Siple Station Research Articles

The excitation of active whistler mode signal paths in the magnetosphere by lightning: Two case studies

In two ∼1 hr case study periods, the properties of whistlers propagating along multiple geomagnetic‐field aligned paths from points of origin in the northern hemisphere were compared to data on the location and intensity of lightning. The whistlers were recorded at the approximately conjugate stations Lake Mistissini, Canada and Siple Station, Antarctica, while the lightning data were acquired by the SUNY‐Albany lightning detection network operating in the eastern United States. In the two studies, which represented times near 0700–0800 LT and relatively quiet magnetospheric conditions, between one quarter and one half of the two‐hop whistlers observed at Lake Mistissini were found to have originated in ground flashes detected by the network. The uncorrelated whistlers are believed to have originated in lightning outside the network viewing area or in undetected ground flashes within the network. However, intracloud flashes within the network area cannot be ruled out as causes of some of the uncorrelated events. We confirm evidence from other workers that lightning can excite ducted whistler paths whose ionospheric endpoints are at ranges up to 2500 km or more from the lightning location. Once in the magnetosphere, the ducted waves were found to “spread” in L value by an interpath coupling process that was operative over the entire L range of detected whistler paths (∼1000 km in the north‐south direction at ionospheric heights in one case and ∼1500 km in the other). In both cases there was a single most active whistler path at L ≃ 4.7, at the high‐L limit of propagation paths within the plasmasphere. In one of the cases, lightning at L ≃ 1.6–1.8 illuminated this path through interpath coupling, after first exciting a path at L ≃ 2.2. In the other case, each of 15 flashes in a storm center at L ≃ 3.5 excited the path directly at a range of roughly 600 km. An approximately linear relation was found between the normalized first stroke peak magnetic field (which is proportional to the peak current) and whistler amplitude observed during a 10‐min period. The detection of whistler waves on a particular magnetospheric path and the relative intensity of the waves in successive whistlers appear to be strongly dependent upon the field strength of the impulsive radio signal from lightning at the point of ionospheric wave injection. Meanwhile, the distribution in space of the multiple paths, the absolute wave levels on the various paths, the duration of propagation in terms of higher order echoes, and the interactions of whistlers and VLF emissions propagating on the same path are found to depend primarily upon magnetospheric propagation conditions. For example, on both days the path transmitting the most whistler wave energy to a ground receiver was the one with entrance point the most distant from the lightning source.

On the possibility of beat excitation of whistler sidebands in the magnetosphere via ponderomotive force

We propose a new mechanism to explain the observed sidebands of whistler waves. The mechanism based on nonlinear interaction has potential application to active experiments in the magnetosphere. The process requires two whistler waves with nearly equal frequency ω1‐ω2 and forward and backward traveling components. They exert a parallel ponderomotive force on the electrons at the difference frequency (ω1‐ω2). The modulation of the plasma density by the nonlinear mixing of these two waves, in conjunction with the oscillatory electron velocity at ω1 and ω2 generates a nonlinear current. This nonlinear current drives the first‐order sidebands. In the presence of low frequency density modulation, the oscillatory velocities at the first‐order sidebands generate second‐order sidebands. The process continues to generate higher order sidebands. The sidebands up to seventh order have been observed in VLF experiment at Siple station. The mechanism is effective for the large amplitude waves.

Longitudinal differences in electron precipitation near L = 4

The differences in the electron precipitation characteristics as seen above Siple Station, Antarctica, and the Kerguelen Islands have been studied. These two sites are both in the southern hemisphere at nearly the same magnetic latitude (L=4). The two stations are at longitudes that place them roughly equal distances east and west of the center of the South Atlantic magnetic anomaly at this L value. With respect to the direction of electron drift, Siple is upstream of the anomaly, and Kerguelen is downstream. The primary data used in the study were counting rates from rocket‐borne, parachute‐deployed, sodium iodide scintillation counters at altitudes of 80 to 30 km and VLF wave data from both ground‐based and rocket‐borne receivers. Nine rocket flights are involved, five from Kerguelen and four from Siple. The flights were made under a range of geomagnetic and VLF activity conditions. The data differ in two major respects. First, the precipitation background at Kerguelen is very low, with high levels of wave activity being required to produce any detectable precipitation, an observation which indicates that there is a range of longitude and/or magnetic local time in which the electron pitch angle diffusion rate is 2 orders of magnitude lower than the long‐term global average rate. Second, X ray microbursts were found to be essentially absent at Kerguelen and very common at Siple. This observation appears to support models of the microburst generation process which predict maximum pitch angle scatterings of only a few tenths of a degree.

Ducted whistler propagation outside the plasmapause

A study has been made of the conditions under which lightning whistlers are observed after propagating along geomagnetic field aligned paths or “ducts” located outside the plasmapause. The study was based in part upon results previously obtained from Antarctic whistler recordings in 1963 and in part upon new data from 45 days of observations at Siple Station, Antarctica (L ≃ 4.3) in 1977 and in 1982. Propagation beyond the plasmapause was found, as expected, to be rare in comparison to propagation within the nearby outer plasmasphere. However, detectable propagation beyond the plasmapause near dawn was found to occur on at least one path on roughly one half of the days studied. The path equatorial radii of propagation tended to cluster in two locations, one at the plasmapause outer edge, and the other in a belt separated from the plasmapause by a region of low activity of order 0.5 RE in extent. The outer edge of the belt was at L ≃ 5.5–6. The probability of whistler detection at any L value outside the plasmapause up to ≃ 6–7 was found to increase with local time across the dayside of the Earth; maximum activity was observed after noon. Whistler activity beyond the plasmapause was least common in the 1800–2400 MLT sector. Spatial variations in lightning source activity may be responsible for large variations with longitude in whistler activity outside the plasmapause and for the many cases in which only a small fraction of whistlers within a given period exhibit evidence of plasmatrough propagation. Ionospheric processes, such as wave damping, defocusing in the mid‐latitude trough, scattering by 10 to 100‐m irregularities, and focusing within 50 to 100‐km blobs, can explain many features of the observed spatial distribution of whistler paths beyond the plasmapause and also the relatively low amplitude of the associated whistlers as compared to plasmasphere events. The large differences in whistler activity between afternoon and premidnight appear to be due to spatial differences in perturbing magnetospheric electric field activity and in electron density in the plasma trough region. The experimental findings serve to make a relatively sharp distinction between the plasmasphere and the region beyond from both the propagation and wave‐particle‐interaction points of view.

VLF wave stimulation experiments in the magnetosphere from Siple Station, Antarctica

The Earth's magnetosphere is host to remarkable very low frequency (VLF) electromagnetic signals of natural origin. One of these, called a whistler, originates in lightning. Others, such as hiss and chorus, originate within the plasma itself. They are important for at least three reasons. First, they reveal the properties of the plasma through which they travel and thus can be used as remote sensing tools. Second, their high intensity and narrow bandwidths indicate the presence of a previously unknown kind of wave particle interaction that converts the kinetic energy of charged particles to coherent electromagnetic radiation. This process is called the coherent wave instability (CWI). Third, energetic charged particles are precipitated into the ionosphere through resonant scattering by these same waves, causing enhanced thermal ionization, X rays, light, and heat. To better understand and use the CWI, controlled VLF signals have been injected into the magnetosphere from Siple Station, Antarctica and received on satellites and near the conjugate point in Quebec, Canada. In addition to reproducing many puzzling natural phenomena, these experiments have provided critical new data on the CWI, laying a foundation for various theories and computer simulations. Key findings are as follows: 1) Coherent VLF signals often exhibit exponential temporal growth (∼30 dB) and saturation at levels estimated to be of order 5 pT. 2) Temporal growth requires that the input signal exceed a threshold that varies widely with time. The probable cause of the growth threshold is in situ background noise that reduces the efficiency of phase bunching by a coherent input signal whose intensity is comparable to the noise level within the frequency band of the interaction (∼100 hz). 3) Narrowband triggered emissions can be entrained by Siple frequency ramps of different slope but of much lower (−20 dB) amplitude. The mechanism of entrainment is not yet understood. 4) For two equal amplitude input waves spaced 20 Hz apart, the temporal growth of each component is almost totally suppressed. For larger spacings, 40–100 Hz, the lower frequency is more suppressed than the upper. For 10 < Δƒ < 100 Hz, unsymmetrical sidebands at integer multiples (up to seventh order) of Δƒ are created, along with subharmonics. The integer sidebands are attributed to emission growth triggered by one beat and suppressed by the next. Taken together, the spectrum of the stimulated sidebands and sub‐harmonics is thus more noise‐like than the transmitted spectrum. 5) Simulated hiss shows coalescence of selected noise wavelets into longer and stronger chorus‐like emissions, suggesting that chorus and hiss originate in the same mechanism. Future objectives of a VLF wave injection facility include (1) new experiments on the physics of wave growth and wave‐induced particle scattering and precipitation, (2) testing of the predictions of theories of VLF wave‐particle interaction, (3) development of new techniques for remote sensing and control of space plasmas using VLF techniques, and (4) improvements in the design and operation of VLF communication and navigation systems.

Rocket investigations of electron precipitation and VLF waves in the Antarctic upper atmosphere

The two Antarctic rocket campaigns conducted by American scientists during the past 10 years are described, and the results of these efforts are reviewed. Although the primary objective of both campaigns, the observation of precipitating electrons triggered by the Siple VLF transmitter, was not achieved, significant advancements in understanding upper atmosphere phenomena were made. Standing wave patterns in the ionosphere were observed for the first time by wave experiments flown aboard Nike‐Tomahawk rockets. The same detectors were able to monitor a continuous signal from the transmitter through the neutral atmosphere and into the ionosphere. This in situ observation of a radio wave traversing a neutral and a plasma environment provided unique data for comparison to theoretical studies of wave propagation. SuperArcas measurements provided two distinct types of energetic electron precipitation data: one type was continuous on the time scale of > 1 min, and the other was in X ray microbursts created by bursts of energetic electrons stopping in the upper atmosphere with durations < 1 s. The burst data were compared to theoretical models of wave‐particle interactions. The observations of continuous electron precipitation with sufficient energy to penetrate the atmosphere to an altitude of 60–80 km were found to be consistent with a model of electron transport from the trapped population to the atmosphere in which gradient and curvature drift in the drift cone plays a principal role. The model requires slow electron pitch angle diffusion of the geomagnetically trapped population and predicts that almost all energetic electron precipitation occurs in the South Atlantic Anomaly. The electron precipitation model is shown to organize all of the SuperArcas data, which include not only measurements at Siple Station, Antarctica, but also those at Kerguelen, another L = 4 location in the southern hemisphere, and at northern locations at higher latitudes. A prediction of the model which involves the effect of the dawn‐to‐dusk convection electric field on energetic electron precipitation is described: diurnal modulation is predicted with the maximum electron precipitation intensity when the South Atlantic Anomaly is at local dawn. A test of this prediction made by SuperArcas flights in 1981 provided support for the slow diffusion model. The SuperArcas results on energetic electron precipitation are shown to disagree with a satellite result at L = 4 which requires fast diffusion. Reasons for the discrepancy between the two data sets are discussed, and steps which could be taken to resolve the apparent controversy are suggested. Finally, the suggestion is made that the two rocket campaigns conducted by American scientists in Antarctica during the past decade be regarded as a feasibility study leading to an active rocket research program in Antarctica for the remainder of the century.

Self‐Consistent Theory of Triggered Whistler Emissions

A kinetic theory of triggered VLF whistler emissions is given that is capable of predicting from a small‐scale numerical implementation the observed emission forms and frequency‐time characteristics. The present paper focuses on the theoretical developments and the explanation of the triggering process, complete with a demonstration of the threshold behavior (sometimes known as the dot‐dash anomaly) and the generation of specific falling frequency emissions whose falling rate compares quite favorably with typical observations made in the controlled experiments based in Siple Station, Antarctica. The theory that gives these results is a fully self‐consistent nonlinear treatment based on kinetic theory and valid in the asymptotic limit when several trapping periods occur within the interaction region. For a typical set of parameters (l = 4.3, n = 400 cm−3, and amplitude BT ∼ 1.6 pT) for the input wave magnetic field there are about seven trapping periods in the triggering signal, and one would expect good results from the asymptotic limit. In this limit the nonlinear dynamics can be reduced to the determination of the time τ that the resonant particles are trapped. The nonlinear currents can be expressed in terms of this function τ by simple integrals over the trapped particles’ perpendicular velocities alone. Most of the features of the emission process can be determined analytically, such as the rate of change of frequency, related to magnetospheric parameters. For more quantitative predictions, such as predictions of amplitude and frequency waveforms, a small numerical code which integrates the nonlinear wave equations is used. The theoretical picture of the triggering mechanism contains the observed threshold behavior wherein short triggering pulses of nominal amplitude BT ∼ 1 pT and duration 100 ms or less cannot generate an emission, whereas those in the range of 200 ms and longer do. Similar sensitivity is found with respect to initial frequency, where in some cases, 5.5‐kHz signals can trigger, but 5.0‐kHz signals cannot. Local saturation yielded gains in the range of 20‐30 dB with initial temporal growth rates in the range 100‐200 dB/s. The emission process requires an inverted population in the perpendicular velocity distribution function, but not necessarily linear instability. A sufficient number of high‐energy electrons are required for the driven currents to offset convection, but provided this is satisfied, a sufficiently long triggering signal will always generate a self‐sustaining emission. Interestingly, however, the self‐sustaining emission does not depend on the number of these high‐energy particles or on the details of the velocity distribution function but only on bulk magnetospheric parameters such as magnetic field, gradient scale length, and plasma density in the plasmapause. Also, the marginal signals generated just above the threshold are always fallers, as is observed. These features have not been explained by existing theories. Unresolved issues include the mechanism of termination, the generation of risers and hooks for more intense emission situations, and the detailed synchronization of the frequency‐time and amplitude waveforms. This last point requires the solution of a peculiar singular two‐way wave equation for the phase which is presently solved in a subsidiary asymptotic limit valid within the interaction region. These unexplained points could well be explained within the present theory by inclusion of the variable resonant velocity in the projection of the electron distribution function.

Structure of VLF whistler mode sideband waves in the magnetosphere

An accurate determination is made of the sideband structure of ducted VLF whistler mode waves transmitted from Siple Station and observed at Lake Mistissini, Quebec. Single‐ and double‐frequency experiments are described. It is shown that sideband spectra can be explained by nonlinear interactions between two or more lines in the magnetosphere, the line intensities required for such interactions being low. If one of the lines is a transmitted carrier, the other line can be as much as 40 dB lower in amplitude, implying that power line radiation (PLR) can be an important factor in sideband generation. It is shown that “single‐line” sidebands are due to interactions between the input wave and harmonics of 60 Hz present within the duct (assumed to be PLR), and that double‐line spectra are affected by their position relative to the same 60‐Hz system of lines. A study of double‐line spectra as a function of line separation and line amplitude ratio is made.

Shallow-Core Analysis and Pit Studies at Siple Station, Antarctica: Implications for Extraction of a 500 Year Proxy Climate Record (Abstract)

Two cores, 302 and 132 m long, were drilled 1.5 km up-wind from Siple Station (75°55'S, 84°15'W) during the 1985–86 austral summer. These cores are expected to contain an annually resolvable 500 year record of atmospheric constituents. The entire length of these cores will be analyzed for insoluble particulate concentrations, liquid conductivity, and oxygen-isotope ratios. Anion concentrations (Cl−, SO42−) will be measured in selected sections. The extraction of a proxy climate history from these cores requires an assessment of the temporal quality and the spatial variability of the preserved record. This paper presents this assessment, which is based upon an extensive investigation of both pit and shallow-core (20 m) records.To complement the two deeper cores, eight 20 m cores were drilled and three pits were sampled. The pits were excavated up-wind of the drill site at distances between 0.5 and 1.0 km. In the central pit, 2.8 m deep, walls A and C (each 1 m wide) were positioned parallel to the prevailing wind, with wall B (4 m wide) perpendicular to the prevailing wind. The stratigraphy of wall B was mapped and continuous vertical profiles of samples were collected for microparticle concentrations (MPC), oxygen-isotopic ratios (δ18O), anion chemistry (AC) and liquid conductivity (LC). Three 20 m cores, each associated with a vertical profile of pit samples, were drilled 0.5 m behind wall B. In addition, part of one core was sampled for beta radioactivity in order to isolate the 1965–66 horizon which resulted from atmospheric thermonuclear testing. Pits 2 and 3 were both 2 m deep and consisted of two 1 m wide perpendicular walls. Each wall was sampled for MPC, δ18O, LC and AC, and one 20 m core was drilled behind each wall. The visible stratigraphy of both walls in pit 3 was mapped. Densities were measured in each pit. The results from the pit and shallow-core analyses lead to the following conclusions.The high frequency of storm events, often associated with winds in excess of 50 knots, leads to substantial drifting. The result is that stratigraphic features in the uppermost annual layer (e.g. crusts, mass-loss layers, hard layers, etc.) exhibit substantial vertical deviations and, on occasion, are laterally discontinuous. Substantial variation may occur over very short distances: e.g. a 2 cm thick melt feature pinched out over a distance of 10 cm. Interestingly, the stratigraphy in firn older than 1 year exhibits much less lateral variation. In fact, several distinct visible stratigraphic features can be correlated among most of the shallow cores and the two deeper cores. Examples will be presented. These features provide excellent time-stratigraphic markers for cross-core comparisons. Of special interest is the presence of a major melt feature, first reported by Swiss investigators in 1983. This feature, present at four different drill sites, allows time-stratigraphic correlation and yields nearly identical 10 year accumulation rates for all four sites. Such spatial continuity means that accumulation histories reconstructed from the two deeper cores should be representative of this area.The MPC measurements (particles per ml of water) are the lowest that have been measured. No measurements of atmospheric particulate concentrations are available, but the low concentrations probably result from the high annual accumulation (∼0.55 m a−1, water equivalent), which dilutes the atmospheric input signal of particulates preserved in the firn. Swiss investigators reported similar results for the dilution of volcanic acids (e.g. SO42−), which muted the conductivity signals in a 201 m Siple Station core. The low concentrations make particles difficult to use for dating Siple cores. On the other hand, the low background levels of particulates may make Siple Station an excellent site for monitoring major atmospheric turbidity events (e.g. prolonged periods of desertification or volcanic activity).The high annual accumulation leads to the excellent preservation of the δ18O annual signal. The annual δ18O signal ranges from 15‰ at the surface to 8‰ at 122 m, so it is probable that the annual δ18O signal will be preserved over the entire 500 year record, allowing nearly absolute dating of the core.The potential of a nearly absolute time-scale, coupled with detailed particulate, liquid-conductivity and δ18O histories, should provide the first proxy climate record from this area. This 500 year history will complement similar records anticipated from the Antarctic Peninsula (∼73° S) and that recently constructed from the James Ross Island ice cap. A more global climatic picture of the last 500 years may be obtained when these high-resolution Antarctic ice-core records are integrated with non-polar records (e.g. the Quelccaya ice cap, Peru, and the Dunde ice cap, China) of similar quality.

VLF wave-injection experiments from Siple Station, Antarctica

The background of VLF wave-particle experiments from Siple Station, Antarctica, is briefly reviewed. Most of these results are interpreted in terms of Doppler shifted cyclotron interaction between energetic electrons and the VLF waves near the equatorial plane. Single frequency ducted signals that exceed a certain ‘threshold’ intensity are observed at the conjugate point (Roberval, or Lake Mistissini, Quebec) to be amplified 30–50 dB, with temporal growth rates of 30–200 dB/s. Following saturation, variable frequency emissions are triggered. With the addition of a second signal, separated in frequency from the first by Δƒ < 100 Hz, signal growth at both frequencies is reduced and sidebands are generated at frequencies separated from the carriers by integer multiples (up to seven) of Δƒ. The sidebands are attributed to short emissions triggered by the beats between the two input carriers. Midlatitude magnetospheric hiss is crudely simulated by a sequence of 10 ms pulses whose frequencies are chosen randomly within a 400 Hz band. Results show that certain combinations of 10 ms pulses link together to form chorus-like elements, suggesting a common origin for hiss and chorus. In contrast to the equatorial region phenomena are low altitude (1000–8000 km) satellite observations of narrowband Siple pulses showing increases in bandwidth of up to 1000 Hz with impulsive triggered emissions of similar bandwidth; these phenomena are interpreted in terms of a new type of quasi-electrostatic plasma instability in the subauroral region of the magnetosphere, involving 200 m wide field aligned irregularities and occurring at frequencies near the lower hybrid resonance frequency. A new crossed dipole antenna at Siple Station that generates left-hand, right-hand, or linear polarization is used to increase the effective radiated power and to perform polarization experiments.

Shallow-Core Analysis and Pit Studies at Siple Station, Antarctica: Implications for Extraction of a 500 Year Proxy Climate Record (Abstract)

Two cores, 302 and 132 m long, were drilled 1.5 km up-wind from Siple Station (75°55'S, 84°15'W) during the 1985–86 austral summer. These cores are expected to contain an annually resolvable 500 year record of atmospheric constituents. The entire length of these cores will be analyzed for insoluble particulate concentrations, liquid conductivity, and oxygen-isotope ratios. Anion concentrations (Cl−, SO4 2−) will be measured in selected sections. The extraction of a proxy climate history from these cores requires an assessment of the temporal quality and the spatial variability of the preserved record. This paper presents this assessment, which is based upon an extensive investigation of both pit and shallow-core (20 m) records. To complement the two deeper cores, eight 20 m cores were drilled and three pits were sampled. The pits were excavated up-wind of the drill site at distances between 0.5 and 1.0 km. In the central pit, 2.8 m deep, walls A and C (each 1 m wide) were positioned parallel to the prevailing wind, with wall B (4 m wide) perpendicular to the prevailing wind. The stratigraphy of wall B was mapped and continuous vertical profiles of samples were collected for microparticle concentrations (MPC), oxygen-isotopic ratios (δ18O), anion chemistry (AC) and liquid conductivity (LC). Three 20 m cores, each associated with a vertical profile of pit samples, were drilled 0.5 m behind wall B. In addition, part of one core was sampled for beta radioactivity in order to isolate the 1965–66 horizon which resulted from atmospheric thermonuclear testing. Pits 2 and 3 were both 2 m deep and consisted of two 1 m wide perpendicular walls. Each wall was sampled for MPC, δ18O, LC and AC, and one 20 m core was drilled behind each wall. The visible stratigraphy of both walls in pit 3 was mapped. Densities were measured in each pit. The results from the pit and shallow-core analyses lead to the following conclusions. The high frequency of storm events, often associated with winds in excess of 50 knots, leads to substantial drifting. The result is that stratigraphic features in the uppermost annual layer (e.g. crusts, mass-loss layers, hard layers, etc.) exhibit substantial vertical deviations and, on occasion, are laterally discontinuous. Substantial variation may occur over very short distances: e.g. a 2 cm thick melt feature pinched out over a distance of 10 cm. Interestingly, the stratigraphy in firn older than 1 year exhibits much less lateral variation. In fact, several distinct visible stratigraphic features can be correlated among most of the shallow cores and the two deeper cores. Examples will be presented. These features provide excellent time-stratigraphic markers for cross-core comparisons. Of special interest is the presence of a major melt feature, first reported by Swiss investigators in 1983. This feature, present at four different drill sites, allows time-stratigraphic correlation and yields nearly identical 10 year accumulation rates for all four sites. Such spatial continuity means that accumulation histories reconstructed from the two deeper cores should be representative of this area. The MPC measurements (particles per ml of water) are the lowest that have been measured. No measurements of atmospheric particulate concentrations are available, but the low concentrations probably result from the high annual accumulation (∼0.55 m a−1, water equivalent), which dilutes the atmospheric input signal of particulates preserved in the firn. Swiss investigators reported similar results for the dilution of volcanic acids (e.g. SO4 2−), which muted the conductivity signals in a 201 m Siple Station core. The low concentrations make particles difficult to use for dating Siple cores. On the other hand, the low background levels of particulates may make Siple Station an excellent site for monitoring major atmospheric turbidity events (e.g. prolonged periods of desertification or volcanic activity). The high annual accumulation leads to the excellent preservation of the δ18O annual signal. The annual δ18O signal ranges from 15‰ at the surface to 8‰ at 122 m, so it is probable that the annual δ18O signal will be preserved over the entire 500 year record, allowing nearly absolute dating of the core. The potential of a nearly absolute time-scale, coupled with detailed particulate, liquid-conductivity and δ18O histories, should provide the first proxy climate record from this area. This 500 year history will complement similar records anticipated from the Antarctic Peninsula (∼73° S) and that recently constructed from the James Ross Island ice cap. A more global climatic picture of the last 500 years may be obtained when these high-resolution Antarctic ice-core records are integrated with non-polar records (e.g. the Quelccaya ice cap, Peru, and the Dunde ice cap, China) of similar quality.

Air Mixing in Firn and the Age of the Air at Pore Close-Off

The air trapped in the bubbles of natural ice is not the same age as the surrounding ice. This is due to the fact that the air is enclosed in isolated bubbles only at the depth of the firn–ice transition. Within the overlying porous firn layer the air is able to mix and to exchange to a certain degree with the atmosphere. The age difference between ice and air is given by the age of the ice at pore close-off, less the mixing delay. Also, there is an age distribution due to diffusive smoothing and due to the gradual enclosure of the air at the firn–ice transition. Knowledge of this age relation is necessary for the interpretation of climatic parameters measured on ice cores. This work concentrates on the effect of diffusive mixing. We report on measurements of the diffusivity of CO2 and O2 (in N2) in firn samples from Siple Station, Antarctica. It is shown that the dominant mixing process is molecular diffusion. The diffusion coefficient depends approximately linearly on the porosity. A one-dimensional diffusion model has been used to calculate the air mixing in firn at Siple Station (Antarctica), at the South Pole, and at Station Crête (Greenland). An exchange time of between 10 and 50 years is obtained.

Air Mixing in Firn and the Age of the Air at Pore Close-Off

The air trapped in the bubbles of natural ice is not the same age as the surrounding ice. This is due to the fact that the air is enclosed in isolated bubbles only at the depth of the firn–ice transition. Within the overlying porous firn layer the air is able to mix and to exchange to a certain degree with the atmosphere. The age difference between ice and air is given by the age of the ice at pore close-off, less the mixing delay. Also, there is an age distribution due to diffusive smoothing and due to the gradual enclosure of the air at the firn–ice transition. Knowledge of this age relation is necessary for the interpretation of climatic parameters measured on ice cores. This work concentrates on the effect of diffusive mixing. We report on measurements of the diffusivity of CO2 and O2 (in N2) in firn samples from Siple Station, Antarctica. It is shown that the dominant mixing process is molecular diffusion. The diffusion coefficient depends approximately linearly on the porosity. A one-dimensional diffusion model has been used to calculate the air mixing in firn at Siple Station (Antarctica), at the South Pole, and at Station Crête (Greenland). An exchange time of between 10 and 50 years is obtained.

The Dilemma of the Rapid Variations in CO2 in Greenland Ice Cores (Abstract)

The Dilemma of the Rapid Variations in CO₂ in Greenland Ice Cores (Abstract)

Seasonal Variations in Hydrogen Peroxide in Polar Ice Cores

Hydrogen peroxide is present in polar snow and ice in remarkably high concentrations. With values up to 300 ppb, H2O2 is one of the most concentrated impurities in polar ice. We present a continuous H2O2 firn record from Siple Station (Antarctica); it covers the last 83 years with a resolution of 10–20 samples per year. A very strong seasonality is present in this record. This seasonality is also observed in a Greenland ice core from Dye 3, where we have continuously measured the top 10 m with the same resolution. The maximum concentrations correspond to summer snow layers and can exceed winter snow concentrations by a factor of 10. This property makes H2O2 a useful tracer for dating suitable cores by counting annual layers. The different steps needed to relate the atmospheric to the ice–core H2O2 concentration are discussed. As with isotopic tracers, diffusion in the firn smooths the original H2O2 concentration profile.

Seasonal Variations in Hydrogen Peroxide in Polar Ice Cores

Hydrogen peroxide is present in polar snow and ice in remarkably high concentrations. With values up to 300 ppb, H2O2 is one of the most concentrated impurities in polar ice. We present a continuous H2O2 firn record from Siple Station (Antarctica); it covers the last 83 years with a resolution of 10–20 samples per year. A very strong seasonality is present in this record. This seasonality is also observed in a Greenland ice core from Dye 3, where we have continuously measured the top 10 m with the same resolution. The maximum concentrations correspond to summer snow layers and can exceed winter snow concentrations by a factor of 10. This property makes H2O2 a useful tracer for dating suitable cores by counting annual layers. The different steps needed to relate the atmospheric to the ice–core H2O2 concentration are discussed. As with isotopic tracers, diffusion in the firn smooths the original H2O2 concentration profile.

The Dilemma of the Rapid Variations in CO2 in Greenland Ice Cores (Abstract)

The Dilemma of the Rapid Variations in CO2 in Greenland Ice Cores (Abstract)

Electron precipitation near L=4: Longitudinal variation

Observations of penetrating electron precipitation at two L ∼ 4 stations in the southern hemisphere, Kerguelen and Siple Station, Antarctica, are summarized. The nearly complete absence of precipitation at Kerguelen and its frequent occurrence at Siple Station is explained in terms of their location with respect to the center of the South Atlantic geomagnetic anomaly at L ∼ 4. Conclusions which result from this explanation are discussed along with questions which remain concerning penetrating electron precipitation.

Ice core record of the 13C/12C ratio of atmospheric CO2 in the past two centuries

The release of carbon into the atmosphere due to the activities of humans has caused an increase in concentration as well as a change in the isotopic composition of atmospheric carbon dioxide. CO2 derived from fossil fuel combustion and from biomass destruction have δ13C values of ∼−25‰ (compared to the atmospheric value of ∼−7‰) and are thus depleted in 13C. We have measured δ13C of CO2 separated from air trapped in bubbles in ice samples from an ice core taken at Siple Station in Antarctica, in which it has been possible to demonstrate the atmospheric increase of CO2 (ref. 1) and methane2 with high time resolution. The isotopic results, together with the CO2 record from the same ice core, yield information on the sources of excess carbon dioxide and provide a data base for testing the consistency of global carbon cycle models.

The local time, substorm, and seasonal dependence of electron precipitation at L ≈ 4 inferred from riometer measurements

We have examined the variations of electron precipitation at L ≈ 4 as inferred from riometer measurements of cosmic radio noise absorption made during 1975 at Siple Station and Halley Bay, Antarctica. The results are presented in the form of annual and seasonal averages of ½‐hourly values for two geomagnetic activity subsets, AE &gt; 140 nT (disturbed) and AE ≦ 140 nT (quiet). Monthly quiet day curves were used to remove the diurnal and seasonal variations in the background noise levels. Generally, the local time characteristics of the absorption were the same at both stations; the highest absorption occurred in the 0400–1600 MLT sector during disturbed conditions and in the 1200–2000 MLT sector during quiet conditions. For high AE, the highest correlation was obtained at a lag equal to the magnetic local time difference (1.5 hours) between the two stations. On the other hand, for low AE, the highest correlation occurred for a lag of 3.0 hours, nearer the local solar time difference (3.8 hours). Consistently higher absorption was measured at Halley on the average during both levels of magnetic disturbance and in all seasons. At both locations, and for both geomagnetic activity subsets, more absorption was observed in summer and equinox than in winter. This is in contrast to earlier studies for L ≧ 6, and suggests that a meridional reversal of seasonal behavior occurs between L = 4 and L = 6. In a further comparison with Siple ELF/VLF data, the 2‐ to 4‐kHz emission intensity was found to maximize in the morning hours during disturbed conditions, as was found for the absorption. For quiet conditions, the emission intensity was near minimum levels during the afternoon absorption peak. The VLF intensity is maximum in winter in contrast to the riometer result. This is consistent with increased attenuation in summer as a result of increased ionospheric opacity at that time. The local time characteristics of the absorption and wave data can be explained by a model of electron drift and precipitation under quiet and disturbed conditions that was first suggested by Thorne et al. (1974) to account for the local time variation of ELF emission intensity with substorm activity. Thus, the afternoon‐evening absorption peak during quiet conditions is attributed to cyclotron resonance interactions of drifting energetic electrons with plasmaspheric hiss, an emission largely unobservable from the ground. We associate the correlation time delay of ∼3 hours with the time difference between encounters of the Halley and Siple flux tubes with the plasmasphere bulge. The morning maximum of ionospheric absorption and 2‐ to 4‐kHz wave intensity during disturbed conditions is consistent with cyclotron resonance interactions between substorm‐injected electrons and chorus emissions occurring outside the plasmapause. For these conditions the absorption data suggest that maximum precipitation is ordered in MLT.