Short Scale Variability Research Articles

Current interaction in large-scale wave models with an application to Ireland

A nested large-scale wave model that covers most of the Atlantic Ocean and focuses on generating accurate swell conditions for Ireland is presented and validated over the two years 2016−2017. A sensitivity analysis for the βmax parameter appearing in the wind input formulation is carried out, optimising the extreme significant wave heights from the model against altimetry data in the North-East Atlantic region around Ireland. An optimal value of 1.75 is found using the ERA5 reanalysis corrected beforehand for a well-known negative bias for the high wind speed values. The impact of currents is studied using the surface currents from the GLORYS12 product. Currents slightly reduce by 1% the root-mean-square and bias errors for significant wave height compared to altimetry data. However, they explain most of the wave energy at scales less than 50m. The variability induced by currents is found to be more noticeable when looking at the satellite tracks. Track following observations indicate that wave refraction and advection induced by mesoscale eddies are correctly captured by the model, inducing a spatial variability reaching up to 0.5m. However, this modulation is of the same order as a shorter-scale variability appearing in the altimetry data. An averaging operator is used to smooth the data and remove those high-frequency variations. The main finding is that a spatial shift seems to appear between the model and the observations, attributed to the relatively poor resolution of the current product used for the scales observed.

MAVEN/NGIMS wind observations in the martian thermosphere during the 2018 planet encircling dust event

During the 2018 planet encircling dust event (PEDE) at Mars, the Mars Atmosphere and Volatile Evolution (MAVEN) spacecraft’s Neutral Gas and Ion Mass Spectrometer (NGIMS) was able to take a series of neutral wind observations in the thermosphere at around 150–220 km altitude. These MAVEN observations are the first in-situ measurements of thermospheric wind velocities during a Mars global dust storm. From these measurements, it was found that the observed averaged horizontal thermospheric wind speeds peaked to nearly 200 m/s during the onset of the dust storm before decreasing to about 140 m/s near the peak of the storm. The maximum averaged wind speed observed during the dust storm is not significantly greater than averaged wind speeds at other times, but does represent a local maximum that coincides with the timing of the dust storm. Additionally, throughout the mature and early decay stage of the PEDE, NGIMS wind observations show increased short-scale variability that distinguishes this time period from most other observational periods.

Using GPR to analyze regeneration success of cork oaks in the Maâmora forest (Morocco)

Under limited precipitation, coarse-textured soils may strongly limit tree regeneration success and stand productivity, yet their effective impact depends on layer thickness and the nature of the underlying soil material. Within that context, we have evaluated the performance of two ground-penetrating radar (GPR) techniques to assess the short-scale variability in physical soil properties, using the Maâmora forest and cork oak as a case study. A frequency-domain GPR was developed to map the soil surface water content of a 1 ha, 18-year old cork oak stand. In addition, a time-domain GPR was used to detect soil textural transitions in depth between the sand and clay textures of the same cork oak stand. Intensive GPR measurements were made along transects, together with detailed local stand and soil characterization. The volumetric water contents (ϴv) varied between 0.01 m³/m³ and 0.1 m³/m³, which represent available water storage of coarse sands to sandy loams [0.04 m³/m³ - 0.12 m³/m³]. The estimated transition depths detected by the second GPR varied between 1.8 m and 4.0 m deep. Soil water content was found to be a relevant indicator of successful cork oak regeneration as it was strongly related to the dominant heights of the cork oaks. The most successful areas were characterized by higher surface moisture and these higher values were mostly explained by soil texture. Moreover, the layer thickness also influenced cork oak growth. Hence, both radar systems provided useful predictors for the regeneration success of cork oaks. Using GPR for future tree planting would allow foresters to detect the best areas in terms of hydric and edaphic conditions where trees should be installed in order to ensure a successful establishment.

Surface soil clay content mapping at large scales using multispectral (VNIR–SWIR) ASTER data

ABSTRACT The potential of Visible Near-Infrared and Short-Wave Infrared (VNIR–SWIR, 400 nm–2500 nm) hyperspectral imagery for use in multivariate approaches and geostatistical techniques for mapping topsoil properties has been previously demonstrated. However, the use of VNIR–SWIR hyperspectral imagery remains costly, which limits the spatial scales over which it can be applied. This paper aims to evaluate the potential for substituting the more accessible Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) VNIR–SWIR multispectral data for hyperspectral imagery in mapping surface soil clay contents. This study used ASTER multispectral data (nine bands in the VNIR–SWIR spectral domain) acquired over the Cap-Bon region in Tunisia (2000 km2) and 262 surface soil samples collected within the ASTER scene that were subjected to laboratory analysis of the clay fraction (soil particles less than 2 µm). The approach followed two steps: i) estimation of surface soil clay contents for bare soil areas using a Multiple Linear Regression (MLR) model built from the 9 ASTER VNIR–SWIR bands and ii) spatial interpolation (co-kriging) of the soil sampling of measured points and the ASTER-estimates over the whole study area. The MLR model for estimating clay contents using ASTER multispectral data performed correctly ( = 0.60). In addition, this performance is only slightly lower than that obtained using hyperspectral imagery (specifically, an Airborne Imaging Spectrometer for Applications (AISA-DUAL) dual hyperspectral sensor) in a previous study. Moreover, the co-kriging process appeared to yield encouraging results for capturing the large range of variability of clay content values, although it was not able to represent the short scale variability ( = 0.43). Finally, the ASTER multispectral data, despite being underused in the mapping of soil properties, may open up new ways to perform more extensive mapping of surface soil properties in semi-arid contexts characterized by extensive bare and dry soil surfaces.

Primary productivity and phytoplankton dynamics in a subtropical estuary: a multiple timescale approach

Temporal variability of physicochemical parameters and phytoplankton primary productivity, abundance and composition were investigated at Babitonga Bay, southern Brazil, using a hierarchical sampling design on the scales of months, weeks and days during spring 2012 and summer 2013. Only temperature, respiration rates and concentrations of dissolved oxygen and silicate exhibited significant differences in the greatest timescale (seasons: spring; summer). In contrast, most physicochemical parameters, such as salinity and the concentration of nitrogen compounds, varied mainly among weeks and days. This short-scale variability was similarly observed for the microphytoplankton abundance, ranging from 0.04 to 1.7.106 cells L–1 during a bloom of the diatom Skeletonema costatum. Two major phytoplankton assemblages were associated with high primary production rates: > 30 μm centric diatoms such as Cyclotella spp. and Cymatodiscus sp., in spring; and a summer assemblage dominated by smaller, bloom-forming diatoms (S. costatum and Eucampia cornuta). Primary production ranged from 10.5 to 1793 mg C m–2 d–1 and varied significantly between days and months, being associated with the photosynthetic active radiation level and weather conditions on the sampling day. Abundance of specific plankton taxa appeared controlled by trophic interactions, as revealed for the mixotrophic, toxic dinoflagellate Dinophysis acuminata complex and its prey, the ciliate Mesodinium rubrum.

Soil organic carbon stock explored by geo-spatial modeling at Sagar Island, India

Soil organic carbon (SOC) and organic carbon density (OCD) is an active indicator of soil fertility and productivity, and it fluctuates spatially and temporally. The main objective of the present study is to evaluate the spatial variability of SOC and OCD using geostatistics model as well as field methods at Sagar Island (Sundarban), India. Soil samples at the depths of 0–20 cm were collected from twenty plots at different season to estimate organic carbon concentrations and carbon density in the laboratory. Spatial distributions of SOC and OCD concentrations were estimated using ordinary kriging method. Remote sensing imagery and ground truth helped in identifying different features as well as land use/land cover correctly. Landuse are characterized by coastal water, sand, coastal wetland, salt marsh, agricultural fallow land, sparse vegetation, crop/paddy field, and mixed settlement area. The average value SOC were recorded as 2.80, 1.83, and 2.069 % in post-monsoon, pre-monsoon and monsoon season respectively. Semivariograms models of SOC and OCD in all seasons illustrated strong positive nugget values indicated short scale variability and sampling error. The coefficient of determination (R2) of the SOC model denoted as 0.71, 0.68 and 0.74 and for OCD calculated as 0.78, 0.67 and 0.83 at pre-monsoon, monsoon and post-monsoon season respectively. The results illustrated that mangrove forest portrayed highest amount of SOC concentration in both the pre-monsoon (1.05 kg/m2) and post-monsoon (0.94 kg/m2) season.

High-resolution simulation of the circulation in the Bahamas and Turks and Caicos Archipelagos

We present the first high-resolution simulation of the Bahamas and Turks and Caicos archipelagos implemented with the Regional Oceanic Modeling System (ROMS). Our model resolves tides and our simulation consists of two nested grids of respectively 4 and 2km horizontal resolution. This oceanic region is constituted of the juxtaposition of very large and shallow banks, and very deep basins. Our modeling study shows that the dominant influence is the salinity from the sub-tropical gyre, which varies latitudinally with seasons. Less salty surface waters entered the archipelago from September to November from the north and through the Old Bahama (OBC) and the North West Providence (NWPC) channels as the western Atlantic saline waters moved south. Saltier Atlantic waters arrived from the south in December and moved north till the next fall.This salinity flux is mitigated by the atmospheric forcing on the banks as well as the water temperature, which varied diurnally, what modified the water density causing this water to be unique to the banks, yielding its own seasonality as seen in their SST. The banks were colder in winter and warmer in summer than their surrounding waters. Fronts were formed at every ebb tide in winter, spring, and fall. However the influence of bank waters on the surrounding waters was shown to be limited because of the dominant influence of the large-scale flow and of the local production of water masses by wind and cyclone driven upwelling. Bank waters were contained at the front where they mixed with the surrounding waters as shown by the short time presence of the fronts. Local wind, in addition to the tide was also shown to be a significant forcing which is responsible for some of the short scale variability 5–30days in Florida Current transport, the NWPC, the OBC, the Windward Passage (WP), and the Antilles Current (AC). Finally, the study also confirmed the mesoscale eddy influence at seasonal scales (100days) on the AC transport as well as on the FC, through a possible leakage of eddy flux through the NWPC. A similar influence was found south of the archipelago, where mesoscale eddy could also influence WP transport.

Influence of estimation method of compression index on spatial distribution of consolidation settlement in Songdo New City

This paper describes the influence of the method that estimates the compression index (Cc) on the spatial distribution of consolidation settlement (sc) for Songdo New City. The spatial distribution of consolidation settlement can be evaluated by using the spatial distribution (Case 1), mean value (Case 2), and probability distribution (Case 3) of soil properties. For Case 1, ordinary cokriging was adopted to estimate the spatial distribution of Cc as it provides more reliable estimates than ordinary kriging. It is observed that the spatial distribution of sc for Case 1 has significantly shorter scale variability than that for Case 2 because the short scale variability of the spatial distribution of Cc and void ratio (eo) affect the spatial distribution of sc for Case 1. It is also shown that the ratio of the area of sc>100cm (design criterion) to the total area for Case 1 (7.7 %) is larger than that for Case 2 (0.5 %) because the ordinary cokriging estimate of Cc is larger than the mean value of Cc in some regions. The probabilistic analysis (Case 3) shows that the ratio of the area of P(sc>100cm)>probabilistic criterion (α) to the total area increases as the coefficient of variation (COV) of Cc/(1+e0) increases and α decreases. For Songdo New City, the area ratio for the probabilistic design criterion (α) of 0.30–0.45 is found to be in the range of 2.9–19.4 %.

The combined effect of wave–current interaction and mud-induced damping on nonlinear wave evolution

The development of a phase-resolving nonlinear frequency-domain model with both wave–current interaction and viscous mud-induced energy dissipation is discussed. The model is compared to dissipation rates deduced from experimental data, with favorable results. The model is then run with cnoidal waves over a finite mud patch with both opposing and following currents. It is determined that wave height dissipation by mud is exacerbated by opposing currents and reduced by following currents, in agreement with previous work. It is shown that mud-induced damping affects the cnoidal wave shape; under significant damping, the resulting waveform resembles a sine wave, with some short-scale variability as phase-locking between the harmonics breaks down. In addition, the effect of uncompensated subharmonic interactions, a cause of high frequency damping over mud, is also evident with wave–current interaction. Finally, random wave spectra are used to initialize the model and allowed to evolve over a flat bottom with a mud patch, with and without co-flowing currents. As before, the dissipation of the random waves is enhanced by opposing currents and reduced by following currents. The degree of spectral broadening seen in wave–current interaction in non-dissipative environments is also seen here with mud-induced dissipation. High spectral frequencies strongly damped by bottom mud recover some energy (at the expense of low frequencies) in the lee of the finite mud patch. This recovery is evident even with substantial damping across the majority of the frequency range of the spectrum.

Bayesian estimation of basal conditions on Rutford Ice Stream, West Antarctica, from surface data

AbstractThe determination of basal properties on ice streams from surface data is formulated as a Bayesian statistical inference problem. The theory is applied to a flowline on Rutford Ice Stream, West Antarctica. Estimates of bed topography and basal slipperiness are updated using measurements of surface topography and the horizontal and vertical components of the surface velocity. The surface topography is allowed to vary within measurement errors. We calculate the transient evolution of the surface until rates of surface elevation change are within limits given by measurements. For our final estimation of basal properties, modelled rates of elevation change are in full agreement with estimates of surface elevation changes. Results are discarded from a section of the flowline where the distribution of surface residuals is not consistent with error estimates. Apart from a general increase in basal slipperiness toward the grounding line, we find no evidence for any spatial variations in basal slipperiness. In particular, we find that short-scale variability (<10 × ice thickness) in surface topography and surface velocities can be reproduced by the model by variations in basal topography only. Assuming steady-state conditions, an almost perfect agreement is found between modelled and measured surface geometry, suggesting that Rutford Ice Stream is currently close to a steady state.

XMM-NewtonRGS observation of the warm absorber in Mrk 279

The Seyfert 1 galaxy Mrk 279 was observed by XMM-Newton in November 2005 in three consecutive orbits, showing significant short-scale variability (average soft band variation in flux ~20%). The source is known to host a two-component warm absorber with distinct ionisation states from a previous Chandra observation. We aim to study the warm absorber in Mrk 279 and investigate any possible response to the short-term variations of the ionising flux, and to assess whether it has varied on a long-term time scale with respect to the Chandra observation. We find no significant changes in the warm absorber on neither short time scales (~2 days) nor at longer time scales (two and a half years), as the variations in the ionic column densities of the most relevant elements are below the 90% confidence level. The variations could still be present but are statistically undetected given the signal-to-noise ratio of the data. Starting from reasonable standard assumptions we estimate the location of the absorbing gas, which is likely to be associated with the putative dusty torus rather than with the Broad Line Region if the outflowing gas is moving at the escape velocity or larger.

Symmetric-embedding prediction of the CATS benchmark

We present a general strategy for filling the missing data of the CATS benchmark time series prediction competition. Our approach builds upon a time-symmetric embedding of this time series and the use of a one-shot forecasting for each missing value inside the gaps from distant-enough delayed and forwarded predictors. In the extrapolation region we perform standard, non-iterated forward predictions. For modeling purposes we consider bagging of multi-layer perceptrons (MLPs). We discuss two different implementations of this strategy: The first one is based on a simultaneous modeling of both large- and short-scale dynamics information, using (suitably delayed and forwarded) original CATS values and their first differences as inputs to MLPs. The second one follows a two-stage strategy, in which behaviors at different scales are modeled separately. First, the overall behavior at large scales is fitted with a smooth curve obtained by repeated application of a Savitzky-Golay filter. Then, the remaining short-scale variability is approximated using bagged MLPs. Expected error levels for these two implementations are provided according to performance on test data.

A Comparison of Snow Telemetry and Snow Course Measurements in the Colorado River Basin

Abstract Temporal and spatial differences in snow-water equivalent (SWE) at 240 snow telemetry (SNOTEL) and at 500 snow course sites and a subset of 93 collocated sites were evaluated by examining the correlation of site values over the snow season, interpolating point measurements to basin volumes using hypsometry and a maximum snow extent mask, and variogram analysis. The lowest correlation at a point (r = 0.79) and largest interpolated volume differences (as much as 150 mm of SWE over the Gunnison basin) occurred during wet years (e.g., 1993). Interpolation SWE values based on SNOTEL versus snow course sites were not consistently higher or lower relative to each other. Interpolation rmse was comparable for both datasets, increasing later in the snow season. Snow courses correlate over larger distances and have less short-scale variability than SNOTEL sites, making them more regionally representative. Using both datasets in hydrologic models will provide a range of predicted streamflow, which is potentially useful for water resources management.

Multi-scale variability in tropical soil nutrients following land-cover change

The effects of land-cover change on soil properties have reached regional proportions in the southwestern Amazon, and can be detected in their patterns at the scale of hundreds of thousands of square kilometers. Spatial analysis of an extensive soil profile database revealed four nested spatial scales at which different patterns in soil properties occur: 68 km. The shortest scales account for the majority of the spatial variability in the biogeochemically important properties (phosphorus, nitrogen, organic carbon), while the longest scales explain the most for pedologic properties (pH, aluminum, percent clay and sand). The magnitude and sign of correlations among soil properties changed with scale, indicating major shifts in distribution and soil dynamics depending on the scale of observation and analysis. Precipitation, substrate composition, topography, and biological influences vary over large areas, leading to variable distribution of soil properties, and complicating the evaluation of the effect of deforestation on biogeochemical cycling. We modeled the relationships among scale-specific patterns in soil properties and the soil forming state factors to tease out the relative impact of changing land cover. Soil property short-scale variability corresponded with the distribution of land cover and terrain attributes, medium scale variation with geology and state soil map classifications, and long scale with geology and precipitation. The strength of these relationships may be partially attributable to the resolution of the maps used as proxies for the soil forming factors. Land-cover change has already left a discernable imprint on broad-scale soil nutrient patterns, although it is still not the dominant process.

BME-based uncertainty assessment of the Chernobyl fallout

The vast territories that have been radioactively contaminated during the 1986 Chernobyl accident provide a substantial data set of radioactive monitoring data, which can be used for the verification and testing of the different spatial estimation (prediction) methods involved in risk assessment studies. Using the Chernobyl data set for such a purpose is motivated by its heterogeneous spatial structure (the data are characterized by large-scale correlations, short-scale variability, spotty features, etc.). The present work is concerned with the application of the Bayesian Maximum Entropy (BME) method to estimate the extent and the magnitude of the radioactive soil contamination by 137Cs due to the Chernobyl fallout. The powerful BME method allows rigorous incorporation of a wide variety of knowledge bases into the spatial estimation procedure leading to informative contamination maps. Exact measurements (“hard” data) are combined with secondary information on local uncertainties (treated as “soft” data) to generate science-based uncertainty assessment of soil contamination estimates at unsampled locations. BME describes uncertainty in terms of the posterior probability distributions generated across space, whereas no assumption about the underlying distribution is made and non-linear estimators are automatically incorporated. Traditional estimation variances based on the assumption of an underlying Gaussian distribution (analogous, e.g., to the kriging variance) can be derived as a special case of the BME uncertainty analysis. The BME estimates obtained using hard and soft data are compared with the BME estimates obtained using only hard data. The comparison involves both the accuracy of the estimation maps using the exact data and the assessment of the associated uncertainty using repeated measurements. Furthermore, a comparison of the spatial estimation accuracy obtained by the two methods was carried out using a validation data set of hard data. Finally, a separate uncertainty analysis was conducted that evaluated the ability of the posterior probabilities to reproduce the distribution of the raw repeated measurements available in certain populated sites. The analysis provides an illustration of the improvement in mapping accuracy obtained by adding soft data to the existing hard data and, in general, demonstrates that the BME method performs well both in terms of estimation accuracy as well as in terms estimation error assessment, which are both useful features for the Chernobyl fallout study.

Geostatistical Ground Water Monitoring of a Point Source Plume Entering a Restored Riparian Zone

concentrations associated with a point-source ground water plume were monitored as the plume entered a restored riparian zone (RRZ) on the University of Idaho campus in Moscow, Idaho. Seasonal data collected for a network of piezometers installed in the RRZ were analyzed geostatistically and then modeled spatially using the conditional simulation method of sequential Gaussian simulation (SGS). SGS was utilized to predict the spatial distribution of ground water concentrations at unsampled locations in the RRZ and to quantify the uncertainty associated with the prediction. Maps prepared using SGS results illustrate the short-scale variability, or patchiness, expected of concentration distributions in a riparian zone. Manipulation of SGS output provided graphical and quantitative estimates of the likelihood of exceeding a specified concentration threshold at a given confidence level for any location within the RRZ. Geostatistical simulation tools for quantifying uncertainty also provide a potential risk assessment methodology for making remediation decisions and reducing remediation costs.

Variability of the baroclinic Rossby radius in the Gulf of Finland

The internal Rossby radius, R 1, and its variability over the Gulf of Finland has been determined on the basis of 1911 hydrographic soundings made during the years 1993–1999. It was found that R 1 has a large spatial and temporal variability. The largest values of R 1 were observed in summer in deep parts of the Gulf. In shallow coastal regions and in the eastern Gulf of Finland, R 1 is smaller than in the open deep areas. Average values of R 1 were 2.1 km in shallow areas (depth <36 m) and 4.1 km in the open sea. The smallest R 1 values were of the order of some hundreds of metres, whilst the largest were over 7 km. Typical values (approximately 70% of all the observations) of R 1 in the Gulf of Finland fall within the range 2–4 km; which was less than half of the corresponding values in the Gotland Deep, where reference data provided an average value of 10.9 km. The seasonal variability of R 1 was also pronounced. The largest values occur in late summer, when the temperature stratification is at its strongest. Short-scale variability of R 1 on a time-scale of some hours was also identified.

Changes in a tidepool fish assemblage on two scales of environmental variation: Seasonal and El Niño Southern Oscillation

Intertidal organisms are influenced by the tidal, daily, and seasonal environmental variability of their habitat. Interannual variability, although often less severe than shorter‐scale variability, may also be important in structuring intertidal systems. This study compares the magnitude of changes in a rocky intertidal fish guild occurring on a seasonal scale with those occurring during an El Niño Southern Oscillation (ENSO). I examined tidepool fish assemblage structure and habitat use in two southern California sites approximately every 3 months from 1996 to 2000, a period including non‐ENSO conditions, the 1997–1998 El Niño, and the 1998–1999 La Niña. During each sampling period, I censused fish abundance in 105 tidepools of differing intertidal height, depth, and surface area. Several aspects of habitat use varied seasonally for the four most common species: Clinocottus analis (woolly sculpin), Girella nigricans (opaleye), Gobiesox rhessodon (California clingfish), and Hypsoblennius gilberti (notch‐brow blenny). All four species migrated vertically within the intertidal zone on a seasonal scale, corresponding to seasonal changes in sea level. The assemblage dominant, C. analis, occupied tidepools of different sizes depending on season. Although seasonality in habitat use suggests an influence of environmental variability on seasonal scales, fish habitat was generally not altered by temperature and sea level changes imposed by the El Niño. Species assemblage, however, differed among climate conditions. C. analis declined in abundance during the El Niño because of lack of recruitment but increased immediately after its conclusion. Paraclinus integripinnis (reef finspot), usually rare, was more abundant during the El Niño. Effect of the El Niño on the other four species was not detected. Assemblage changes suggest that although intertidal fishes regularly experience large tidal, daily, and seasonal environmental fluctuations, interannual changes in environmental factors, even when relatively small in magnitude, can perturb the system. Perturbations in the present system, however, did not persist beyond the end of the El Niño event as they often do in lower‐latitude nearshore areas.

Faraday resonance in thin sedimentary layers

SUMMARY It is well established that earthquake waves passing through soft alluvial valleys and sedimentary basins will generally amplify, so that signals are stronger, longer and more complex on such sites than on adjacent crystalline rocks. One anomalous feature of the amplification there is the rather significant variability in the recorded strong ground motion, especially over relatively short length-scales. This paper explains such short-scale variability in the amplification factor by proposing a novel seismic resonance mechanism in thin flat layers. The variability is attributed to the subharmonic generation of standing elastic waves on the surface of the deposit. Nodal regions of little horizontal shaking, or little destruction, are separated by antinodal regions of strong horizontal shaking, i.e. significant damage. Vertical ground shaking follows a similar pattern. The subharmonic surface waves are forced by the vertical component of an input seismic shaking at the base of the deposit, through a linear instability mechanism in the layer. A perturbation theory is worked out to describe the conditions for resonance, and a small shaking-tank experiment is conducted to confirm some of the theoretical findings.

From Short-Scale Atmospheric Variability to Global Climate Dynamics: Toward a Systematic Theory of Averaging

Abstract Traditionally, climate is defined by the properties of the averages of the meteorological fields over an appropriate time interval. In this paper the properties of the time-averaged observables of a red noise atmosphere and of a simplified model of thermal convection are investigated both analytically and numerically and are compared to those of the original finescale variables. It is shown that averaging tends to reduce the domain of variability and the attractor dimension favors persistence of initial correlations and enhances predictability. The implications of these findings in the real-world atmosphere are briefly assessed.