Sea Level Pressure Time Series Research Articles

Fuzzy time series clustering using autoencoders neural network

<p>This study presents a novel approach that employs autoencoders (AE)—an artificial neural network—for the nonlinear transformation of time series to a compact latent space for efficient fuzzy clustering. The method was tested on atmospheric sea level pressure (SLP) data towards fuzzy clustering of atmospheric circulation types (CTs). CTs are a group of dates with a similar recurrent SLP spatial pattern. The analysis aimed to explore the effectiveness of AE in producing and improving the characterization of known CTs (i.e., recurrent SLP patterns) derived from traditional linear models like principal component analysis (PCA). After applying both PCA and AE for the linear and nonlinear transformation of the SLP time series, respectively, followed by a fuzzy clustering of the daily SLP time series from each technique, the resulting CTs generated by each method were compared to assess consistency. The findings reveal consistency between the SLP spatial patterns from the two methods, with 58% of the patterns showing congruence matches greater than 0.94. However, when examining the correctly classified dates (i.e., the true positives) using a threshold of 0.8 for the congruence coefficient between the spatial composite map representing the CT and the dates grouped under the CT, AE outperformed PCA with an average improvement of 29.2%. Hence, given AE's flexibility and capacity to model complex nonlinear relationships, this study suggests that AE could be a potent tool for enhancing fuzzy time series clustering, given its capability to facilitate the correct identification of dates when a given CT occurred and assigning the dates to the associated CT.</p>

A Teleconnection between the West Siberian Plain and the ENSO Region

The Walker circulation is linked to extratropical waves that are deflected from the Northern Hemisphere polar regions and travel southeastward over central Asia toward the western Pacific warm pool during northern winter. The wave pattern resembles the east Atlantic–west Russia pattern and influences the El Niño–Southern Oscillation (ENSO) region. A tripole pattern between the West Siberian Plain and the two centers of action of ENSO indicates that the background state of ENSO with respect to global sea level pressure (SLP) has a significant negative correlation to the West Siberian Plain. The correlation with the background state, which is defined by the sum of the two centers of action of ENSO, is higher than each of the pairwise correlations with either of the ENSO centers alone. The centers are defined with a clustering algorithm that detects regions with similar characteristics. The normalized monthly SLP time series for the two centers of ENSO (around Darwin, Australia, and Tahiti) are area averaged, and the sum of both regions is considered as the background state of ENSO. This wave train can be detected throughout the troposphere and the lower stratosphere. Its origins can be traced back to Rossby wave activity triggered by convection over the subtropical North Atlantic that emanates wave activity toward the West Siberian Plain. The same wave train also propagates to the central Pacific Ocean around Tahiti and can be used to predict the background state over the ENSO region. This background state also modifies the subtropical bridge between tropical eastern Pacific and subtropical North Atlantic leading to a circumglobal wave train.

A daily Azores-Iceland North Atlantic Oscillation index back to 1850.

We present the construction of a continuous, daily (09:00 UTC), station‐based (Azores–Iceland) North Atlantic Oscillation (NAO) Index back to 1871 which is extended back to 1850 with additional daily mean data. The constructed index more than doubles the length of previously existing, widely available, daily NAO time series. The index is created using entirely observational sea‐level pressure (SLP) data from Iceland and 73.5% of observational SLP data from the Azores – the remainder being filled in via reanalysis (Twentieth Century Reanalysis Project and European Mean Sea Level Pressure) SLP data. Icelandic data are taken from the Southwest Iceland pressure series. We construct and document a new Ponta Delgada SLP time series based on recently digitized and newly available data that extend back to 1872. The Ponta Delgada time series is created by splicing together several fractured records (from Ponta Delgada, Lajes, and Santa Maria) and filling in the major gaps (pre‐1872, 1888–1905, and 1940–1941) and occasional days (145) with reanalysis data. Further homogeneity corrections are applied to the Azores record, and the daily (09:00 UTC) NAO index is then calculated. The resulting index, with its extended temporal length and daily resolution, is the first reconstruction of daily NAO back into the 19th Century and therefore is useful for researchers across multiple disciplines.

North Sea near-surface wind climate and its relation to the large-scale circulation patterns

The North Sea 10-m wind speed (WS10) climate is compared and related to circulation patterns based on the sea level pressure (SLP) extracted from three reanalysis and one high-resolution model dataset. The mean magnitude and the trends of WS10 depend considerably on the selected reanalysis. The variability of WS10 among the three reanalysis datasets is highly correlated in the recent period (1980–2000) but less so in the past period (1960–1980). The WS10 over the North Sea is well represented by the relatively low reanalysis resolution when compared to the high-resolution WS10 model data partially owing to the high spatial correlation of WS10. Exceptions are observed only at the coastal areas. The dominant mode of WS10 explains coherent variability of WS10 over the North Sea and is related to a SLP pattern similar to the North Atlantic oscillation (NAO). The increase of the magnitude of the dominant WS10 pattern is related to the increase of the magnitude of the NAO-like SLP pattern from 1960s to mid-1990s. The second dominant WS10 pattern—a dipole in WS10 to the north and south of Great Britain—is related to the differences in SLP between Scandinavia and Iceland. The relation between the second WS10 and SLP patterns is more prominent in the recent period. The extreme WS10 in the German Bight is related to the low SLP over Scandinavia. The extreme WS10 is strongly increasing from the early 1980s to the beginning of 1990s, which is not observed in the corresponding SLP time series over Scandinavia.

Fuzzy descriptor systems and spectral analysis for chaotic time series prediction

Predicting future behavior of chaotic time series and systems is a challenging area in the literature of nonlinear systems. The prediction accuracy of chaotic time series is extremely dependent on the model and the learning algorithm. In addition, the generalization property of the proposed models trained by limited observations is of great importance. In the past two decades, singular or descriptor systems and related fuzzy descriptor models have been the subjects of interest due to their many practical applications in modeling complex phenomena. In this study fuzzy descriptor models, as a more recent neurofuzzy realization of locally linear descriptor systems, which have led to the introduction of intuitive incremental learning algorithm that is called Generalized Locally Linear Model Tree algorithm, are implemented in their optimal structure to be compared with several other methods. A simple but efficient technique, based on the error indices of multiple validation sets, is used to optimize the number of neurons as well as to prevent over fitting in the incremental learning algorithms. The aim of the paper is to demonstrate the advantages of fuzzy descriptor models and to make a fair comparison between the most successful neural and neurofuzzy approaches in their best structures according to prediction accuracy, generalization, and computational complexity. The Mackey–Glass time series, Lorenz time series (as two well-known classic benchmarks), Darwin sea level pressure time series and long-term prediction of Disturbance Storm Time index, an important index of geomagnetic activity (as two natural chaotic dynamics) are used as practical examples to evaluate the power of the proposed method in long term prediction of chaotic dynamics.

Interannual variability associated with Semiannual Oscillation in southern high latitudes

The annual and interannual variations of sea level pressure (SLP) over the Southern Ocean are investigated using the NCEP/NCAR reanalysis data from 1968 to 2005. Results show that the weakening of the Semiannual Oscillation (SAO) has continued to present day but does not occur homogeneously across all longitudes. While the second harmonic of SLP has decreased over the South Pacific Ocean, its signal is still strong over the Atlantic sector up to 2005. This suggests that the mechanism influencing the amplitude of the SAO over South Atlantic Ocean is different to that in the Pacific and Indian oceans. Spectral analysis of zonally averaged SLP at 50°S and 65°S shows that significant energy is associated with the period of 4 years. The same period is found for the time series of the zonally averaged position of the circumpolar trough and for the meridional temperature gradient at 500 hPa. This interannual signal is even stronger in the time series of the SLP difference between those two latitudes: the SAO index. Wavelet analysis of the SLP time series at mid and high southern latitudes showed significant interannual energy, also with a 4 year period, at 50°S, 65°S, and in the time series of their difference. When analyzed by ocean basins, spectral analysis reveal that there is no interannual variability at the 4 year period for the Atlantic, indicating that the known weakening of the SAO is basin‐dependent.

Calibrations of Bermuda corals against large‐scale sea surface temperature and sea level pressure pattern time series and implications for climate reconstructions

Proxy records from stony reef corals are established tools for reconstructing climate features. However, validation procedures for proving the calibrations are often insufficient or even lacking. We address this problem by the establishment and validation of statistical regression models for the reconstruction of large‐scale climate patterns from local coral records (upscaling). We apply the upscaling technique to the annual growth rates of two coral species from Bermuda, representing a period of about 140 years. The coral records are related to reanalyzed sea surface temperature (SST) and sea level pressure (SLP) fields. In the western North Atlantic, the correlation maps between the SSTs and the single coral records and their leading principal component show similar structures. On the basis of regression patterns for limited areas, upscaling models are established. Validation of such models for the SST yields an explained variance of 30% of the low‐pass‐filtered data using the first principal component of both corals. For the detrended SLP time series, representing an almost hemispheric SLP pattern, 21% of the interannual variability is explained by means of only one coral. Thus different corals from neighboring sites represent similar climate signals, however, to slightly varying degrees, leading to different reconstruction skills. Problems remain in reconstructing long‐term trends. We conclude that reconstructions from multiple corals are not a priori predominant to those from single corals. A calibration to what extent a single coral represents a climate parameter should be the first step before combining multiple corals as the basis for a reconstruction.

Interpretation of North Pacific Variability as a Short- and Long-Memory Process*

A major difficulty in investigating the nature of interdecadal variability of climatic time series is their shortness. An approach to this problem is through comparison of models. In this paper a first-order autoregressive [AR(1)] model is contrasted with a fractionally differenced (FD) model as applied to the winter-averaged sea level pressure time series for the Aleutian low [the North Pacific (NP) index] and the Sitka winter air temperature record. Both models fit the same number of parameters. The AR(1) model is a “short-memory” model in that it has a rapidly decaying autocovariance sequence, whereas an FD model exhibits “long memory” because its autocovariance sequence decays more slowly. Statistical tests cannot distinguish the superiority of one model over the other when fit with 100 NP or 146 Sitka data points. The FD model does equally well for short-term prediction and has potentially important implications for long-term behavior. In particular, the zero crossings of the FD model tend to be farther apart, so they have more of a “regimelike” character; a quarter century interval between zero crossings is 4 times more likely with the FD than the AR(1) model. The long-memory parameter δ for the FD model can be used as a characterization of regimelike behavior. The estimated δs for the NP index (spanning 100 yr) and the Sitka time series (168 yr) are virtually identical, and their size implies moderate long-memory behavior. Although the NP index and the Sitka series have broadband low-frequency variability and modest long-memory behavior, temporal irregularities in their zero crossings are still prevalent. Comparison of the FD and AR(1) models indicates that regimelike behavior cannot be ruled out for North Pacific processes.

Wind‐forced pressure response of the Gulf of Maine

Bottom pressure measurements, coastal synthetic subsurface pressures, and wind stresses, derived from operational wind measurements throughout the Gulf of Maine, are used to explore the seasonal wind‐forced response of the gulf. Barotropic pressure fluctuations throughout the gulf are highly correlated all year (>0.75) and even more so in the winter (>0.85), with more than 90% of the variance explained by the relevant mode 1 empirical orthogonal functions (EOF). Both the pressures and the wind stress variability are concentrated in the 2‐ to 10‐day “weather band,” with typical winter wind stresses (0, 10 Pa) about 5 times more energetic than those in summer (0.05 Pa). Pressure fluctuations are highly correlated with 255°–75°T wind stress fluctuations. Typical gulf‐scale westward wind stress fluctuations force general gulf‐scale increases in the sea level pressure of about 10 eq cm in winter and about 4 eq cm in summer. The spatially averaged gulf pressure response to westward wind stress is 92±6 eq cm Pa−1 in winter and 83±8 eq cm Pa−1 in summer. The Gulf of Maine wind stress‐forced pressure variability was numerically modeled using the Dartmouth linear, homogeneous, three‐dimensional (3‐D), finite element diagnostic model (FUNDY 5). A series of 36‐hour‐averaged, spatially variable, optimally interpolated operational wind stress patterns was used to force model sea level pressure time series, which looked like observations and explained about 40% of the observed variance. The model provided detailed information on the linkage among the fluctuating near‐westward (eastward) wind stress‐forced surface Ekman transport into (out of) the gulf, the compensatory outflows (inflows) around Georges Bank, and the intensified flow along the western coast of the gulf.

Detecting and Removing Inhomogeneities from Long-Term Monthly Sea Level Pressure Time Series

Abstract A method for the detection of and adjustment for multiple discontinuities in pressure time series is presented. The method relies on determining a reference pressure series using improved interpolation techniques and objectively comparing the observed and reference time series. The ability of the method to locate discontinuities in time and to estimate the magnitude of the discontinuity are examined using randomly generated time series with known properties. The sea level pressure time series for Darwin (WMO 94120) is analyzed, and ten discontinuities are noted for the period 1869–1988. These are compared with previous examinations of the Darwin pressure time series.