Learning Center
Plans & pricing Sign in
Sign Out

Slow But Not Quite Silent



                                                       15 M
                                                                                                     The fates of stars. In this Hertzsprung-Russel diagram, the main diagonal line de-
                                  10,000                                                             notes “main sequence” stars, which, like the Sun, burn hydrogen in their cores. Lines
       Luminosity (solar units)

                                                              9M                                     moving away from the main sequence are followed by stars after they have ex-
                                                                                                     hausted their hydrogen fuel supply. After a brief period of helium burning, most stars
                                    100                                                              eventually reach the white dwarf cooling track—the last part of this evolution for
                                                                                 δ Scuti
                                                          Hot                    roApp               stars of less than ~8 solar masses (M ). Asteroseismology has produced insights in-
                                                          subdwarfs                                  to the interior and evolution of a growing variety of stars (see shaded areas). The lat-
                                                                                         Sun and
                                                                                1M       solarlike   est stars to yield some of their secrets are the β-Cepheids (red area)—massive main-
                                             White dwarf                                             sequence stars that are destined to become supernovae.
                                             cooling sequence                    0.8 M
                                    0.01                         and stellar models, they infer                                           may seem like a long time, yet to the uni-
                                                                 that the interior of the star ro-                                        verse it is but the blink of an eye. Thanks to
                                                                 tates at different rates at dif-                                         the observers who collected data for future
              200     100   60     30 20      10   6     3       ferent depths, and that the tur-                                         analysis since the 1980s, the next few years
                        Surface temperature (1000 K)             bulent process of convection                                             should see a rapid growth in the seismolo-
                                                                 in its central nuclear furnace                                           gy of HD 129929 and other massive stars
   With more than 90% of all stars sharing drives mixing of material beyond the “classi-                                                  as the data continue to accumulate.
   fates similar to that of the Sun, more mas- cal” limits of the convective core. Both of
   sive stars (which will eventually explode as these results had been hinted at by earlier                                                   References and Notes
   supernovae) escaped this detailed scruti- studies of the evolution of massive stars, but                                                1. C. Aerts et al., Science 300, 1926 (2003); published
                                                                                                                                              online 29 May 2003 (10.1126/science1084993).
   ny—but not anymore.                             Aerts et al. (1) shed new light on these                                                2. D. Gough, Science 291, 2325 (2001).
       Aerts et al. monitored the β-Cephei star processes.                                                                                 3. T. Bedding et al., Astrophys. J. Lett. 549, 105 (2001).
   HD 129929 (with a mass of 9.5 solar mass-         For their work to succeed, Aerts et al.                                               4. A. Landolt, Astrophys. J. 153, 151 (1968).
   es) for more than 21 years. Earlier studies required precision photometry, which was                                                    5. R. E. Nather et al., Astrophys. J. 361, 309 (1990).
                                                                                                                                           6. S. D. Kawaler, in Variable Stars as Essential Astrophysical
   identified some oscillation modes of this made possible by advances in instrumenta-                                                        Tools, C. Ibanoglu, Ed. (NATO Science Series C, Kluwer,
   star, but gaps in data sampling precluded un- tion in the second half of the 20th century.                                                 Dordrecht, Netherlands), vol. 544, p. 511 (2000).
   ambiguous assignments. Aerts et al. now But no instrument can remove the final                                                          7. M. Breger, Baltic Astronomy 9, 149 (2000).
   show that HD 129929 is indeed a multiperi- constraint—time. Rewards for these past                                                      8. D. W. Kurtz, Astrophys. Space Sci. 284, 29 (2003).
                                                                                                                                           9. D. Kilkenny et al., Mon. Not. R. Astron. Soc. 285, 640
   odic star. They resolve six independent oscil- advances in instrumentation are now final-                                                  (1997).
   lation modes in the star. From these periods ly being realized. For us humans, 20 years                                                10. S. Charpinet et al., Astrophys. J. Lett. 471, L103 (1996).

                                                                                                                                          Quantification of the stress caused by the
                                                                                                                                          deep, slow earthquakes requires knowledge
                                           Slow But Not Quite Silent                                                                      of the precise location and amount of the
                                                                                                                                          slow slip. Herein lies a problem. Static sur-
                                                     Timothy I. Melbourne and Frank H. Webb                                               face deformation from deep faulting pro-
                                                                                                                                          vides only a blurry image of creep at depth.
          aults at subduction zones—regions                                          Cascadia subduction zone are not silent.             Moreover, the vertical deformation that is

   F      where one tectonic plate dives beneath
          another—generate the world’s largest
    earthquakes, which rapidly release strain
                                                                                     Their geodetic deformation signature corre-
                                                                                     lates with a characteristic seismic tremor
                                                                                     that bears the telltale signature of forced flu-
                                                                                                                                          most useful for locating the creep is the least
                                                                                                                                          resolvable with GPS. As a result, stress drops
                                                                                                                                          have remained largely unconstrained, and
                             over large areas of                                     id flow. This correlation opens up a more            the loading of the seismogenic zone by slow
  Enhanced online at         the plate interface.                                    facile avenue for studying slow earthquakes.         earthquakes has not been well quantified. In recent years, a                                             Isolated reports of slow earthquakes                 Such was the state of affairs until last
  content/full/300/5627/1886 much slower form of                                     have been around for decades (2). But un-            year, when Obara discovered nonvolcanic
                             strain release has                                      til a few years ago, the geophysical net-            tremor associated with subduction of the
    been detected in many subduction zones                                           works needed to resolve subtle signatures            Philippine sea plate beneath southeast Japan
    throughout the world. It involves episodes                                       of slow earthquakes did not exist. It took           (5). With the ultralow noise, bore-hole Hi-
    of fault slip that resemble conventional                                         the deployment of dense global positioning           Net array, Obara was able to detect long-pe-
    earthquakes, except that faulting occurs                                         system (GPS) arrays around the world in              riod seismic tremor at levels that on any
    slowly, often lasting weeks or months.                                           the 1990s for transient slow faulting to be          conventional network would have gone un-
        Such sluggish faulting should not by it-                                     recognized as a widespread and fundamen-             noticed or been attributed to anthropogenic
    self produce shaking at frequencies or inten-                                    tal phenomenon. Japan, with its state-of-            or other nontectonic sources. The signals
    sities that can be detected with seismome-                                       the-art arrays of seismic and geodetic in-           Obara recognized were previously only
    ters. Hence, “slow earthquakes” were held                                        strumentation, has led the way in identify-          found within active volcanoes, where they
    to be seismically quiet, or aseismic. But on                                     ing transient slow faulting events (3, 4).           are generated by flow-induced resonance in
    page 1942 of this issue, Rogers and Dragert                                          A common characteristic of slow earth-           magma-carrying conduits (6). Obara’s
    (1) show that slow earthquakes in the                                            quakes in subduction zones is that they are          tremor, however, appeared to come from
                                                                                     deep. They occur along the deeper reaches            deep regions, at depths of at least 35 km, and
                                                                                     of the plate interface, below the seismo-            well away from any known volcanic source.
       T. I. Melbourne is in the Department of Geological                            genic region that breaks every few hundred               Like their volcanic cousins, the signals
       Sciences, Central Washington University, Ellensburg,
       WA 98926, USA. E-mail: F.
                                                                                     years to produce great earthquakes. Like             described by Obara are emergent, that is, they
       H. Webb is in the Jet Propulsion Laboratory, California                       tickling the dragon’s belly, the slow fault-         mostly lack any isolated seismic P or S waves
       Institute of Technology, Pasadena, CA 91125, USA.                             ing stress loads the seismogenic regions.            that can be used to locate their origin.

1886                                                                   20 JUNE 2003             VOL 300    SCIENCE
Interseismic deformation from                                                     WILL                          ures at Earth’s surface. With nearly 2000
subduction of the Juan de Fuca                                                                                  new geophysical instruments coming on-
plate. The deformation vectors                                                                                  line with EarthScope (10), the future
reverse themselves for 2 to 6
                                                        HOLB                                                    promises better seismic locations, energy
weeks every 14.5 ± 1 months
                                                                                 WSLR                           estimates, and source mechanisms, as well
during slow earthquakes. Tremor                                             NANO
                                                                                                DRAO            as tighter constraints on along-strike prop-
correlated with the vector re-                                   UCLU
                                                                                                                agation of tremor and slip.
versals is detected to the north                                   NEAH
                                                                           ALBH                                     It may therefore be only a matter of
of the Olympic Peninsula.                              uc
                                                          a                        SEAT                         time before the initiation of regular earth-
                                                    eF n            PABH                                        quakes is itself tied definitively to fault
                                                 n d otio
                                                a m                    SATS                    LIND
Through cross-correlation of                 Ju                                                                 fluid flow, an idea that has been around for
                                           ar ica                                       JRO1
their filtered signal en-               /ye mer                         FTS1                                    years. If this idea is proven to be correct, it
                                     mm h A                                      K
velopes, however, Obara was        34 Nort                              CHZZ                                    probably applies to faults beyond those at
able to estimate that their                                                                                     subduction zones. Free-flowing brine has
hypocenters fall along the                                           NEWP        CORV
                                                                                                                been detected in faults at depths below 10
35- to 40-km depth contour                                                                                      km in the deepest boreholes on Earth (11).
within the subducting Phil-                                                       DDSN                          Like many other aspects of earthquake
                                                                           CABL                        BURN
ippine sea plate. At precise-                                                                  MDMT             physics, discoveries first made in subduc-
ly this depth, the water-re-                                                                                    tion-zone faults may prove to be applicable
                                        51 th A

                                                                                                MODB SHLD
                                          mm me

leasing dehydration from                                                  P S
                                                                          PTSG    YBHB                          to all active faults—particularly those on
                                              /ye rica

basalt to eclogite is ex-                                                   TRND                                which many of our cities are built.
                                                 ar mo

pected to occur (7). It thus
                                                    Pa tio

                                                       ci f n

seems likely that the                                                                           QUIN                References

tremor originates from           10 mm/year                                                                      1. G. Rogers, H. Dragert, Science 300, 1942 (2003); pub-
                                                                                                                    lished online 8 May 2003 (10.1126/science.
the forced flow of fluids       25 mm/year                                                                          1084783).
that are released near the                                       200 km
                                                                                                                 2. J. Beavan, E. Hauksson, S. R. McNutt, R. Bilham, K. H.
plate interface during                                                                                              Jacob, Science 222, 322 (1983).
                                                        nism on how the fluid flow enables slow                  3. I. Kawasaki et al., J. Phys. Earth 43, 105 (1995).
metamorphic dehydration. But how is the                                                                          4. H. Hirose et al., Geophys. Res. Lett. 26, 3237 (1999).
tremor related to slow earthquakes?                     slip remains unclear, but may prove as sim-              5. K. Obara, Science 296, 1679 (2002).
    Obara’s data show clearly that tremor               ple as hydraulic pressure unclamping the                 6. K. Aki, M. Fehler, S. Das, J. Volcanol. Geotherm. Res. 2,
occurs in regions of known slow earth-                  fault walls that sandwich the fluid.                        259 (1977).
                                                                                                                 7. S. M. Peacock, K. Wang, Science 286, 937 (1999).
quakes, but is absent in areas where no slow                The correspondence established by                    8. B. Julian, Science 296, 1625 (2002).
earthquakes have been detected. However,                Rogers and Dragert (1) provides an impor-                9. M. M. Miller, T. I. Melbourne, D. J. Johnson, W. Q.
he did not show that tremor and slow earth-             tant new tool with which to study the slow                  Sumner, Science 295, 2423 (2002).
quakes occur simultaneously. As Julian has              earthquake process. Tremor can potentially              10. J. H. Whitcomb, Seismol. Res. Lett. 71, 251 (2000).
                                                                                                                11. E. Huenges, W. Kessels, J. Kueck, in International Union
pointed out (8), the Cascadia subduction                be used to locate slow slip at depth more                   of Geodesy and Geophysics; XXI General Assembly
zone off the western coast of North                     precisely than can static deformation meas-                 (Univ. of Colorado, Boulder, 1995), vol. 21, p. 145.
America, with its periodic and predictable
slow earthquakes (see the figure) (9), is ide-          E C O L O G Y A N D E VO L U T I O N
al for addressing the relation between slow
earthquakes and Obara-type tremor. After
detailed analysis of 10 years of seismic
recordings from Vancouver Island, Rogers
                                                                Desperately Seeking Similarity
and Dragert now conclude not only that                                                    Janis L. Dickinson and Walter D. Koenig
slow earthquakes and tremor are highly
correlated, but that one is the hallmark of                   ver since W. D. Hamilton pointed out              cial to the evolution of helping behavior.
the other. Cascadia slow earthquakes are
not silent; rather, they are accompanied by
tremor that is notably absent when slow
                                                        E     that cooperation is facilitated by genet-
                                                              ic relatedness (1), “kin selection” has
                                                        held a central place in the study of social be-
                                                                                                                The importance of kinship relative to other
                                                                                                                more direct benefits of group living and co-
                                                                                                                operation is the subject of much debate. Two
faulting is not occurring.                              havior. Although most cooperative societies             studies on pages 1947 and 1949 of this issue
    The slow earthquakes in the Cascadia                comprise close relatives (2), there has al-             (3, 4) shed new light on this problem.
subduction zone, and by extension else-                 ways been a caveat to the logical conclusion                In support of kin selection, a preponder-
where around the world, thus seem to be                 that kin selection is the driving force in their        ance of cooperative species have groups
moderated by fluid flow in or near the plate            evolution. Consider cooperative breeders                consisting of close relatives. One particular-
interface. As in southwest Japan, the                   such as the meerkat (Suricata suricatta) that           ly notable experiment in a British bird
Cascadia tremor peaks between 1 and 5 Hz,               have “helpers” providing care to young that             species showed that returning helpers pre-
persists for days to weeks, migrates tens of            are not their own. Although helpers are usu-            ferred to assist relatives over unrelated pairs
km horizontally along the fault plane, and              ally a breeder’s offspring from prior years,            at closer nests (5). Critics argue that cooper-
appears to both trigger and be triggered by             close genetic relatedness between the giver             ation among relatives arises as a side effect
adjacent conventional earthquakes. The                  and recipient of aid is not necessarily cru-            of delayed dispersal, which causes off-
tremor is not caused by near-simultaneous                                                                       spring to remain near kin. This viewpoint is
slip of large regions, as in conventional                                                                       advocated in recent reviews highlighting
                                                         The authors are at the Hastings Reservation and
earthquakes, but probably by brine resonat-              Museum of Vertebrate Zoology, University of
                                                                                                                gains for helpers independent of aiding rel-
ing the walls of the conduits through which              California, Berkeley, CA 93924, USA. E-mail: sialia@   atives (6, 7) and severe competition that re-
it episodically bursts. The precise mecha-     ,        duces or eliminates kin-selected benefits

                                               SCIENCE         VOL 300     20 JUNE 2003                                                             1887

To top