference for what might have contributed to the opposite sea ice trends in Arctic and Antarctic and highlight the important influence of large-scale sea surface temperature anomalies on the trends in the two polar regions.
The potential mechanisms underlying the observed increasing trend in Antarctic summertime sea ice cover for the 1979-2017 period have been investigated using a relatively novel method called the self-organizing map (SOM). Among the nine nodes generated to explain the variability of Antarctic sea ice cover, two (nodes 3 and 7) exhibit a statistically significant linear trend in the time series of the frequency of the SOM pattern occurrence that together explain 40% of the total trend in the sea ice cover. These two nodes have completely opposite spatial patterns and directions of trend and are associated with different atmospheric circulation patterns. Node 3, which represents an increase in sea ice over the Weddell Sea and the eastern Ross Sea and a decrease over the other coastal seas of West Antarctica, appears to be related to the positive phase of the southern annular mode (SAM) linked with the La Niña pattern of sea surface temperature (SST) over the tropical Pacific Ocean. The opposite spatial pattern and trend represented by node 7 is associated with a wave train originating over northern Australia. The anomalous wind field, surface downward longwave radiation, and surface air temperature generated by these circulation patterns are consistent with the spatial pattern and overall trends in the Antarctic sea ice cover.
Robust projections of climate change impacts on regional hydrology are crucial for water resources management, especially in data-sparse regions. Impact projections have to: (a) be available at gauged and ungauged locations; (b) consider changes in watershed behaviour for different future climates and land uses; and (c) include estimates of uncertainty. We apply a novel water resources modelling framework which combines signature regionalization with the trading of space for time to obtain hydrological projections in ungauged basins and under climate change for the Olifants basin in southern Africa, a UNESCO HELP basin. We find that using statistically downscaled general circulation model (GCM) vs observed climate data leads to only slight deterioration in reliability of runoff projections for the historical period. The framework projects a decrease in regional runoff by –8.7% and –3.9% for A2 and B1 scenarios, respectively, by the end of the century, with, 80% and 67% of GCMs agreeing on the decrease in runoff, respectively.
To predict the rate and consequences of shrinkage of the earth's mountain glaciers and ice caps, it is necessary to have improved regional-scale models of mountain glaciation and better knowledge of the subglacial topography upon which these models must operate. The problem of estimating glacier ice thickness is addressed by developing an artificial neural network (ANN) approach that uses calculations performed on a digital elevation model (DEM) and on a mask of the present-day ice cover. Because suitable data from real glaciers are lacking, the ANN is trained by substituting the known topography of ice-denuded regions adjacent to the ice-covered regionsof interest, and this knowntopographyis hidden by imagining it to be ice-covered. For this training it is assumed that the topography is flooded to various levels by horizontal lake-like glaciers. The validity of this assumption and the estimation skill of the trained ANN is tested by predicting ice thickness for four 50 km 3 50 km regions that are currently ice free but that have been partially glaciated using a numerical ice dynamics model. In this manner, predictions of ice thickness based on the neural network can be compared to the modeled ice thickness and the performance of the neural network can be evaluated and improved. From the results, thus far, it is found that ANN depth estimates can yield plausible subglacial topography with a representative rms elevation error of 670 m and remarkably good estimates of ice volume.
This study uses a statistical downscaling method based on self-organizing maps (SOMs) to produce high-resolution, downscaled precipitation estimates over the state of Pennsylvania in the mid-Atlantic region of the United States. The SOMs approach derives a transfer function between large-scale mean atmospheric states and local meteorological variables (daily point precipitation values) of interest. First, the SOM was trained using seven coarsely resolved atmospheric variables from the National Centers for Environmental Prediction (NCEP) reanalysis dataset to model observed daily precipitation data from 17 stations across Pennsylvania for the period 1979-2005. Employing the same coarsely resolved variables from nine general circulation model (GCM) simulations taken from the historical analysis of the Coupled Model Intercomparison Project, phase 3 (CMIP3), the trained SOM was subsequently applied to simulate daily precipitation at the same 17 sites for the period 1961-2000. The SOM analysis indicates that the nine model simulations exhibit similar synoptic-scale features to the (NCEP) observations over the 1979-2007 training interval. An analysis of the sea level pressure signatures and the precipitation distribution corresponding to the trained SOM shows that it is effective in differentiating characteristic synoptic circulation patterns and associated precipitation. The downscaling approach provides a faithful reproduction of the observed probability distributions and temporal characteristics of precipitation on both daily and monthly time scales. The downscaled precipitation field shows significant improvement over the raw GCM precipitation fields with regard to observed average monthly precipitation amounts, average monthly number of rainy days, and standard deviations of monthly precipitation amounts, although certain caveats are noted.
Trends in the satellite-derived Antarctic sea ice concentrations
(1979-2002) show pronounced increase (decrease) in the central Pacific
sector (Bellingshausen/western Weddell sector) by ~4-10% per decade.
Confidence levels for these regional trends exceed 95%. Positive
polarities of the Antarctic Oscillation (AAO) lead to more (less) ice in
the eastern Ross/Amundsen sector (Bellingshausen/northern Weddell
sector), which are qualitatively opposite to the impacts of positive
polarities of the El Niño-Southern Oscillation (ENSO). The
mechanisms responsible for the covariability between the ice and the (a)
AAO and (b) ENSO are demonstrated. Over the last 24 years, a positive
AAO trend and a slightly negative ENSO trend produce a spatial pattern
of ice changes similar to the regional ice trends. However, the
magnitude of the ice changes associated with the AAO and ENSO is much
smaller than the regional ice trends. More local (or less understood
large) scale processes should be investigated for the explanations.
Using monthly Arctic sea ice concentration data (1953-95) and the National Centers for Environmental Prediction-National Center for Atmospheric Research (NCEP-NCAR) reanalysis dataset (1958-99), possible feedbacks of sea ice variations in the Greenland and Barents Seas to the atmosphere are investigated. Winter (February-April) sea ice anomalies in the Greenland and Barents Seas display important feedback influences on the atmospheric boundary layer in terms of both thermodynamic and dynamic processes. The vertical gradients of potential pseudo-equivalent temperature (use) between 850 and 700 hPa are greater over sea ice than over open water, implying that a more stable boundary layer forms below 700 hPa over sea ice. The effects of temperature advection are shown to account for a relatively small percentage of the temperature variance in area of sea ice feedbacks. Horizontal and vertical variations of the effects of sea ice on temperature in the Nordic and Barents Seas create the potential for dynamical influences on the atmospheric boundary layer through vertical motion induced by the pressure anomalies in the boundary layer.
Self organizing maps (SOMs) are used to locate archetypal points that describe the multi-dimensional distribution function of a gridded sea level pressure data set for the northeast United States. These points-nodes on the SOM-identify the primary features of the synoptic-scale circulation over the region. In effect, the nodes represent a non-linear distribution of overlapping, non-discreet, circulation types. The circulation patterns are readily visualized in a 2-dimensional array (the SOM) that places similar types adjacent to one another and very different types far apart in the SOM space. The SOM is used to describe synoptic circulation changes over time, and to relate the circulation to January station precipitation data (for State College, Pennsylvania) in the center of the domain. The paper focuses on the methodology; however, the analysis suggests that circulation systems that promote precipitation have decreased over the last 40 yr-although January precipitation at State College has actually increased. Further analysis with the SOM indicates that this is due to a change in precipitation characteristics of the synoptic-scale circulation features, rather than to their frequency of occurrence.
The anomalies in the climate and sea ice cover of the Southern Ocean and their relationships with the Southern Oscillation (SO) are investigated using a 17-yr dataset from 1982 to 1998. The polar climate anomalies are correlated with the Southern Oscillation index (SOI) and the composites of these anomalies are examined under the positive (SOI > 0), neutral (0 > SOI > 1), and negative (SOI < 1) phases of SOI. The climate dataset consists of sea level pressure, wind, surface air temperature, and sea surface temperature fields, while the sea ice dataset describes its extent, concentration, motion, and surface temperature. The analysis depicts, for the first time, the spatial variability in the relationship of the above variables with the SOI. The strongest correlation between the SOI and the polar climate anomalies are found in the Bellingshausen, Amundsen, and Ross Seas. The composite fields reveal anomalies that are organized in distinct large-scale spatial patterns with opposing polarities at the two extremes of SOI, and suggest oscillations that are closely linked to the SO. Within these sectors, positive (negative) phases of the SOI are generally associated with lower (higher) sea level pressure, cooler (warmer) surface air temperature, and cooler (warmer) sea surface temperature in these sectors. Associations between these climate anomalies and the behavior of the Antarctic sea ice cover are evident. Recent anomalies in the sea ice cover that are clearly associated with the SOI include the following: the record decrease in the sea ice extent in the Bellingshausen Sea from mid-1988 to early 1991; the relationship between Ross Sea SST and the ENSO signal, and reduced sea ice concentration in the Ross Sea; and the shortening of the ice season in the eastern Ross Sea, Amundsen Sea, far western Weddell Sea and lengthening of the ice season in the western Ross Sea, Bellinghausen Sea, and central Weddell Sea gyre during the period 1988-94. Four ENSO episodes over the last 17 years contributed to a negative mean in the SOI (0.5). In each of these episodes, significant retreats in ice cover of the Bellingshausen and Amundsen Seas were observed showing a unique association of this region of the Antarctic with the Southern Oscillation.
As a contribution toward improving our ability to identify robust patterns of variability in complex, noisy climate datasets, we have compared a relatively new technique, Self-Organizing Maps (SOMs), to the well-established method of principal component analysis (PCA). Recent results suggest that SOMs offer advantages over PCA for use in climatological and other studies. Here each analysis technique was applied to synthetic datasets composed of positive and negative modes of four idealized North Atlantic sea-level-pressure fields, with and without noise components, to identify the predefined patterns of variability. PCA, even with component rotation, fails to adequately extract the known spatial patterns, mixes patterns into single components, and incorrectly partitions the variance among the components. The SOMs-based analyses are more robust and, with a sufficiently large set of generalized patterns, are able to isolate all the predefined patterns with correct attribution of variance. With PCA, it is difficult, if not impossible, to detect pattern mixing without prior knowledge of the patterns being mixed.
The pack ice of Earth's polar oceans appears to be frozen white desert, devoid of life. However, beneath the snow lies a unique
habitat for a group of bacteria and microscopic plants and animals that are encased in an ice matrix at low temperatures and
light levels, with the only liquid being pockets of concentrated brines. Survival in these conditions requires a complex suite
of physiological and metabolic adaptations, but sea-ice organisms thrive in the ice, and their prolific growth ensures they
play a fundamental role in polar ecosystems. Apart from their ecological importance, the bacterial and algae species found
in sea ice have become the focus for novel biotechnology, as well as being considered proxies for possible life forms on ice-covered
extraterrestrial bodies.
by Sir Ernest Shackleton. "With eighty-eight illustrations and diagrams."
During the Austral summer of 2001/02 exceptionally heavy sea ice conditions were experienced over the eastern Weddell Sea. Satellite microwave imagery showed that large negative (positive) ice anomalies were present from October 2001 to January 2002 over the north-western Weddell Sea (off the coast of Dronning Maud Land). These were a result of anomalously high (low) atmospheric pressure over the South Atlantic (southern Weddell Sea and Bellingshausen Sea), which gave strong north to north-westerly cyclonic flow over the northern and eastern Weddell Sea. This resulted in convergence of sea ice into the southern Weddell Sea and inhibited ice advection along the coast of Dronning Maud land. The atmospheric anomalies around the Weddell Sea were part of an Antarctic-wide amplification of the mean wavenumber 3 pattern resulting in more intrusions of mid-latitude air into the interior of the continent, giving rise to near-record warm temperatures at several locations.
The Self-Organising Map (SOM) algorithm was introduced by the author in 1981. Its theory and many applications form one of the major approaches to the contemporary artificial neural networks field, and new technologies have already been based on it. The most important practical applications are in exploratory data analysis, pattern recognition, speech analysis, robotics, industrial and medical diagnostics, instrumentation, and control, and literally hundreds of other tasks. In this monograph the mathematical preliminaries, background, basic ideas, and implications are expounded in a manner which is accessible without prior expert knowledge.
A climatology of Antarctic sea ice extent based on 20 years of data (1973-1992) is presented, including measures of interannual variability and extrema. We have used this dataset to explore the possible links between Antarctic sea ice extent and the Southern Oscillation Index (SOI). To do this we have caculated their correlation for all pairings of calendar months, as well as with the SOI taken from the year before and subsequent to that of the time of the sea ice data. Guided by the results and hypotheses of earlier authors, we examined the possibility of links between the SOI and sea ice in four key sectors around the Antarctic coast, namely, the southwest Indian Ocean, the southwest and southeast Pacific Ocean, and a sector to the west of the Ross Sea. -from Authors
The sea ice surface, not open water, is the dominant source of sea salt to aerosol and ice cores in coastal Antarctica. Here, we show that it may also form the dominant source for central Antarctica. We can then explain higher concentrations in the winter and last glacial maximum (LGM) as being due to increased sea ice production. This suggests that ice core sea salt can indicate at least the timing of changes in Antarctic sea ice production. The pattern of sea salt in ice cores is consistent with marine evidence about sea ice changes in the Holocene and LGM. Sea salt shows no change at the initial CO2 increase out of the last glacial, making it unlikely this was primarily due to changing sea ice cover. The sea salt record should not be treated as an indicator of meridional transport.
Ice cores have, in recent decades, produced a wealth of palaeoclimatic insights over widely ranging temporal and spatial scales. Nonetheless, interpretation of ice-core-based climate proxies is still problematic due to a variety of issues unrelated to the quality of the ice-core data. Instead, many of these problems are related to our poor understanding of key transfer functions that link the atmosphere to the ice. This study uses two tools from the field of artificial neural networks (ANNs) to investigate the relationship between the atmosphere and surface records of climate in West Antarctica. The first, self-organizing maps (SOMs), provides an unsupervised classification of variables from the mid-troposphere (700 hPa temperature, geopotential height and specific humidity) into groups of similar synoptic patterns. An SOM-based climatology at annual resolution (to match ice-core data) has been developed for the period 1979–93 based on the European Centre for Medium-Range Weather Forecasts (ECMWF) 15-year reanalysis (ERA-15) dataset. This analysis produced a robust mapping of years to annual-average synoptic conditions as generalized atmospheric patterns or states. Feed-forward ANNs, our second ANN-based tool, were then used to upscale from surface data to the SOM-based classifications, thereby relating the surface sampling of the atmosphere to the large-scale circulation of the mid-troposphere. Two recorders of surface climate were used in this step: automatic weather stations (AWSs) and ice cores. Six AWS sites provided 15 years of near-surface temperature and pressure data. Four ice-core sites provided 40 years of annual accumulation and major ion chemistry. Although the ANN training methodology was properly designed and followed standard principles, limited training data and noise in the ice-core data reduced the effectiveness of the upscaling predictions. Despite these shortcomings, which might be expected to preclude successful analyses, we find that the combined techniques do allow ice-core reconstruction of annual-average synoptic conditions with some skill. We thus consider the ANN-based approach to upscaling to be a useful tool, but one that would benefit from additional training data. Copyright © 2005 Royal Meteorological Society.
South: the story of Shackleton's last expedition 1914-1917 / Sir Ernest Shackleton Note: The University of Adelaide Library eBooks @ Adelaide.
A 30-year satellite record of sea ice extents derived mostly from satellite microwave radiometer observations reveals that the Arctic sea ice extent decreased by 0.30+0.03 x 10(exp 6) square kilometers per 10 yr from 1972 through 2002, but by 0.36 plus or minus 0.05 x 10(exp 6) square kilometers per 10yr from 1979 through 2002, indicating an acceleration of 20% in the rate of decrease. In contrast, the Antarctic sea ice extent decreased dramatically over the period 1973-1977, then gradually increased. Over the full 30-year period, the Antarctic ice extent decreased by 0.15 plus or minus 0.08 x 10(exp 6) square kilometers per 10 yr. The trend reversal is attributed to a large positive anomaly in Antarctic sea ice extent in the early 1970's, an anomaly that apparently began in the late 1960's, as observed in early visible and infrared satellite images.
An algorithm for the analysis of multivariate data is presented along with some experimental results. The algorithm is based upon a point mapping of N L-dimensional vectors from the L-space to a lower-dimensional space such that the inherent data "structure" is approximately preserved.
Arctic Climate Impact Assessment (ACIA) Arctic climate impact assessment: scientific report
REFERENCES
Arctic Climate Impact Assessment (ACIA). 2004. Arctic climate
impact assessment: scientific report. Cambridge, etc., Cambridge University Press.
Climate change 2001: the scientific basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change
- J Houghton
Houghton, J.T. and 7 others, eds. 2001. Climate change 2001: the
scientific basis. Contribution of Working Group I to the Third
Assessment Report of the Intergovernmental Panel on Climate
Change. Cambridge, etc., Cambridge University Press.
Interpretation of recent Antarctic sea ice variability
- J Liu
- J A Curry
- D G Martinson
Liu, J., J.A. Curry and D.G. Martinson. 2004. Interpretation of recent
Antarctic sea ice variability. Geophys. Res. Lett., 31(2), L02205.
(10.1029/2003GL018732.)