Shallow discontinuities

Figure 6.1 shows the depths of reflection points at the depth range from 20 - 150 km.

Figure 6.1: Results for reflections from depths of $\sim$30 km to $\sim$150 km. The depths of the reflectors are displayed as columns at the location of the geometrical PP reflection point. The height of the columns indicate the depth. Additionally, the depth is shown in different colours of the columns. The Hawaii-Emperor seamount chain is displayed as solid line and the locations of the Hawaiian Islands, Kurile Islands and Aleutians are marked. The dashed line is the location of a vertical cross section displayed in Figure 6.2.

37 events have reflections from discontinuities at these depths. Some events show several reflections in this depth range. The depths are summarized in Table A.7. The depths of the reflection points are shown as column height and they are additionally colour coded. The colour scale stretches from 30 - 170 km with blue colours indicating shallower depths and red colours showing deeper depths. The columns are located at the position of the geometrical PP reflection point. The solid line marks the strike of the Hawaii-Emperor seamount chain. The location of the Hawaiian Islands, the Aleutians and the Sea of Okhotsk are marked. The dashed line in Figure 6.1 shows the location of a vertical cross section computed for this depth range. The length of this profile is much longer than the width of the corridor in which the event bounce points are located. Nevertheless, depths changes of the discontinuity perpendicular to the profile occur, resulting in a 2 dimensional structure of the reflector. These perpendicular depth changes look like a random distribution of the depths in the cross sections. The depth as a function of profile length is shown in Figure 6.2.

Figure 6.2: Vertical cross section for the depths of the shallow reflectors. The location of the profile is marked in Figure 6.1 as a dashed line. The profile stretches from the Sea of Okhotsk to the Hawaiian Islands. The depth of the reflector at the geometrical PP reflection point is projected onto the profile. The mean depths of the two reflectors identified and the errors of the mean depths are marked as solid and dashed lines, respectively.

The profile stretches from the Sea of Okhotsk to the southeast, to the tip of the Hawaii-Emperor-Seamount chain.
No continuous reflector can be found. The points seem to be totally uncorrelated and scatter through the whole depth range.
In Chapter 2.1 two distinct seismic discontinuities at these shallow depths have been explained. The Hales discontinuity at depths between 60 - 90 km due to a phase change from spinel to garnet can be identified by an impedance increase. The Gutenberg discontinuity at depths of 50 - 150 km is the upper boundary of a low velocity layer and therefore, it is connected to a negative impedance contrast. The hypothesis of two discontinuities at these shallow depths can be tested by comparing the signals originating from these boundaries with the major phase PP.
The polarity (e.g. up or down) of the wavelets can be used to study the impedance contrast at the reflecting boundaries. This is done by comparing the polarity of P$^d$P and PP. The sign of the velocity jump (fast $\rightarrow$ slow or slow $\rightarrow$ fast when looking along the PP ray path) for the PP reflection is known as a reflection at a fast $\rightarrow$ slow boundary. If the P$^d$P wavelet shows the same polarity as PP it is also reflected at a discontinuity separating fast velocities below the discontinuity from slow velocities above. A negative polarity of P$^d$P relative to PP indicates a slow $\rightarrow$ fast transition.
The polarities of P$^d$P and PP are studied by using the 4th-root vespagram, due to the invisibility of the onsets in the raw traces. Additionally, the 3-component stations of YKA are used to compute the particle motion in the P$^d$P and PP time windows.
An example of the 4th-root vespagram analysis of the polarities and the particle motion study is given in Figure 6.3 for the event of 08-feb-1990.

Figure 6.3: Examples of the wavelet polarity study for PP and P$^d$P, shown for the event of 08-feb-1990 with the PP wavelet and the P$^{153}$P wavelet.
a) Slowness traces of 4th-root-vespagram appropriate for PP and P$^{153}$P, respectively. The top trace shows P$^{153}$P with a first up movement and the lower trace PP with a first down movement indicating an opposite polarity of PP and P$^{153}$P. The traces are aligned on the PP and P$^{153}$P onset.
b) Particle motion study using a 3 component station of YKA. The particle motion in the Z-R plane is displayed. Prior to computing of the particle motion the traces were filtered using the 0.5 Hz to 1.4 Hz 4th-order band-pass filter. The circle indicates the beginning of the particle motion. The left panel shows the PP wavelet with a downward movement and the right panel the P$^{153}$P wavelet with an up movement.
Both methods come to the same result that the impedance contrasts across both boundaries show an opposite sign.

Figure 6.3 a) shows the results of the polarity study using the 4th-root-vespagram. The traces are computed for the appropriate slownesses for PP and P$^{153}$P. The two traces show only the short time windows around the PP and P$^{153}$P wavelets. The upper trace shows the P$^{153}$P wavelet. The first movement can easily be identified as an upward movement. The lower trace shows PP with a downward directed movement. Both traces show very similar waveforms although the 4th-root vespa process deforms the waveforms intensively and the waveforms should not be interpreted. Sometimes the first deflection is difficult to identify. As a control of the polarities found by the 4th-root vespagrams, the particle motions of the PP- and precursor-time-windows were studied. Figure 6.3 b) shows the particle motions of PP and P$^{153}$P for the same event discussed in Figure 6.3 a). The particle motions were computed using the Z- and R-components of the 3-component stations at YKA. The examples show the particle motion of P-waves. The time windows are the same as were used for the fk-analysis. The circle in the particle motion diagrams indicates the beginning of the motion. Corresponding to the results of the vespagram analysis, PP shows a downward and P$^{153}$P shows an upward motion indicating different impedance contrasts across the boundaries in agreement with Figure 6.3 a).
Figure 6.4 shows the results of the polarity study of all shallow P$^d$P wavelets. A plus sign indicates that the P$^d$P wavelet shows the same polarity as PP (i.e. fast $\rightarrow$ slow transition for PP) and a minus sign means inverse polarity. A question mark indicates that no definite decision on the polarity can be made.

Figure 6.4: Polarity study of shallow reflectors. A + sign indicates that the P$^d$P and PP wavelets have the same polarity and a - sign that the wavelets have an inverse polarity. The ? denotes events where no polarity can be determined. The same polarity of the wavelets indicates a reflection of P$^d$P at an impedance contrast with the same sign as for PP (fast $\rightarrow$ slow). An inverse polarity is the result of a negative reflector with a velocity change from slow $\rightarrow$ fast.

Figure 6.4 demonstrates that the reflections from shallower depths ($\le$ 80 km) mostly have the same polarity as PP, whereas deeper reflections reveal mostly a negative polarity in comparison to PP. Due to the small amplitudes of P$^d$P, some errors exist within this general trend.
The polarity study indicates two discontinuities in the depth range from 20 - 150 km. The shallow discontinuity with a mean depth of 60 km is marked by a transition from fast to slow material and might be the Hales discontinuity. At this depth, the seismic velocities change due to the phase change from spinel to garnet.
The deeper reflection points, which are characterized by a negative polarity, result most likely from the upper boundary of the low velocity layer and were described as Gutenberg discontinuity. In the region studied here, the Hales discontinuity shows a mean depth of 60.2 $\pm$ 15.8 km with the shallowest depths northwest of the bend of the Hawaii-Emperor seamount chain and a high near the Hawaiian Islands and the subduction zone. The temperature dependence of the spinel-to-garnet transition is complicated (Wood and Yuen, 1983; Jenkins and Newton, 1979). Therefore, a correlation of the depth with the mantle temperature or temperature disturbances is difficult.
The deeper reflections which show negative polarizations can be divided into two different reflectors. One is located at a mean depth of 98.6 $\pm$ 6.3 km and one at 140.0 $\pm$ 8.5 km. The deeper reflector is defined by only 5 points and is therefore uncertain. The well defined reflector with a mean depth of 98.6$\pm$6.3 km is identified as the Gutenberg discontinuity, whereas the source of the reflector near 140 km depth remains unsolved.