Embed Size (px)
Citation preview
Paleomagnetism and Geochemistry of Tertiary Intrusions and Flows Associated with the Kootznahoo Formation near Kake, Southeast Alaska, and Implications for the Alexander Terrane
Jordan Epstein
Department of Geology, Carleton College, Northfield, MN USA
March 10, 2010
TABLE OF CONTENTS:
INTRODUCTION 1
GEOLOGIC SETTING 1
PETROGRAPHY 2
METHODS OF GEOCHEMICAL ANALYSIS 4
MAJOR-ELEMENT CHEMISTRY 6
TRACE-ELEMENT CHEMISTRY 11
IMPLICATIONS OF GEOCHEMISTRY FOR PALEOMAGNETISM 15
PALEOMAGNETIC METHODS AND RESULTS 15
ANALYSIS OF DIRECTIONS 16
DISCUSSION OF REMNANT MAGNETISM 18
TECTONIC IMPLICATIONS 22
KEKU ISLETS 25
HARE ISLANDS AND HOUND ISLAND OVERPRINT 26
POINT CAMDEN GABBRO AND CLARK ISLAND 27
DISCUSSION OF THE REGROUPING OF THE HAEUSLLER ET AL. (1992) DATA 28
CONCLUSION 31
ACKNOWLEDGEMENTS 31
REFERENCES 32
ABSTRACT
Major and trace element geochemistry of basaltic sills, dikes, and flows from 16
sites associated with the Paleogene-Neogene Kootznahoo Formation near Kake, Alaska
plot in either "within-plate" or "ocean-floor" fields on the Ti-Zr-Y diagram of Pearce and
Cann (1973), suggesting at least two basalts with distinct histories. The mean in-situ
characteristic remnant magnetism (ChRM) of 49 cores from 8 sites with MAD <7.0° and
α95 <10° is D =331.3° ± 7.07°, I = 66.5° ± 3.1, α95=3.1, k = 43.95. Applying a structural
correction based on the local strike and dip of the Kootznahoo Formation, the mean
corrected direction for all sites is D = 339.0° ± 10.7°, I =76° ± 2.4°, α95 = 2.4°, k = 70.75,
which suggests a moderate 20.6° ± 9.8° of vertical axis counterclockwise rotation and no
displacement when compared with a North American pole at ~23 Ma, the approximate
crystallization age of the gabbros and basalts (Haeussler et al., 1992). However, close
inspection of the paleomagnetic results from this study combined with previously
published paleomagnetic data from Haeussler et al. (1992) show systematic regional
differences in directions that suggest local tilting and vertical-axis rotations vary across
the region from no vertical axis rotation up to 46.3° ± 8.3° counterclockwise rotation at
some locations. This suggests Neogene deformation in the Alexander terrane east of the
Queen Charlotte and Clarence Strait faults could be locally significant and should be
considered when determining paleomagnetic directions for older rocks in the region.
Keywords: southeast Alaska, paleomagnetism, Alexander terrane, Tertiary tilt and
counterclockwise rotation
INTRODUCTION
Shallower than expected paleomagnetic directions from the Coast Plutonic
Complex have been interpreted to indicate large-scale northward transport of the
Wrangellia Composite Terrane of about 4000 km from its present-day location (Beck,
1976, 1981; Irving et al. 1985, 1996), and this interpretation has been coined the Baja BC
hypothesis (Irving et al., 1985, Cowan et al., 1996). However, Butler et al. (1989, 2006)
note that this challenges the conventional view of the Cretaceous North American
paleogeography, and instead suggest that these directions indicate tilting of the Coast
Plutonic Complex about 30° to the southwest. Resolution of this controversy is important
for the understanding of the tectonic history of the North American Cordillera in
particular, and on convergent boundaries in general. Whether systematic post-mid-
Cretaceous tilting and rotation of crustal blocks has occurred in the region is central to
disentangling the Baja British Columbia hypothesis (Cowan et al., 1996).
This study focuses on the geochemistry and paleomagnetism of sills, dikes, and
volcanic flows associated with the Tertiary Kootznahoo Formation near Kake, Alaska.
These observations document heterogeneous crustal rotation and tilting of the Alexander
terrane post late-Paleogene, and suggest a complex local response to Tertiary deformation
in Southeast Alaska.
GEOLOGIC SETTING
The Kootznahoo Formation is a nonmarine to marginal marine clastic unit that
crops out in Southeastern Alaska, most notably in the Zarembo-Kuiu Islands region
(Dickinson, 1979; Brew et al., 1984) and on Admiralty Island (Lathram et al, 1965). The
formation consists mostly of arkosic conglomerate, sandstone, and rare shale and is
1
intruded by gabbroic sills and dikes ranging up from one to hundreds of meters in
thickness. At Port Camden beach, the Kootznahoo Formation grades into volcaniclastic
rocks, where at least three distinct flows of three meters in thickness are recognized,
separated by hundreds of meters of the clastic Kootznahoo Formation.
Muffler (1967) suggests the sills associated with the Kootznahoo Formation have
a ‘gentle’ dip and are up to 500 meters thick, and groups the gabbro on Hamilton Island,
Kuiu Island, and Big John Bay together, with a suggested age of Miocene-Oligocene
(Gehrels and Berg, 1992). Muffler (1967) also notes that the large gabbroic sills crop out
roughly parallel with the Kootznahoo Formation, and that some samples have a magmatic
foliation that is roughly parallel with the trend of the dip. On the basis of its inferred age
and geographic location, Brew and Morrel (1983) consider these large sills to be the
northernmost intrusions of the Kuiu-Etolin volcanic-plutonic belt. Heaussler et al. (1992)
dated the intrusion at Point Hamilton using Ar40/Ar39, and obtained a 23 Ma age.
PETROGRAPHY
Thin-section analysis of 20 sites (Fig. 1) indicates two types of intrusions and
volcanism: (1) a common basaltic type, which is quartz poor, and contains
orthopyroxene, clinopyroxene, and plagioclase with rare olivine and biotite, and (2) a
relatively more rare rhyolitic type, which contains significant quartz, plagioclase and
horneblende, which is consistent with descriptions by Gehrels and Berg (1992). Evidence
of alteration is present at almost all sites, with the presence of serpentine, chlorite,
sericite, calcite and quartz veins, which Brew et al. (1984) attribute to deuteric altering.
Conspicuously, at site 09JDE10, virtually no evidence of alteration is present, suggesting
a lack of fluids during cooling. Because of the large variation in grain size and alteration
2
Admiralty Island Dike
PortageBay
Hamil tonBay
Big JohnBay
DavidsonBay
KadakeBay
KuiuIsland
KuiuIsland
KupreanofIsland
Kake Airport
Dak
anee
k Bay
Ro
ck
y P
a ss
Port Cam
de
n
Keku Strai t
Kupreanof Island
HoundIsland
HareIs lands
ClarkIs land
KekuIs lets
PointHamil ton
Hamil tonIs land
PointCamden
133°40’
56°55’
56°50’
56°45’
133°45’133°50’133°55’134°00’
0 3 km
Tmgb
Tmgb
Tmgb
Tmgb
Tmgb
Tmgb
Tmgb
Tmgb
Tmgb
Tmgb
Tmgb
MDc
MDc
MDc
Pp
Pp
Ksm
Tk
Tk
Ksm
Ksm
Pp
Pp
Qs
Tk
Tk
Tk
TkQTf
QTf
QTf
QTf
QTf
QTf
QTb
QTb
QTb
QTb
QTb
QTb
QTc
QTc
QTc
QTc
vRT
cRT
hRT
hRT
kRT
vRT
vRT
vRT
vRT
vRT
QTc
Qs
QTf
Geologic Map of the Keku Straight Area, Southeast Alaska.Modified from Muffler (1967) and Brew et al. (1984)
Surficial deposits (Quaternary): Alluvium, glaciofluvial deposits, and tidal mud flats.
Felsic volcanics (Quaternary and Tertiary): Grey to buff altered volcanic flows, tuff, or shallow intrusives with myrolitic cavities.
Basalt (Quaternary and Tertiary): Dark grey to black aphanitic basalt.
Hound Island volcanics (Triassic): Dark green to black basaltic pillow breccia and pillow lava with some interbedded limestone.
Hamilton limestone (Triassic): Dark grey thinly bedded limestone.
Qs
QTf
QTb
QTc
Tmgb
Tk
Ksm
Pp
MDc
vRT
ALASKA
MAP LOCATION
hRT
cRT
kRTVolcaniclastic deposits (Quaternary and Tertiary): Grey to buff conglomerate and lithic sandstone. Gradational contact with Tk.
Gabbro (Tertiary): Grey to black phaneritic gabbro with olivine and clinopyroxene.
Kootznahoo Formation (Tertiary): Grey to buff arkosic sandstone, conglomerate, and black shale. Coal, fossil leaves and wood present.
Turbidites (Early? Cretaceous): Rhythmically layered grey to black sandstone, siltstone, and mudstone with carbonate concretions.
Cornwallis limestone (Triassic): Grey medium to thick bedded oolitic limestone.
Keku volcanics (Triassic): Altered felsic flows and breccia, basalt, volcanoclastics and limestone.
Pybus Formation (Permian): White to light grey limestone, dolostone and chert.
Cannery Formation (Mississipian and Devonian): Dark grey to bluish green thinly bedded volcanic argillite and greywacke with chert.
Sample locationStrike and dip of bedding
09JDE01A&B
09JDE01
15
10
15
135
12
16
10
10
10
5
25
15
25 15
15
1510
20
1515
09JDE04
09JDE10A
09JDE11
09JDE07
09JDE13
09JDE12
14
12
20
20
15
15
09JDE02
09JDE03
09JDE05
09JDE09
09JDE14
09JDE15&1609JDE08
09JDE06
09JDE10BHE69
HE60
HE25&26
HE74HE75
HE76
HE77
HE73
HE71&72HE50&51
HE52HE53
HE21&22
HE24HE23
HE70HE68
HE67
HE58&59
HE20
HE55-57
HE17
HE19HE18 HE14HE13HE15
HE61HE63HE65
HE34&35HE32&33
HE31HE30
HE36HE37
HE39-41
HE42-45
HE46HE47
HE48HE49
HE85
HE78
HE83&84HE79-81
HE82
HE66
Hamilton Bay Dike
Admiralty Island Dike
Pt. & Port Camden Dikes
Pt. Camden SillPt. Camden Flow
Pt. Hamilton Sills + Dikes
Dellenite
Geochemistry Sites
Paleomagnetism Sites
QTb
Fig. 1
HorseshoeIs land
Recent dikesPt. Hamilton gabbroPort Camden flowPt. Camden gabbroLower Kuiu Is. overprintUpper Kuiu Is. overprintKeku Islets gabbroKeku Islets overprintHound Island overprintHare Islands gabbroHamilton Is. overprintUnknown component
3
among sites, it is not possible to discern systematic differences in mineralogy between
sites based on petrographic observations alone.
Of the three rhyolitic samples, sites 09JDE05 and 09JDE09 contain large quartz
grains that appear sub-equant and hornblende, suggesting a magmatic origin, consistent
with Muffler’s (1967) field description of dellenite. Sample 09JDE09 contained
miarolitic cavities, with quartz grains clearly growing into the empty space, which
constrains emplacement to a shallow depth. Conversely, while 09JDE11 has what
appears to be some characteristics of a magmatic texture, including extremely large,
euhedral laths of plagioclase, the laths are surrounded by very small grains of quartz,
suggesting a mobilization and subsequent relatively rapid cooling of quartz. Sample
09JDE11 is sampled from the margin of the Point Camden gabbro, a sill hundreds of
meters thick, and such an interpretation is consistent with localized melting of quartz-
rich sandstone at the margin of a large sill. This high quartz-content margin is 3.5 meters
thick, which suggests either prolonged heating or extremely high temperatures at time of
intrusion.
METHODS OF GEOCHEMICAL ANALYSIS
Representative samples were collected from sixteen sites associated with the
Kootznahoo Formation: six mafic sills (Point Hamilton, Big John Bay, Point Camden,
Port Camden), six mafic dikes (Point Hamilton, Hamilton Bay, Point Camden, Port
Camden, Admiralty Island), two mafic flows (Port Camden), and two felsic flows or
intrusions (Davidson Point, Horseshoe Island) (Fig 1).
4
Samples were prepared and analyzed at Macalaster College using the procedure
of Vervoort et al. (2007). Their method, repeated verbatim, is as follows. “Samples were
crushed in a Bico-Braun jaw crusher, separated into two aliquots using a sample splitter,
and ground to fine powder (<75 μm) using a shatterbox with steel and tungsten carbide
(WC) bowls. The aliquot ground in WC was used for preparation of samples for major
element analyses, and the powders ground in the steel mill were used for trace element
and isotope analyses. The concentrations of major elements (SiO2, TiO2, Al2O3, Fe2O3,
MnO, MgO, CaO, Na2O, K2O, and P2O5) were determined by XRF from fused lithium
borate glass beads. Loss on Ignition (LOI) values were calculated from the percent
weight loss after ignition (1000° C for 1 hour) relative to the dried (105° C for 2 hours)
sample powders.
One gram of ignited rock powder was mixed with five grams of dried lithium
tetraborate/metaborate (12:22) flux, and 0.01 gram ammonium nitrate (NH3NO4). The
sample mixture was then fused in platinum alloy (95% Pt—5% Au) crucibles with two
drops 50% hydrobromic acid solution (HBr) using a Claisse® Fluxy. Fused samples were
cast in 32mm diameter platinum alloy molds which produced glass discs approximately
5mm thick.
Trace-element concentrations (Co, Ni, Zn, Ga, Ba, Rb, Sr, Y, Zr, Nb, Ce, Pb, Th,
and U) were determined from pressed powder pellets. Ten grams of rock powder were
combined with 15–20 drops of 2% polyvinyl alcohol (PVA). The mixture was then
placed in a 40mm diameter stainless steel die and formed into a 5mm thick pellet by
applying six tons pressure for 60 seconds in a manual press.
5
Major and trace-element concentrations were determined using a Philips PW-
2400 X-ray Fluorescence Spectrometer with a Rh-anode, end window X-ray tube, and
Philips Super-Q analytical software at Macalester College. Elemental concentrations
were determined by comparing X-ray intensities for each element in a sample unknown
with those from reference samples from the US Geological Survey” (Vervoort et al.,
2007).
MAJOR-ELEMENT CHEMISTRY
SiO2 content of samples ranges widely, from about 35 -75 weight percent (Table
1); however, analyses with weight percent lower than 45 should be treated with suspect,
as these samples tended to have extremely high loss on ignition (LOI), with some as high
as 14 weight percent. Comparison with geochemical analysis by Ford et al. (1996) of the
Admiralty Island volcanics had similarly high LOI but no quartz deficient samples, which
suggests a processing error on my part. Discounting these data, these rocks plot as either
basalt or rhyolite, as shown on the total alkali and silica (TAS) diagram (Fig. 2) (Lebas et
al., 1986). Discounting the sample from Admiralty Island and the chill margin at site
09JDE11, there is a SiO2 gap from about 48-75 weight percent. The lack of rocks of
intermediate composition makes determination of source material, fractional
crystallization, or crustal assimilation difficult; however, the Point Hamilton gabbro has
higher MgO and lower Na2O content than all other sites (Fig. 3); however, Na2O is highly
mobile so should be interpreted with care. Aside from this, the major-element chemistry
of basalts is largely homogenous across sites and samples, with the only exception that
dikes (regardless of location) having slightly higher TiO and lower Al2O3 content than
other sites (Fig. 3).
6
Major-element contents of the rhyolite samples 09JDE05 and 09JDE09 are nearly
identical, which were mapped by Brew et al. (1984) as dellenite and granite, respectively.
This suggests that the two are genetically related, and that the dellenite is simply the
extrusive equivalent of the granite. Comparison with the other “rhyolite” sample
09JDE11, the margin of the Point Camden sill, shows significantly lower SiO2 and a
distinct major element composition. This is consistent with the interpretation of localized
melting of a quartz-rich arkosic sandstone.
7
Tabl
e 1.
Maj
or-e
lemen
t che
mica
l com
posit
ion
(weig
ht p
erce
nt) o
f sam
ples
reno
rmal
ized
by
the
volat
ile fr
ee a
mou
ntEl
emen
t09
JDE0
1A09
JDE0
1B09
JDE0
209
JDE0
309
JDE0
409
JDE0
509
JDE0
709
JDE0
809
JDE0
909
JDE1
009
JDE1
0G09
JDE1
109
JDE1
409
JDE1
509
JDE1
609
JDE1
8Si
O2
46.4
946
.10
34.9
846
.43
46.5
975
.00
46.8
543
.81
73.8
648
.06
47.3
069
.17
44.0
541
.95
38.3
650
.22
TiO
21.
151.
053.
761.
321.
300.
111.
621.
520.
252.
151.
380.
432.
212.
241.
661.
89A
l 2O3
16.6
016
.83
15.1
516
.30
16.1
412
.61
16.1
618
.71
9.80
15.3
218
.09
14.0
314
.53
15.1
913
.98
14.9
1Fe
2O3
11.0
410
.98
16.4
811
.43
11.7
11.
568.
958.
555.
8413
.41
9.61
4.49
12.4
212
.32
10.5
99.
93M
nO0.
180.
220.
190.
180.
190.
010.
160.
180.
060.
220.
160.
080.
190.
200.
210.
18M
gO8.
619.
203.
298.
077.
950.
102.
365.
490.
015.
196.
400.
393.
613.
795.
083.
40C
aO11
.52
11.0
110
.36
10.7
811
.32
0.64
10.8
511
.20
0.11
9.39
10.1
51.
9812
.35
9.28
12.0
36.
44N
a 2O
2.50
2.26
2.95
2.57
2.79
4.34
3.61
3.90
5.48
3.41
3.96
4.62
3.09
3.43
3.09
3.59
K2O
0.21
0.37
0.86
0.69
0.21
5.32
0.59
0.32
4.53
0.71
0.45
4.07
0.62
0.32
0.73
0.98
P 2O
50.
100.
100.
600.
120.
110.
020.
220.
190.
000.
350.
160.
070.
360.
250.
370.
46LO
I1.
581.
8911
.38
2.11
1.68
0.29
8.63
6.13
0.07
1.79
2.34
0.68
6.56
11.0
213
.91
8.00
Sum
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
100.
0010
0.00
8
Picro-basalt
Basalt
Basalticandesite
AndesiteDacite
Rhyolite
Trachyte
TrachydaciteTrachy-andesite
Basaltictrachy-andesiteTrachy-
basalt
TephriteBasanite
Phono-Tephrite
Tephri-phonolite
Phonolite
Foidite
35 40 45 50 55 60 65 70 750
2
4
6
8
10
12
14
16
Na2O+K2O
SiO2
Hamilton Bay Dike
Admiralty Island Dike
Pt. & Port Camden Dikes
Pt. Camden SillPt. Camden Flow
Pt. Hamilton Sills + Dikes
Dellenite
Figure 2. Total alkali versus silica diagram, after Lebas et al. (1986). The Point Hamilton sills and dikes have significantly lower total alkali content than all other samples. There is a silica gapfrom 48 to 70 wt. percent, exlcuding the Admiralty Island dike.
9
0
5
10
MgO
12
17
22
Al2O3
0.0
0.5
1
1.5
2.0TiO2
CaO
K O2 Na O2
6
4
2
30 75 30 75SiO2
15
0
Figure 3. Harker Diagrams for Majors. Point Hamilton sills and dikes have higher MgO and lower Na Othan all other samples. Dikes appear to have higher TiO and lower Al O . Sample 09JDE11 has lower silica and distinct major elment composition from the other rhyolites.
Sample 9 has 9 wt. % Al2O3
10
5
2
3
4
0
SiO2
6
4
2
0
2
2 3
10
TRACE-ELEMENT CHEMISTRY
Trace-element contents of basaltic samples (Table 2) show variation that appears
independent of SiO2 (Fig. 4). This suggests that the basalts are not related by fractional
crystallization, as it would be expected that incompatible elements be concentrated in
basalts of higher SiO2 content, and compatible elements concentration likewise be
reduced. However, because SiO2 content of reasonable analyses spans only 45-48 weight
percent, it is unlikely that any sign of fractionalization would be apparent. Trace-element
content of the rhyolitic samples (excluding 09JDE11) have extremely low concentrations
of the compatible elements Ni, Cr, and Sr, and high concentrations of the incompatible
elements Rb and Zr, as expected, although Ba concentration is more variable than
anticipated. The similarity between samples 09JDE05 and 09JDE09 supports the
conclusion that they are genetically related, and sample 09JDE09 has significantly higher
elevated levels of Zr, as expected of a plutonic rock.
It is not possible to distinguish between bimodal volcanism, fractionalization, or
assimilation from this data alone—however, given the paucity of rocks of intermediate
composition, I would favor bimodal volcanism as found throughout Southeast Alaska and
on the nearby Zarembo and Admiralty Islands (Lindline et al., 2004, Ford et al., 1996).
Finally, elevated trace element concentration in of La, Nb, Rb, and Zr in sample 09JDE11
11
Tabl
e 2.
Trac
e ele
men
t con
tent
(in
parts
per
milli
on)
Elem
ent
09JD
E01A
09JD
E01B
09JD
E02
09JD
E03
09JD
E04
09JD
E05
09JD
E07
09JD
E08
09JD
E09
09JD
E10
09JD
E10G
09JD
E11
09JD
E14
09JD
E15
09JD
E16
09JD
E18
Ba10
726
267
761
510
570
544
722
137
719
201
990
365
328
456
4187
Ce11
1044
108
7129
2215
434
1864
4722
4440
Co53
5216
350
502
3348
136
324
4348
4728
Cr27
027
853
260
242
412
224
55
8019
83
105
6821
213
Cu42
5845
422
523
1119
619
206
2412
880
41Ga
1717
3119
1718
2119
3420
1922
2222
1918
La3
417
66
4015
1291
198
3720
1220
16N
b4
521
45
1910
997
169
2115
1214
16N
i11
111
959
8394
1231
135
1437
4015
3782
141
25Pb
bd7
bdbd
bd17
2bd
141
bd13
0bd
bdbd
Rb8
814
215
152
52
137
85
114
103
1219
Sc34
3341
3938
332
370
2926
638
2127
24Sr
234
255
525
262
224
6931
529
13
396
559
144
450
348
529
863
Th3
3bd
23
163
219
20
112
40
bdU
10
6bd
bd9
21
10bd
bd7
13
bdbd
V21
520
441
625
623
72
257
253
224
219
95
298
276
286
280
Y26
2349
3029
4632
2711
140
2054
4026
2331
Zn69
145
449
5281
3111
079
213
121
7286
132
107
8811
2Zr
8077
294
9990
158
171
125
943
206
113
355
219
151
164
238
12
NiO
200
0 0
500
1000
Ba
Note Sample 18 has a Ba of4402 ppm
Cr
0
50
100Rb
0
100
200
300Zr
Note Sample 9 has a Zr of1006 ppm
0
500
1000
Sr
30 7530 75
400
0
300
200
100
300
100
150
200
400
500
SiO2 SiO2
Figure 4. Harker Diagrams for trace elements. Selected trace elements (in ppm) against silica (wt. percent.). Trace element concentration of basalts appears independent of silica. Rhyolites show high concentrationof Zr, Rb, and low concentrations of Ni, Cr, Sr.
13
Zr Y*3
Ti/100
C
DA
B
Island-arc A B
Ocean-!oor B
Calc-alkali B C
Within-plate D
Alkali basalt
Tholeiitic basalt
0.00 0.05 0.10 0.150
1
2
3
4
TiO 2
Zr/(P 2O 5*104)
200 400
Figure 6. Tectonic discrimination plotsfor basalts with 1220 wt. percent CaO + MgO, after Pearce and Cann (1973). Note that the Point Hamilton sills and dikes plot exclusivley in theoceanfloor basalt, while all other samples plot as withinplate or calcalkaline basalts.
Figure 5. Alkali vs.Tholetic discrimination diagram for basalts, after Floyd and Winchester (1976).All basalts from the Keku Strait appear tholeitic.
14
is consistent with the hypothesis of localized melting and incorporation of the nearby
poorly developed arkosic sandstone formation.
IMPLICATIONS OF GEOCHEMISTRY FOR PALEOMAGNETISM
The geochemistry of these samples suggests that there are at least two genetically
distinct basalts: the Point Hamilton gabbroic sill, and all other sites. While all basalts are
tholeitic (Fig. 5), the Point Hamilton sill has a geochemical signature consistent with
ocean floor basalts, while all other samples plot in the within-plate or calc-alkaline fields
(Fig. 6). In thin section, it is difficult to distinguish between basalts—thus, geochemically
and petrologically distinct basalts might easily be inferred to be identical based on
mineralogy alone. Reconnaissance of this area of Southeast Alaska has previously
grouped all sills and dikes together under the label “Tertiary Gabbros” (Muffler, 1967;
Brew et al., 1984; Gehrels and Berg, 1992). Geochemical evidence in this paper suggests
that at least some of these sills and dikes are distinct and that the assumption of a
homogenous “Tertiary Gabbro” be revisited. This further suggests that paleomagnetic
data from one location of intrusions cannot necessarily be applied to or averaged with
another without further geochemical or geochronological work demonstrating they have
related compositions or crystallized at the same time.
PALEOMAGNETIC METHODS AND RESULTS
Seventy-two paleomagnetic samples were collected from 12 sites of basaltic
dikes, sills, and volcanic flows exposed on shorelines of islands west of Kake in southeast
Alaska (Fig. 1). Six or more oriented samples were collected from each site. A portable
coring device was used to collect samples in the field, which were oriented using an
inclinometer and magnetic compass. At the Institute of Rock Magnetism at the University
15
of Minnesota, a ten-step alternating field demagnetization was performed on each
sample. All measurements were made using a 2G three axis cryogenic magnetometer in a
shielded room. For 8 sites (49 cores), characteristic remnant magnetism (ChRM) was
easily isolated, and little evidence of thermal or chemical overprinting was present. The
magnetic mineralogy appears to be magnetite or titanomagnetite, based on the
coercivities of less than 100 mT for most samples, and in thin section visible equant and
euhedral to subeuhedral opaque grains are ubiquitous.
ANALYSIS OF DIRECTIONS
The demagnetization behavior for most of the samples in this study suggests one
stable component of magnetization for which a ChRM was isolated. The high coercivity
and stable component is typically well defined, and yields four primary directions of
magnetism. The first group of magnetizations is steeply dipping down (I = 67°) and
oriented north-northwest (D = 335°); the second, dipping moderately down (I = 55°) and
oriented northwest (D = 320°); the third dipping moderately down (I = 59°) and oriented
north (D=2°, and the fourth, dipping steeply down (I = 76°), and oriented north (D =7°)
(Fig. 7A-D). A fifth unstable magnetization was found at four sites for which no ChRM
could be confidently established (Fig. 7 E, F). All sites within this fifth group of
characteristic magnetizations displayed extremely low inclinations either up or down (I ≈
± 10°), and oriented either south or north (D ≈ 0° or 180°). These samples also had
extremely low coercivities, suggesting variations in magnetic mineralogy, range of
magnetic grain sizes, or lack of a strong magnetic field at time of cooling through the
blocking temperature (Haeussler et al., 1992). Regardless, these directions are not further
treated in this paper.
16
NRM
5 101520
30
40
50
75100
N, Up
W
NRM5N, Up
E
1015
2030
40
50
NRM
510
1520
30 405075
N, Up
W
NRM
5
101520 30
40
N, Up
W
NRM
5
N, Up
E
1015
2030
A. B. C.
D. E. F.
Fig. 7. A-E. Orthogonal vector plots of alternating field demagnetization results. Filled symbols correspond to the horizontalvector component (declination), and open symbols correspond to the vertical component projected onto the plane of thefigure (inclination). Alternating field steps in mT. Plots show examples of A.) normal polarity Point Hamilton sill,B.) reverse polarity Port Camden flow, C.) Point Camden sill, D.) recent dike at Port Camden, E.) a lack of stablemagnetization. Note the extremely low coercivity and the loss of nearly all remnant magnetism in the smallstep from 15 to 20 mT. F.) is an equal area plot of the directions of 09JDE14, showing extremely shallow dips and both normal and reversed polarities within the same core. Such data is suggestive of a lack of a strong magnetic fieldat the time of cooling through the blocking temperature, possibly due to a transitioning pole. Such directions are not treatedfurther in this paper.
09JDE015E
09JDE014F
09JDE08D
09JDE11B09JDE4C
09JDE014F
17
For individual samples, I calculated the orientation of the ChRM by a least
squares regression of the portion of the vector diagram that decays univectorally toward
the origin of the plot, using a weighted analysis and forcing through the origin. Between
3 and 10 points, with an average of 5.0 points, were used to define this straight-line
segment. No least squares regression reported in this paper had a maximum angular
deviation (MAD) of greater than 7.0°, and the average MAD was 1.81°. Site means were
calculated using Fisher statistics, and all sites had an α95 confidence interval of <10°, with
an average of 5.27°. The Fisher precision parameter k for all sites is greater than 50, with
an average of 218.6 and a maximum as high as 645. The coercivities of sites is quite
variable, with some samples losing >90% of their initial magnetization when exposed to
fields of 20 μT, while other sites displayed extremely high coercivities and showed
considerable remnant magnetization even after exposure to 100 μT.
DISCUSSION OF REMNANT MAGNETISM
Paleomagnetic data from the Keku Strait is organized by locality and presented in
Table 3 and Fig. 8. Magnetic directions are compared to the 20 Ma reference pole of
Hagstrum et al. (1987) at 87.4° N, 129.7° E, A95 = 3.0°, with expected direction of I =
73.5° ± 1.5°, D = 359.6° ± 5.9°. Using the Oligocene reference pole at 84.0° N, 168.0°,
A95 = 4.0° (Diehl, 1988), the expected Oligocene direction differs by only I = +1.5° and
D = +7°, so the choice of reference pole has largely negligible effect for paleomagnetic
comparison. Magnetic directions for individual sites are compared to expected values,
and values for inclination flattening (F ± ΔF) and rotation of declination (R ± ΔR) are
calculated using the methods of Beck (1980) and Demarest (1983). Site 09JDE15 is the
only site to record a magnetization direction within error of the expected direction; all
18
Abbreviations as follows: M — inferred magnetization group, with: (red) Point Hamilton gabbro group, (orange) Point Camden gabbro, (yellow) Port Camden flows, and (green) recent dike; N — indicates the number of cores used out of the total collected; P — polarity, whether normal (n), reversed (r) or no good(ng); R — the vector length of the mean direction; α95 — indicates the 95% confidence interval; k — the Fisher precision parameter; I — Inclination; D — Declination; R — Rotation (positive for clockwise) as compared to the expected declination based on the 20 Ma North American pole of Hagstrum et al.(1987); F — Flattening, after Beck et al. (1981) and Demarest (1983); Bedding S & D— best estimate for strike and dip of the Kootznahoo Formation near site.
Group In-Situ Directions BeddingM Group N P r α95 k I D R ΔR F ΔF Variable
All Data 49/72 n/r 48.9 3.1 43.95 66.5 331.3 -28.3 7.8 7.0 2.9Port Camden Flows 19/19 n/r 18.8 3.9 75 62.7 313.3 -46.3 8.3 10.8 3.5Point Hamilton and Big John Bay 21/23 n 20.7 3.6 79 67.6 339.0 -20.6 8.9 5.9 3.2
Group Tilt-Corrected Directions BeddingM Group N P r α95 k I D R ΔR F ΔF Variable
All Data 49/72 n/r 48.9 2.4 75 76.9 339.0 -20.6 9.7 -3.4 2.4Port Camden Flow 19/19 n/r 18.8 3.6 90 74.3 314.9 -44.7 11.5 -0.8 3.2Point Hamilton and Big John Bay 21/23 n 20.7 2.9 118 78.5 353.6 -6.0 12.8 -5.0 2.8
Tilt-Corrected Point Camden sill Compared to Late Cretaceous Pole (Mcelhinney, 1979) BeddingM Site Latitude(°)Longitude(°)N P r α95 k I D R ΔR F ΔF S D
09JDE11 56.79 133.88 5/6 n 5 7.7 81 70.5 9.6 -24.6 19.8 4.3 6.4 80 12
Table 3Epstein (2009) data.Site In-Situ Directions BeddingM Site Latitude(°)Longitude(°)N P r α95 k I D R ΔR F ΔF S D
09JDE3+4 56.86 133.87 9/10 n 8.9 5.3 86 70.5 335.8 -23.8 13.6 3.1 4.5 47 1009JDE06 56.75 133.87 7/7 n 7 2.2 646 62.7 298.3 -61.3 6.1 10.8 2.3 49 1009JDE07 56.76 133.87 6/6 n 6 5.8 111 68.5 312 -47.6 13.7 5.0 4.9 49 1009JDE08 56.77 133.87 6/6 r 6 3.6 290 -55.4 148 -31.6 6.9 18.1 3.2 49 1509JDE10 56.80 133.86 0/6 ng09JDE11 56.79 133.88 5/6 n 5 7.7 81 59.0 2.5 2.9 12.9 14.5 6.3 80 1209JDE12 56.82 133.74 6/6 n 6 3.8 254 69.1 324.3 -35.3 9.9 4.4 3.4 75 1509JDE13 56.80 133.68 6/6 n 5.9 8.4 54 61.4 352.5 -7.1 14.9 12.1 6.9 47 1209JDE14 56.78 133.86 0/7 ng09JDE15 56.78 133.87 4/6 n/r 4 6.3 160 76 7.9 8.3 22.1 -2.5 5.3 49 1609JDE16 56.78 133.87 0/6 ng
Site Tilt-Corrected Directions BeddingM Site Latitude(°)Longitude(°)N P r α95 k I D R ΔR F ΔF S D
09JDE3+4 56.86 133.87 9/10 n 8.9 4.5 119 78.6 356.5 -3.1 19.1 -5.0 3.9 47 1009JDE06 56.75 133.87 7/7 n 7 2.2 669 72.3 291.7 -67.9 7.4 1.2 2.3 49 1009JDE07 56.76 133.87 6/6 n 6 5.8 112 78.4 313.1 -46.5 24.6 -4.9 4.9 49 1009JDE08 56.77 133.87 6/6 r 6 3.4 327 -69.2 159.2 -20.4 9.0 4.3 3.1 49 1509JDE11 56.79 133.88 5/6 n 5 7.7 81 70.5 9.6 9.7 19.4 3.0 6.3 80 1209JDE12 56.82 133.74 6/6 n 6 3.8 258 80.3 339.1 -20.5 19.2 -6.8 3.4 75 1509JDE13 56.80 133.68 6/6 n 5.9 8.4 54 76.1 0.1 0.5 30.2 -2.6 6.9 47 12
19
In-Situ Directions Tilt-Corrected Directions A. B.
C. D.
E. F.
G. H.
Port Camden Flows
Pt. Hamilton
Pt. Camden Sill
N N
N N
N N
N N
Fig 8.
A. In-situ directions for all cores. Black circles are normal and white reversved polarities. The red triangle is the calculated mean pole. The green square is the expected direction as compared to a Miocene pole. Comparison with the expected direction shows directions are 28.3° ± 7.8° counterclockwise and 7.0° ± 2.9° more shallow than expected.B. Tilt-corrected directions for all data. Comparisonwith the expected direction shows declinations are 20.6° ± 9.7°counterclockwise and inclination within error of expected.
All data
Sills & Dikes
Big John Bay
C. In-situ directions for cores converted to lower hemisphere. Yellow circles are individual cores, yellowtriangle is the mean, black circle is the α95 confidence interval.The green square is the expected direction as compared to a Miocene pole. Comparison with the expected direction shows mean is 46.3° ± 8.3° counterclockwise and 10.8° ± 3.5° more shallow than expected.D. Tilt-corrected directions. Comparison with the expected direction shows declinations are still44.7° ± 11.5°counterclockwise and inclination within error of expected.
E. In-situ directions for cores. Red circles are individualcores, red triangle is the mean, black circle is the α95confidence interval.The green square is the expected direction as compared to a Miocene pole. Comparison with the expected direction shows mean is 20.6° ± 8.9° counterclockwise and 5.9° ± 3.2° more shallow than expected.F. Tilt-corrected directions. Comparison with the expected direction shows directions are within error ofexpected.
G. In-situ directions for cores. Orange circles are individual cores, orange triangle is the mean, black circleis the α95 confidence interval.The green square is theexpected direction as compared to a Miocene pole, the blue as compared to a Cretceous pole. Comparison with the expected Miocene direction shows no declination discordance but inclination 14.5° ± 6.3° more shallow than expected.
H. Tilt-corrected directions. Comparison with the expected Miocene direction shows no discordance; comparison with a a late Cretaceous expected directionshows similar inclination but significant counterclockwise rotation of 24.6° ± 19.8° .
20
other sites record shallower than expected inclinations, and declinations west of the
expected north. Because all sites are inferred to be Tertiary and the location of the
Alexander terrane is well constrained to its present day location by that time, significant
northward transport is not viable. All paleomagnetic discordance must therefore be due to
crustal block tilting or rotations.
Structural corrections are performed on all sites by rotating to horizontal based on
the local strike and dip of the Kootznahoo Formation. For the sills and dikes, inclination
changes are particularly sensitive to the dip amount, and to determine confidence
intervals at least an additional ± 2° should be added to reflect the uncertainty in the strike
and dip of the Kootznahoo Formation. By doing so, all structurally corrected dikes and
sills on Kupreanof Island fall within error of the expected paleomagnetic direction (Table
3, Fig. 8 F).
For the Port Camden flows sampled at sites 09JDE06, 09JDE07, and 09JDE08,
the average in-situ declination is about 45° west of expected, and inclination is 10° too
shallow. There are tens of meters of sedimentary and volcaniclastic sediments between
each flow, and site 09JDE08 is reversed, making it likely that a large amount of time is
represented between flows on an order greater than 105 years. This suggests that an
average of the sites’ directions provides an accurate measure of position and successfully
averages out measures of secular variation typical of directions sampled from individual
flows. Correcting the flows based on the bedding of the local Kootznahoo Formation
eliminates the inclination discordance, but still leaves the 45° counterclockwise
declination gap unresolved.
21
Finally, one sill (09JDE16) intruded by dike 09JDE15 was sampled that yielded
unusual directions of magnetism. Three cores yielded the type 5 magnetism with
extremely low coercivity and low inclinations that are characteristic of a transitional pole.
However, the three other cores yielded directions of magnetization which all had reversed
polarities and independently had extremely low MAD, but gave paleomagnetic directions
that were completely inconsistent between cores. Notable at this site were what we
documented in the field to be flame or mullion structures and evidence of soft sediment
deformation. Given this field relationship, it is possible that the sill cooled past its
blocking temperature and then experienced rigid body flexure. This would scatter
otherwise consistent paleomagnetic directions and yield the discordant directions found.
TECTONIC IMPLICATIONS
Paleomagnetic results from this study, which suggest local tilting and vertical-axis
rotations up to 46.3 ± 8.3°, seem to be at odds with previously published paleomagnetic
data from Haeussler et al. (1992), who found insignificant evidence of Tertiary tilting or
counterclockwise rotation anywhere in the Keku Strait. However, close inspection of
their data reveals systematic differences in paleomagnetic directions based on locality,
and it appears only fortuitous that the disparate directions averaged to yield no directional
discordance. Parsing their data by location, data from both studies are consistent with
local tilting and heterogeneous vertical-axis rotations.
The directional data on Point Hamilton of sites 09JDE03&04 are consistent with
those from Haeussler et al. (1992), and are not statistically distinguishable from dikes
09JDE12 and 09JDE13. All sites, including all of the Haeussler et al. (1992) sites, have
paleomagnetic directions that differ from those expected of a Miocene pole, with
22
inclinations too shallow by about 8° and declinations about 18° counterclockwise from
expected. Performing tilt-corrections based on the strike and dip of the local Kootznahoo
Formation yields directions within error of the expected early Miocene directions.
Haeussler et al. (1992) found a 23 Ma age for this sill, which constrains the occurrence of
tilt to the Neogene. This suggests most if not all of the current dip of the Kootznahoo
Formation can be attributed to this tilting. Furthermore, this suggests that the portion of
Kupreanof Island near Point Hamilton and Big John Bay deformed as a unit, and at least
in recent history represents a single crustal block.
Directional data from the basaltic flows of Port Camden beach, sites 09JDE06-08,
are distinct from all other sites and localities except for site HE66, a 100 m thick dike that
cuts through the Kootznahoo Formation across from Port Camden beach. Without
performing a tilt-correction, the data suggests ≈45° of counter-clockwise rotation and a
northward transport of 15° in the Neogene, which is untenable. Performing a tilt-
correction based on the local strike and dip of the Kootznahoo Formation yields a latitude
consistent with that expected, but still leaves the ≈45° of counterclockwise discordance
unresolved.
Without alternative, it appears likely this paleomagnetic discordance is due to
simple counterclockwise rotation about a vertical axis. If so, it is likely this rotation
represents deformation local only to Kuiu Island, as all other sites sampled in this study
do not require any vertical-axis rotation when compared with a Miocene pole. A
comparison of strike directions of the Kootznahoo Formation at Port Camden beach with
the average strike of the Kootznahoo Formation appears to support this conclusion of
variable rotation: while the average strike of the Kootznahoo Formation is 086° (Muffler,
23
1967), the average strike of the sampled area of Port Camden beach is ≈050°, which
suggests a counterclockwise rotation of ≈36°, which is consistent with the rotation
calculated from the paleomagnetic data.
Because all other sites sampled in this study require a tilt-correction but not a
rotation correction, it seems likely that response to tectonic deformation is
heterogeneously accommodated in portions of the Kootznahoo Formation, with the
Kupreanof Island block experiencing tilt-only deformation and the Port Camden block
experiencing both tilt and counter-clockwise rotation.
However, additional complexity must be considered due to the directional data
from the Point Camden sill, including 09JDE11 (Fig. 8 G,H) and HE 67, 68, 70. In-situ
directions show declinations within error of the expected Miocene direction but
inclinations significantly too shallow. Structural correction based on the strike and dip of
the local Kootznahoo Formation corrects data from this study to within the expected
direction but leaves about ≈10° of inclination discordance in the Haeussler et al. (1992)
data. Because this sill is within two kilometers of the Port Camden flows, this suggests
that counterclockwise rotation is localized only to Port Camden flows, or possibly that
the Port Camden flows were extruded and rotated before emplacement of the sill, and
then both were tilted.
Alternatively, an arbitrary comparison of the sill with a Cretaceous pole
(Mcelhinney, 1979) shows insignificant inclination discordance but considerable
counterclockwise rotation 24.6 ± 19.8°, which is consistent with the vertical axis
counterclockwise rotation of the nearby Port Camden flows. Such an interpretation would
suggest crustal block tilting and rotation of the entire portion of the Point Camden
24
Peninsula of Kuiu Island; however, this is contingent upon a late Cretaceous age for the
yet-undated Point Camden gabbro. Field relationships of this sill do not show conclusive
evidence of intrusion into the Kootznahoo Formation, and it is at least possible that the
sill could be Cretaceous in age. In the discussion of the Haeussler et al. (1992) data that
follows, I will make an argument for a late Cretaceous age for the sill.
The re-interpretation and regrouping of the data of Haeussler et al. (1992) satisfyingly
explains directional data of all the Tertiary data on Kupreanof Island, including the Point
Hamilton sills, and dikes and the Big John Bay dikes, and is consistent with directional
data from the Port Camden flows and Point Camden sill. However, a considerable
amount of the Haeussler et al. (1992) data with paleomagnetic directions significantly
different from those just considered are now left in need of explanation. I will now
address this data by locality and address their tectonic implications.
KEKU ISLETS
The paleomagnetic directional data from the gabbros on the Keku Islets are
significantly different from the ChRM of all other locations. However, the directions are
consistent with a thermal overprinting found elsewhere in the Keku Islets, including that
found in the Triassic Hound Island Volcanics and the Permian Pybus Formation that it
intrudes. Because Haeussler et al. (1992) compared the directions of the overprinting to
the average direction of all gabbro in the region and not the directions specific to the
Keku Islets, they considered it less likely that the overprinting was caused by these
gabbro, and instead favored a Cretaceous overprinting. Comparison of the directional
data from the gabbro on the Keku Islets with a Miocene pole suggests neglible
displacement but considerable clockwise rotation ≈34°. Considering Haeussler et al.’s
25
(1992) suggestion that the overprinting might be Cretaceous in age, comparison with a
late Cretaceous pole of 72.3° N, 194.8° W, A95 = 3.7° (Mcelhinney, 1979) shows
directions are consistent within ±1° of the expected direction. Because the youngest rocks
this gabbro intrudes are Triassic, without further data it is at least possible that the gabbro
is Cretaceous in age. We are thus left with two alternatives: (1), the gabbro on Keku
straight is Tertiary in age, and it and the nearby Pybus Formation and Hound Island
volcanic on the Keku Islets have experienced significant ~34° clockwise rotation, or (2)
the gabbro is late Cretaceous in age, and has negligible rotation and displacement.
Resolution of either of these hypotheses requires further geochemical and geochronologic
study.
HARE ISLANDS AND HOUND ISLAND OVERPRINT
Directional data from Hare Islands are consistent with a component of thermal
overprinting of the Hound Island Triassic volcanics found on the north of Hound Island
(Haeussler et al., 1992). Haeussler et al. attribute this overprinting to the aggregate
Tertiary “Point Camden” gabbro principally by noticing that the most local outcrop of the
gabbro, the Hare Islands gabbro, has directions that match the overprint. However,
suggestion that this overprinting is Tertiary in age does not seem valid, as paleomagnetic
declinations on the Hare Islands are almost 100° counterclockwise from those expected
of a Tertiary intrusion. Its seems unlikely that the gabbro on Hare Islands has undergone
≈100° of counterclockwise rotation in the Neogene. However, performing a structural
correction on both the overprint of the Hound Island volcanic and the Hare Island gabbro
based on the local strike and dip of the Hound Island volcanics (~325/12), directional
data exactly matches that of the Keku Islets and of a late Cretaceous pole. Again, it could
26
still be the case that the Keku Islets gabbro is Tertiary and requires a ~34° clockwise
rotation, but if so, it would suggest (1) a regional overprinting caused by the Keku Islet
gabbro, (2) differential tilting of the Hare and Hound Islands but not the Keku Islets
suggesting they are on different blocks, and (3) rotation of both ~34° clockwise, with
both blocks moving together. A Cretaceous age of overprint would not require this
rotation, nor require probable separate blocks first experiencing differential and then
synchronous deformation.
POINT CAMDEN GABBRO AND CLARK ISLAND
Directions from the Point Camden gabbro are consistent with a component of
thermal overprinting found in the Triassic Hound Island volcanics found on Kadake Bay
Point on Kuiu Island (Haeussler et al., 1992). Haeussler et al. (1992) compared this
overprint to the aggregate “Point Camden” gabbro, and finding discordance, suggested
that the overprinting was similar with that of the overprinting signature found at
Turnabout Island and Cape Bendal on Kupreanof Island, which are geographically the
sites farthest away from those on Point Kadake Bay. Using Ar40/Ar39 dating from
Turnabout Island in the Frederick Sound, they calculate an age of overprint to be 90-100
Ma, or late mid-Cretaceous. Grouping sites from the disparate locations together, they
also note that ‘all’ sites are of normal polarity, which they suggest is consistent with the
normal polarity Superchron in the mid-Cretaceous from 120-85 Ma.
However, there are three reversals found in the Point Kadake Bay overprint and
all directions on Point Kadake have directions significantly different from those of
Turnabout Island and Cape Bendal at greater than 99.9% confidence. Indeed, four sites
from the Point Camden sill and Kadake Bay Point overprint (sites HE 54, 61, 63, 65)
27
have reversed polarities, meaning they cannot be from the normal Cretaceous superchron
and not 120-85 Ma.
Keeping with the hypothesis of a Cretaceous overprint, I would like to instead
propose a later date for the overprint—namely, late Cretaceous. Revisiting the hypothesis
that the Point Camden gabbro may be of such an age, I apply the same structural
correction I performed on 09JDE11 based on the average strike and dip of the
Kootznahoo Formation on Kuiu Island to the rest of the Point Camden gabbro and all of
the overprint. Based on my preliminary analysis, such a correction yields directions D ~
10°, I ~ 62°, which is 10° clockwise of an expected Miocene pole but 24°
counterclockwise of a late Cretaceous pole. Furthermore, comparison of the mean strike
and dip of the Hound Island volcanics on Kuiu Island with those on Hound Island show
~45° of counterclockwise rotation. This suggests that a considerable portion of Kuiu
Island may have as a crustal block experienced Neogene tilting and significant
counterclockwise vertical axis rotation. While the large crustal block scenario is
appealing, I suggest radiometric dating of the Point Camden, Hare, Clark, and Keku
Islands gabbro in conjunction with dating of the Hound Island overprint and additional
paleomagnetic sampling, particularly on Kuiu Island, is in order. In particular, the
unresolved ~ 10° inclination discordance present in any scenario presented is
unsatisfying—and hopefully additional data will help resolve the discordance.
DISCUSSION OF REGROUPING OF THE HAEUSLLER ET AL. (1992) DATA
This section details the geochemical, petrographic, and paleomagnetic rationale
for the regrouping of the paleomagnetic data of Haeusller et al. (1992) from the aggregate
“Point Camden” gabbro based on locality.
28
REGARDING COOLING RATES
Haeussler et al. (1992) found that for their samples from “Point Camden” gabbro,
typical laboratory unblocking temperatures were often small, between <40-50° C. They
suggest that if the rocks cool through this temperature range in a few thousand years or
less, then a single specimen or locality would not completely average secular variation
(Haeussler et al., 1992). While sills up to 500 m thick like the “Point Camden” gabbro
(Muffler, 1967) are not expected to cool this quickly, it is possible that hydrothermal
convection of meteoric groundwater through propagated crack systems might have
greatly expedited the cooling process (Dodson et al., 1978). If so, it is possible that the
gabbro cooled through the blocking temperatures of 580° to 500° in as few as 500 years,
making it likely that a particular locality or site would record secular variation and an
average of magnetic directions across disparate localities necessary to compute mean
VGP’s (Haeussler et al., 1992).
Cooling is not likely to be as rapid as Haussler et al. suggest for 3 reasons. First,
while petrographic analysis of most localities indicate significant hydrothermal alteration
as evidenced by the presence of chlorite, iddingsite, and serpentine, it is not possible to
determine if this alteration is primary or secondary. Furthermore, one site, 09JDE10,
shows no evidence of alteration in thin section, suggesting that rapid cooling by
circulating groundwater did not occur at that site.
Second, at three sites, 09JDE10, 09JDE14, 09JDE15, distinct (09JDE10 and
09JDE14 show similar geochemical signature and may be identical) dikes were sampled.
At two of these sites, 09JDE10 and 09JDE14, type 5 magnetizations were recorded,
suggesting transitional poles. At site 09JDE15, however, 6 cores were sampled across the
29
width of the dike, with each core representing a different intrusion event. While two of
the cores yielded type-5 magnetizations typical of transitional poles, four magnetizations
consistent with an Oligocene-Miocene pole were recorded. Most notably, one of these
cores had a reversed polarity (the other three were normal), suggesting that there were at
least two distinct intrusion events, that they did not cause significant overprinting of
ChRM, and that magmatism persisted over a prolonged period of time representing at
least 105 years. Most importantly, the consistency of the magnetization directions of the
four cores, including both reversed and normal polarities, suggests that the intrusions
cooled slowly enough to average secular variation. The fact that these intrusions were on
average only five cm wide further suggests that directions from other sills and dikes in
the region, which varied from one to hundreds of meters thick, also cooled slowly enough
to average secular variation.
Finally, site 09JDE11 was sampled from the chill margin of the Point Camden
gabbro. Geochemical analysis indicated an incredibly high silica content (SiO2 ≈ 70%)
and elevated trace elements of La, Nb, Rb, and Zr, which suggests localized melting and
incorporation of the nearby clastic formation. Petrographic analysis supports this
interpretation, with extremely small quartz grains surrounding all other minerals,
including large euhedral lath plagioclase crystals that are clearly magmatic in origin. This
“chill margin” is 3.5 meters thick, and suggests elevated temperatures well above the
blocking temperature of 580° for prolonged periods of time, making rapid cooling in a
period of fewer than 500 years unlikely. Finally, I should add these elevated temperatures
make a hydrothermal cooling convection system unlikely, as groundwater is unable to aid
30
in cooling until there is a solidified margin in which cracks can open to allow circulation
(Taylor, 1971).
For these reasons, the sills and dikes likely cooled slowly enough to average out
secular variation, and that any significant difference in magnetizations across sites must
be the result of geologic, structural, or paleomagnetic variation.
CONCLUSION
The recent tectonic history of the Keku Strait is decidedly complex. Directional
data found in this study provides evidence of Neogene regional tilting based on the strike
and dip of the Tertiary Kootznahoo Formation, and locally suggest up to 45° of
counterclockwise vertical-axis rotation. A re-evaluation of Hauessler et al. (1992)
supports the conclusion of Neogene tilting, and suggests that deformation in the
Alexander terrane is highly variable and locally significant. Such deformation should be
considered when determining paleomagnetic directions for older rocks in the region,
which have been used to constrain the location of the Alexander terrain in relation to the
Baja BC hypothesis. A final implication of the paleomagnetic results from this study is
that it is unlikely the Kootznahoo Formation had a significant primary dip, making it
unlikely that it was the type of alluvial fan that Loney (1964) suggests.
ACKNOWLEDGEMENTS
I gratefully acknowledge all the people who helped me pull this research
together—Michael Jackson, at the Institute of Rock Magnetism, for his help with
paleomagnetic work, Jeff Thole at Macalester for his assistance with laboratory
processing of majors and traces, Karl Wirth, also at Macalester, for insight into
geochemical analysis, and Tim White, Sue Karl, and Peter Haeussler, for laying down the
31
groundwork that made this research possible. Special thanks goes Maria Princen, Alex
Gonzalez, and Cameron Davidson, who made excellent companions drilling in the field
and in review of this manuscript. Finally, I’d like to thank Sarah Crump, for her help with
paleomagnetic analysis, and Nate Evenson, for his invaluable openness to vetting of
ideas, and to the KECK Consortium, for funding of this project.
REFERENCES Andronicos, C., Hollister, L.S., Davidson, C., and Chardon, D., 1999, Kinematics and tectonic significance
of transpressive structures within the Coast Plutonic Complex, British Columbia: Journal of Structural Geology, v.21, p. 229-243.
Beck, M.E., Jr., 1976, Discordant paleomagnetic pole positions as evidence of regional shear in the western
Cordillera of North America: American Journal of Science, v. 276, p. 694-712. Beck, M.E., Jr., Burmester, R.F, and Schoonover, R., 1981, Paleomagnetism and tectonics of Cretaceous
Mt. Stuart batholith of Washington: Translation or tilt?: Earth and Planetary Science Letters, v. 56, p. 336-342.
Brew, D.A., Ovenshine, A.T., Karl, S.M., and Hunt, S.J., 1984, Preliminary reconnaissance geologic map of the Petersburg and parts of the Port Alexander and Sumdum 1:250,000 quadrangles, southeastern Alaska: U.S. Geological Survey, Report: OF 84-0405, 43.
Butler, R.F., Gehrel, G.E., McClelland, W.C., May, S.R, and Klepacki, D., 1989, Discordant paleomagnetic
poles from the Canadian Coast Plutonic Complex, regional tilt rather than large-scale displacement?: Geology, 17, 691-694.
Butler, R.F, Gehrels, G.E., Hart, W., Davidson, C., and Crawford, M.L., 2006, Paleomagnetism
of late Jurassic to mid-Cretaceous plutons near Prince Rupert, British Columbia: Geological Association of Canada, Special Paper, vol. 46, pp. 171-200.
Butzer, C., Butler, R.F., Gehrels, G.E., Davidson, C., O’Connell, K., and Crawford, M.L., 2004,
Neogene tilting of crustal panels near Wrangell, Alaska: Geology, v. 32, no. 12, p.1061-1064. Coe, R.S., Globerman, B.R., Plumley, P.W., and Thrupp, G., 1985, Paleomagnetic results from Alaska and
their tectonic implications; Tectonistratigraphic Teranes of the Circum0Oacific Region, Earth Science Series, v. 1, edited by D.G. Howell, p. 85-108, Circum Pacific Council for Energy and Mineral Resources, Houston, Texas.
Cowan, D.S., Brandon, M.T. and Garver, J.L., 1997, Geologic tests of hypotheses for large coastwise
displacements—a critique illustrated by the Baja British Columbia controversy: American Journal of Science, v. 297, p. 117-173.
Cowan, D.S., 2003, Revisiting the Baranof-Leech River hypothesis for early Tertiary coastwise
transport of the Chugach-Prince William terrane: Earth and Planetary Science Letters, v. 213, p. 463-475.
Demarest, H. H., Jr., 1983, Error analysis for the determination of tectonic rotation from paleomagnetic
32
Data: J. Geophys. Res., 88, 4321-4328. Diehl, J.F, McClannahan, K>M., Bornhorst, T.J., 1988, Paleomagnetic results from the Mogollon-Datil
volcanic field, southwestern New Mexico, and a refined mid-Tertiary reference pole for North America: Journal of Geophysical Research, v. 93, p. 4869-4879.
Dickinson, K.A., 1979, A uranium occurrence in the Tertiary Kootznahoo Formation on Kuiu Island,
Southeast Alaska, U.S. Geological Survey. Report: OF 79-1427. Dickinson, K.A., and Pierson, C.T., 1988, A statistical analysis of chemical and mineralogic data
from the Tertiary Kootznahoo Formation in southeastern Alaska, with emphasis on uranium and thorium, U.S. Geological Survey Bulletin, Report: B 1851, 19.
Douglass, S.L., Webster, J.H, Burrell, P.D., Lanphere, M.L., and Brew, D.A., 1989, Major element
chemistry, radiometric ages, and locations of samples from the Petersburg and parts of the Port Alexander and Sumdum quadrangles, southeastern Alaska: U.S. Geol. Surv. Open-File Rep., 89-527.
Floyd, P.A., Winchester, J.A., 1975, Magma type and tectonic setting discrimination using immobile elements. Earth Planet Sci Lett 27: 211-218.
Ford, A.B., Palmer, C.A., and Brew, D.A., 1994, Geochemistry of the Andesitic Admiralty Island
Volcanics, An Oligocene Rift-Related Basalt to Rhyolite Volcanic Suite of Southeastern Alaska, Geologic Studies in Alaska by the U.S. Geological Survey, 177-204.
Gehrels, G.E., and Berg, H.C., 1992, Geologic map of southeastern Alaska: U.S. Geological
survey, Report I-1867, 24.
Gehrels, G.E., and Berg, H.C., 1994, Geology of southeastern Alaska: The Geology of North America, v. G-1, The Geology of Alaska, The Geological Society of America, p. 451-467.
Haeussler, P.J.,Coe, R.S., and Renne, p., 1992, Paleomagnetism and Geochronology of 23 Ma Gabbroic
Intrusions in the Keku Strait, Alaska, and Implications or the Alexander Terrane: J. Geophys. Res., v. 97, p. 19617-19639.
Haeussler, P.J.,Coe, R.S., and Renne, p., 1992, Paleomagnetism of the Late Triassic Hound Island
Volcanics—Revisited: J. Geophys. Res., v. 97, p. 19641-19649. Haeussler, P.J., Bradley, D.C., Wells, R.E., Miller, M.L., 2003, Life and death of the
Resurrection plate: Evidence for its existence and subduction in the northeastern Pacific in Paleocene-Eocene time: GSA Bulletin, v. 115, no. 7, p. 867-880.
Hagstrum, J.T., Sawalan, M.G., Hausback, B.P., Smith, J.G., and Gromme, C.S., 1987, Miocene
Paleomagnetism and tectonic setting of the Baja Peninsula, Mexico: J. Geophysical Research., v. 92, 2627-2639.
Irving, E., Woodsworth, G.H., Wyne, P.J., and Morrison, A., 1985, Paleomagnetic evidence for
displacement from the south of the Coast Plutonic Complex, British Columbia: Canadian Journal of Earth Science, v. 22, p. 584-598.
Irving, E., Souther, J.G., and Baker, J., 1992, Tertiary extension and tilting in the Queen Charlotte Islands:
Evidence from dyke swarms and their paleomagnetism: Canadian Journal of Earth Seciences, v. 29, p. 1878-1898.
Karl, S.M., Haeussler, P.J., and McCafferty, A.E., 1999, Reconnaissance Geologic Map of the
Duncan Canal/Zarembo Island Area, Southeastern Alaska: United States Geological Survey, Open-file Report 99-168.
33
Le Bas, M.J., Le Maitre, R.W., Steckeisen, A., and Zanettin, B., 1986. A chemical classification ofvolcanic
rocks base don theTotal Alkali-Silica diagram: Journal of Petrology, v. 27, part 3, p. 745-750. Lindline, J., Crawford, W.A, Crawford, M.L., and Omar, G.I., 2000, Post-accretion magmatism within the
Kuiu-Etolin igneous belt, southeastern Alaska: Canadian Mineralogist, v. 38, p. 951-974. Lindline, J., Crawford, M.L, Crawford, W.A., 2004, BA bimodal volcanic-plutonic system: The Zarembo
Island extrusive suite and the Burnett Inlet ntrusive complex: Canadian Journal of Earth Sciences, v. 41, p. 355-375.
Loney, R.A., 1978, Stratigraphy and Petrography of the Pybus-Gambier area, Admiralty Island, Alaska:
U.S. Geologcal Survey Bulletin, v. 1178. Mcelhinney, M.W., 1979, Paleomagnetism and Plate Tectonics: Cambridge University Press. Monger, J.W.H., 1982, Tectonic accretion and the origin of the two major metamorphic and plutonic belts
in the Canadian Cordillera: Geology, v. 10, p. 70-75. Muffler, L.J.P., 1967, Stratigraphy of the Keku Islets and neighboring parts of Kuiu and Kupreanof Islands,
southeastern Alaska: U.S. Geological Survey Bulletin, 1241-C. Myers, J.D., and Marsh, B.D., 1981, Geology and Petrogenesis of the Edgecumbe Volcanic Field,
SE Alaska: The interaction of Basalt and Silic Crust: Contributions to Mineralogy ad Petrology,
Myers, J.D., Sinha, A.K., and Marsh, B.D., 1984, Assimilation of Crustal Material by Basaltic
Magma: Strontium Isoptopic and Trace Element Data from the Edgecumbe Volcanic Field, SE Alaska: Journal of Petrology, v. 25, no. 1, p.1-26.
Panuska, B.C., 1985, Reconaissance paleomagnetic study of the Eocene Admiralty Island Volcanics,
southeast Aaska: Evidence for pre-late Eocene accretion: Geological Society of America Abstract Programs, v. 17, p. 684.
Riehle, J.R., Budahn, J.R., Lanphere, M.A., and Brew, D.A, 1994, Rare earth element contents
and multiple mantle sources of the transform-related Mount Edgecumbe basalts, southeastern Alaska: Canadian Journal of Earth Sciences, vol. 31, no. 5, p. 852-864.
Stowell, H.H., Green, N.L., and Hooper, R.J., 2000, Geochemistry and tectonic setting of
basaltic volcanism, northern Coast Mountains, southeastern Alaska: Geological Society of America, Special Paper 343.
Vervoort, J.D., Wirth, K., Kennedy, B., Sandland, T., Harpp, K.S., The magmatic evolution of the
Midcontinent rift: New geochronologic and geochemical evidence from felsic magmatism: Precambrian Research, v. 157, p. 235-268.
34
