Supplementary Material for: Upper mantle slab under Alaska: contribution to

anomalous core-phase observations on south-Sandwich to Alaska paths

Authors: Daniel A. Frost1*, Barbara Romanowicz1,2,3, Steve Roecker4

Contains 8 supplementary figures, and 3 supplementary tables.

Supplementary Figures

Supplementary Figure 1. PKPab-df and PKPbc-df travel time anomalies as a

function of ,ξ the angle of the path relative to the rotation axis, showing only data

turning in the upper 450 km of the western hemisphere of the inner core. Data from

the South Sandwich Islands to Alaska path are shown in green, while all other data is

shown in black. (a) Observed travel time anomalies showing the fits to equation S1

using all data (dark red) and only data from outside of the South Sandwich Islands

to Alaska path (red). (b) Data corrected for the red model in a.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Supplementary Figure 2: Predicted ray anomalies for (top) PKPab, (middle)

PKPbc, and (bottom) PKPdf from 3D ray-tracing through our preliminary

tomography model of Alaska, for all 6 events. (left) Travel time residuals. (centre)

Slowness residuals. (right) Back-azimuth residuals. The outline of the Alaskan slab

at 200 km depth (+0.8% dVp) from the preliminary tomography model is shown in

black.

16

17

18

19

20

21

22

23

24

Supplementary Figure 3: Left: Absolute PKPdf travel time anomalies as a function

of distance and for different sections through the slab for event 6 on 2018-12-11.

Observations are shown in blue and predictions from 3D ray-tracing through the

tomography model (as in Figure 3) are shown in red. Note that increasing distance

means moving across section from southeast to northwest. The rough location of the

slab in each cross section is marked by grey shading. Right: Map of the upper

25

26

27

28

29

30

31

mantle tomography model at 200 km depth, with stations shown as black circles.

Azimuths sections shown on the left are labelled on the right, and diamonds show

distances along the section.

32

33

34

35

36

37

Supplementary Figure 4: Comparison of absolute PKPdf ray anomalies from 3D

ray-tracing through standard and perturbed versions of our preliminary

tomography model of Alaska, for all 6 events listed in Suppl. Table 2. (Column 1)

travel time residuals, (Column 2) slowness residuals, (Column 3) back-azimuth

residuals, and (Column 4) travel time residuals as a function of distance from the

source showing observations (blue) and predictions (red). Models and fits

correspond to those in Supplementary Table 2.

38

39

40

41

42

43

44

45

46

Supplementary Figure 5: Observed (left), predicted (right) absolute PKPdf ray

anomalies from 3D ray-tracing through tomography model from Martin-Short et al.,

(2016), for all 6 events. (a, and b) Travel time residuals. (c and d) Slowness

residuals. (e and f) Back-azimuth residuals. The outline of the Alaskan slab at 200

km depth (+0.8% dVp) from the preliminary tomography model is shown in black.

47

48

49

50

51

52

The median observed absolute PKPdf travel time is subtracted from each event to

account for origin time and location errors.

Supplementary Figure 6: Left: Absolute PKPdf travel time anomalies (blue) and

predictions from different models (red) as a function of distance (i.e. moving from

southeast to northwest across the slab) for event 6 on 2018-12-11 for two azimuth

53

54

55

56

57

58

59

slices (labelled 3 and 4 as in Supplementary Figure 2). Right: Map of the upper

mantle tomography model at 200 km depth, with stations shown as black circles.

Azimuths sections shown on the left are labelled on the right, where color-scaled

diamonds show distances along the section. Grey shading indicates locations

mentioned in the main text and are shown on the model that best fits the

observations: A – distances sampling the Yakutat showing positive travel time

anomalies, best fit by model scaled by a factor of 2.5; B – sampling over the slab

showing negative travel time anomalies, best fit by standard model; C – sampling at

distances beyond the slab showing increasingly negative travel time anomalies with

distance, not well fit by any model but best fit by standard model.

60

61

62

63

64

65

66

67

68

69

70

Supplementary Figure 7: Observed (left), predicted (middle) and comparison

(right) of absolute teleseismic P wave ray anomalies from 3D ray-tracing through

our preliminary tomography model of Alaska, for P wave 3 events (see

Supplementary Table 3). (a, b and c): travel time residuals. (d, e, f): slowness

residuals; (g, h, i) back-azimuth residuals. The outline of the Alaskan slab at 200 km

depth (+0.8% dVp) from the preliminary tomography model is shown in black. The

median observed absolute P wave travel time is subtracted from each event to

account for origin time and location errors.

71

72

73

74

75

76

77

78

79

80

Supplementary Figure 8: Observed and predicted absolute PKPdf ray anomalies

for all 6 events showing the effect of the ICA correction on the observations. (a)

Predicted travel time anomalies from 3D ray-tracing through our preliminary

tomography model of Alaska. PKPdf travel time anomalies as a function of location

(b) without and (c) with the ICA correction. PKPdf travel time anomalies as a

function of distance showing predictions in red and observations in blue (d) without

the ICA correction and (e) with the ICA correction.

Supplementary Table 1: Source parameters of events used in this study for

analysis of PKPdf waves. Locations from IRIS catalogue. The median residual time

for each event likely represents errors in location or event timing since we use

absolute PKPdf times without a reference.

Number Event Date Lat Lon Depth Magnitude Median

81

82

83

84

85

86

87

88

89

90

91

92

93

(km) observed

PKPdf

residual (s)

1 2016-05-

28

-56.24 -26.94 78.0 6.0 0.0

2 2016-08-

21

-55.28 -31.75 10.0 6.4 -1.5

3 2017-05-

10

-56.43 -25.78 17.4 6.5 0.1

4 2017-09-

04

-57.79 -25.58 35.0 6.0 0.1

5 2018-08-

14

-58.11 -25.26 35.0 6.1 -2.1

6 2018-12-

11

-58.60 -26.47 163.7 7.1 3.8

Supplementary Table 2: Slopes (m) and R-squared fit (R2¿1−¿where refers to

the average of all observation) between observations and predictions for each event

for different upper mantle tomography models. t refers to travel time residuals, u to

slowness residuals and to back-azimuth residuals. SR refers to the standard model

of Roecker et al., (2018), while RMS refers to the model of Martin-Short et al., (2016)

cut at 400 km and 800 km depth, respectively. A slope of 1 indicates that the model

predictions are of the same scale as the observations. For most events, the slowness

94

95

96

97

98

99

100

101

and back-azimuth gradients reach 1 for models with lower scaling factors than

necessary for the travel time gradient to reach 1. Red coloured text indicates the

scaling factor for which the gradient (m) is closest to 1, and thus the predictions best

match the observations.

Event Date Type Model

Number SR SR sat SR scl 2 SR scl 2.5 SR scl 3

m m m m m

2016/05/28 t 0.46 0.73 0.68 0.74 0.85 0.67 1.08 0.67 1.21 0.69

1 u 0.65 0.52 0.90 0.38 0.99 0.33 0.47 0.07 0.47 0.06

θ 0.26 0.13 0.17 0.03 0.38 0.13 0.25 0.04 0.20 0.03

t 0.37 0.51 0.66 0.58 0.72 0.48 0.91 0.48 1.06 0.53

2 2016/08/21 u 0.33 0.24 0.52 0.25 0.53 0.21 0.51 0.17 0.64 0.24

θ 0.40 0.31 0.46 0.27 0.54 0.28 0.64 0.35 0.79 0.41

t 0.40 0.68 0.63 0.69 0.74 0.62 0.94 0.62 1.09 0.65

3 2017/05/10 u 0.67 0.66 0.85 0.52 1.06 0.51 1.02 0.35 1.02 0.30

θ 0.26 0.15 0.29 0.13 0.13 0.02 0.16 0.03 0.25 0.06

t 0.39 0.53 0.56 0.47 0.73 0.50 0.93 0.50 0.98 0.45

4 2017/09/04 u 0.24 0.14 0.37 0.14 0.47 0.14 0.56 0.19 0.55 0.16

θ 0.23 0.19 0.27 0.17 0.22 0.09 0.30 0.12 0.37 0.15

t 0.35 0.54 0.50 0.43 0.68 0.49 0.86 0.49 1.00 0.48

5 2018/08/14 u 0.35 0.28 0.49 0.27 0.55 0.21 0.60 0.19 0.59 0.16

θ 0.22 0.18 0.19 0.10 0.29 0.14 0.36 0.14 0.38 0.14

t 0.35 0.55 0.52 0.45 0.65 0.48 0.83 0.49 0.93 0.49

6 2018/12/11 u 0.38 0.24 0.39 0.13 0.46 0.13 0.52 0.12 0.47 0.09

θ 0.19 0.11 0.08 0.01 0.23 0.07 0.19 0.04 0.26 0.07

t 0.37 0.53 0.55 0.46 0.69 0.48 0.88 0.48 1.06 0.48

All u 0.42 0.29 0.61 0.29 0.77 0.28 0.90 0.28 0.91 0.25

θ 0.31 0.18 0.43 0.15 0.52 0.13 0.63 0.13 0.57 0.09

R2 R2 R2 R2 R2

102

103

104

105

106

107

108

109

Supplementary Table 3: Source parameters of events used in this study for

analysis of teleseismic P waves. Locations from IRIS catalogue. The median residual

time for each event likely represents errors in location or event timing since we use

absolute PKPdf times without a reference.

Number Event

Date

Lat Lon Depth

(km)

Magnitude Median observed P

wave residual (s)

1 2018-

01-10

17.47 -83.52 10 7.5 4.3

2 2018-

08-21

10.78 -62.91 146.2 7.3 2.5

3 2019-

09-24

19.08 -67.27 10 6.0 -1.4

110

111

112

113

114

115