Jonathan M. Lees

Robust estimation of background noise and signal detection in climatic time series

We present a new technique for isolating climate signals in time series with a characteristic ‘red’ noise background which arises from temporal persistence. This background is estimated by a ‘robust’ procedure that, unlike conventional techniques, is largely unbiased by the presence of signals immersed in the noise. Making use of multiple-taper spectral analysis methods, the technique further provides for a distinction between purely harmonic (periodic) signals, and broader-band (‘quasiperiodic’) signals. The effectiveness of our signal detection procedure is demonstrated with synthetic examples that simulate a variety of possible periodic and quasiperiodic signals immersed in red noise. We apply our methodology to historical climate and paleoclimate time series examples. Analysis of a ≈ 3 million year sediment core reveals significant periodic components at known astronomical forcing periodicities and a significant quasiperiodic 100 year peak. Analysis of a roughly 1500 year tree-ring reconstruction of Scandinavian summer temperatures suggests significant quasiperiodic signals on a near-century timescale, an interdecadal 16–18 year timescale, within the interannual El Niño/Southern Oscillation (ENSO) band, and on a quasibiennial timescale. Analysis of the 144 year record of Great Salt Lake monthly volume change reveals a significant broad band of significant interdecadal variability, ENSO-timescale peaks, an annual cycle and its harmonics. Focusing in detail on the historical estimated global-average surface temperature record, we find a highly significant secular trend relative to the estimated red noise background, and weakly significant quasiperiodic signals within the ENSO band. Decadal and quasibiennial signals are marginally significant in this series.

Monitoring volcanic eruptions with a wireless sensor network

This paper describes our experiences using a wireless sensor network to monitor volcanic eruptions with low-frequency acoustic sensors. We developed a wireless sensor array and deployed it in July 2004 at Volcan Tingurahua, an active volcano in central Ecuador. The network collected infrasonic (low-frequency acoustic) signals at 102 Hz, transmitting data over a 9 km wireless link to a remote base station. During the deployment, we collected over 54 hours of continuous data which included at least 9 large explosions. Nodes were time-synchronized using a separate GPS receiver, and our data was later correlated with that acquired at a nearby wired sensor array. In addition to continuous sampling, we have developed a distributed event detector that automatically triggers data transmission when a well-correlated signal is received by multiple nodes. We evaluate this approach in terms of reduced energy and bandwidth usage, as well as accuracy of infrasonic signal detection.

Geochemical evidence for the melting of subducting oceanic lithosphere at plate edges

Most island-arc magmatism appears to result from the lowering of the melting point of peridotite within the wedge of mantle above subducting slabs owing to the introduction of fluids from the dehydration of subducting oceanic crust. Volcanic rocks interpreted to contain a component of melt (not just a fluid) from the subducting slab itself are uncommon, but possible examples have been recognized in the Aleutian islands, Baja California, Patagonia and elsewhere. The geochemically distinctive rocks from these areas, termed ‘adakites’, are often associated with subducting plates that are young and warm, and therefore thought to be more prone to melting. But the subducting lithosphere in some adakite locations (such as the Aleutian islands) appears to be too old and hence too cold to melt. This implies either that our interpretation of adakite geochemistry is incorrect, or that our understanding of the tectonic context of adakites is incomplete. Here we present geochemical data from the Kamchatka peninsula and the Aleutian islands that reaffirms the slab-melt interpretation of adakites, but in the tectonic context of the exposure to mantle flow around the edge of a torn subducting plate. We conclude that adakites are likely to form whenever the edge of a subducting plate is warmed or ablated by mantle flow. The use of adakites as tracers for such plate geometry may improve our understanding of magma genesis and thermal structure in a variety of subduction-zone environments.

Mantle flow at a slab edge: Seismic anisotropy in the Kamchatka region

The junction of the Aleutian Island and the Kamchatka peninsula defines a sharp turn in the boundary of the Pacific and North American plates, terminating the subduction zones of the northwest Pacific. The regional pattern of shear-wave birefringence near the junction indicates that trench-parallel strain follows the seismogenic Benioff zone, but rotates to trench-normal beyond the slab edge. Asthenospheric mantle is inferred to flow around and beneath the disrupted slab edge, and may influence the shallowing dip of the Benioff zone at the Aleutian junction.

Plugs and chugs—seismic and acoustic observations of degassing explosions at Karymsky, Russia and Sangay, Ecuador

Frequent degassing explosions, occurring at intervals of minutes to tens of minutes, are common at many active basaltic and andesitic volcanoes worldwide. In August 1997, April 1998, and September 1998 we recorded seismic and acoustic signals generated at two andesitic volcanoes with ‘Strombolian-type’ activity. Despite variations in explosion frequency (5‐15 h 21 at Karymsky as opposed to 1‐3 h 21 at Sangay), the signatures of the explosions are remarkably similar at these two, diverse field sites. In all explosions, gas emission begins rapidly and is correlated with an impulsive acoustic pressure pulse. Seismic waveforms are emergent and begin 1‐2 s before the explosion. We classify explosion events at the two volcanoes as either short-duration (less than 1 min) simple impulses or long-duration (up to 5 min) tremor events. Many tremor events have harmonic frequency spectra and correspond to regular 1 s acoustic pulses, often audible, that sound like chugging from a locomotive. Chugging events are intermittent, suggesting that the geometry or geochemistry of the process is variable over short time scales. We attribute the 1 Hz periodic chugs to a resonant phenomenon in the upper section of the conduit. q 2000 Elsevier Science B.V. All rights reserved.

The magma system of Mount St. Helens: non-linear high-resolution P-wave tomography

Abstract High-resolution, three-dimensional images of P-wave velocity anomalies below Mt. St. Helens, Washington, were derived using tomographic inversion. The model is a 27.5 × 21 × 20 km target volume parameterized by blocks 0.5 km per side. The area included 39 stations and 5454 local events leading to 35,475 rays used in the inversion. To diminish the effects of noisy data, the Laplacian was constrained to be small within horizontal layers, providing smoothing of the model. Non-linear effects were compensated for by iterating three-dimensional ray tracing (using pseudo-bending) between inversions and relocating earthquakes relative to the updated three-dimensional model. The structural differences between the linear and non-linear inversions appear to be insignificant, although the amplitudes of the anomalies are larger in the non-linear models. Results indicate a low-velocity anomaly (> 7%), approximately 1 km in lateral extent, from 1.5 to 3 km depths. Between 3 and 6 km depth the anomaly appears to spread out. Below 6 km depth the low-velocity feature changes to a higher-velocity perturbation with lower-velocity perturbations flanking around the perimeter of the volcano. The higher-velocity material, which correlates with the higher seismicity at that depth, is interpreted as being a plug capping the low-velocity magma chamber which begins below 9 km depth.

Crust and upper mantle of Kamchatka from teleseismic receiver functions

Teleseismic receiver functions (RFs) from a yearlong broadband seismological experiment in Kamchatka reveal regional variations in the Moho, anisotropy in the supra-slab mantle wedge, and, along the eastern coast, Ps converted phases from the steeply dipping slab. We analyze both radial- and transverse-component RFs in bin-averaged epicentral and backazimuthal sweeps, in order to detect Ps moveout and polarity variations diagnostic of interface depth, interface dip, and anisotropic fabric within the shallow mantle and crust. At some stations, the radial RF is overprinted by near-surface resonances, but anisotropic structure can be inferred from the transverse RF. Using forward modeling to match the observed RFs, we find Moho depth to range between 30 and 40 km across the peninsula, with a gradational crust–mantle transition beneath some stations along the eastern coast. Anisotropy beneath the Moho is required to fit the transverse RFs at most stations. Anisotropy in the lower crust is required at a minority of stations. Modeling the amplitude and backazimuthal variation of the Ps waveform suggests that an inclined axis of symmetry and 5–10% anisotropy are typical for the crust and the shallow mantle. The apparent symmetry axes of the anisotropic layers are typically trench-normal, but trench-parallel symmetry axes are found for stations APA and ESS, both at the fringes of the central Kamchatka depression. Transverse RFs from east-coast stations KRO, TUM, ZUP and PETare fit well by two anisotropic mantle layers with trench-normal symmetry axes and opposing tilts. Strong anisotropy in the supraslab mantle wedge suggests that the mantle ‘‘lithosphere’’ beneath the Kamchatka volcanic arc is actively deforming, strained either by wedge corner flow at depth or by trenchward suction of crust as the Pacific slab retreats. D 2002 Elsevier Science B.V. All rights reserved.

Multiple-taper spectral analysis: a stand-alone C-subroutine

Abstract A simple set of subroutines in ANSI-C are presented for multiple taper spectrum estimation. The multitaper approach provides an optimal spectrum estimate by minimizing spectral leakage while reducing the variance of the estimate by averaging orthogonal eigenspectrum estimates. The orthogonal tapers are Slepian nπ prolate functions used as tapers on the windowed time series. Because the taper functions are orthogonal, combining them to achieve an average spectrum does not introduce spurious correlations as standard smoothed single-taper estimates do. Furthermore, estimates of the degrees of freedom and F-test values at each frequency provide diagnostics for determining levels of confidence in narrow band (single frequency) periodicities. The program provided is portable and has been tested on both Unix and Macintosh systems.

Interpretation and utility of infrasonic records from erupting volcanoes

Abstract In the most basic seismo–acoustic studies at volcanoes, infrasound monitoring enables differentiation between sub-surface seismicity and the seismicity associated with gas release. Under optimal conditions, complicated degassing signals can be understood, relative explosion size can be assessed, and variable seismo–acoustic energy partitioning can be interpreted. The extent to which these points may be investigated depends upon the quality of the infrasonic records (a function of background wind noise, microphone sensitivity, and microphone array geometry) and the type of activity generated by the volcano (frequency of explosions, bandwidth of the signals, and coupling efficiency of the explosion to elastic energy). To illustrate the features, benefits, and limitations of infrasonic recordings at volcanoes, we showcase acoustic and seismic records from five volcanoes characterized by explosive degassing. These five volcanoes (Erebus in Antarctica, Karymsky in Russia, and Sangay, Tungurahua, and Pichincha in Ecuador) were the focus of seismo–acoustic experiments between 1997 and 2000. Each case study provides background information about the volcanic activity, an overview of visual observations during the period of monitoring, and examples of seismo–acoustic data. We discuss the benefits and utility of the infrasound study at each respective volcano. Finally, we compare the infrasound records and eruptive activity from these volcanoes with other volcanoes that have been the focus of previous seismo–acoustic experiments.

Three‐dimensional attenuation tomography at Loma Prieta: Inversion of t * for Q

Three-dimensional Q−1 variations in the aftershock region of Loma Prieta are derived by tomographic inversion. The data set consists of over 4000 aftershock recordings at 22 PASSCAL (Program for Array Seismic Studies of the Continental Lithosphere) stations deployed after the Loma Prieta mainshock of 1989. Estimates of attenuation are determined from nonlinear least squares best fits to the Fourier amplitude spectrum of P and S wave arrivals. The linear attenuation inversion is accomplished by using three-dimensional velocity variations derived previously in nonlinear velocity inversions. Low Q is observed near the surface and Q generally increases with depth. The southwest side of the San Andreas fault exhibits lower Q than does the northeast side and this feature apparently extends to approximately 7 km depth. The fault zone, as determined by the dipping plane of aftershock activity, is characterized by slightly higher Qp and lower Qs, compared to regions immediately adjacent to the fault. These correlate with high-velocity anomalies associated with seismicity at depth. The results are in agreement with earlier observations regarding the association of high-velocity anomalies, seismicity, and fault zone asperities.