Lecture 14: Paleointensity

Transcription

1 Lecture 14: Paleointensity Key assumptions Paleointensity with TRMs Paleointensity with DRMs Paleointensity with IRMs? 1

2 Key assumptions: The proportionality function between remanence and field is known (usually assumed to be linear) The proportionality constant can be approximated in the laboratory 2

3 Remanent magnetization M lab M NRM B lab B anc Applied magnetic field B anc = M NRM M lab B lab 3

4 Sounds easy - BUT Function may not be linear Specimen may have altered capacity to acquire remanence lab NRM may not be acquired by same mechanism (constant may be different) Anisotropy of remanence acquisition NRM may be multi-component 4

5 Paleointensity with TRMs: experimental design step-wise replacement of NRM with ptrms (Thellier family of methods) step-wise replacement of NRM with microwave induced remanences replacement of NRM with total TRM and check for alteration many other methods (IRM normalization, multi-specimen approaches, new methods invented every month...) 5

6 Normalizing thermal remanence X, Z Natural Induced Y Natural B ancient = Natural Induced B lab Induced

7 Assumptions ptrms are additive (law of additivity) ptrms acquired at Tb are removed at same temperature (Tb=Tub); law of reciprocity ptrm acquired in lab is equivalent to original no lab alteration linearity assumption? cooling rate and anisotropy can be accounted for 7

8 Thermal remanent magnetization ptrms are additive - each cooling interval is independent of all others TRM ptrms Temperature ( o C) T c 8

9 law of reciprocity ptrm acquired between 350 and 370C is removed between 350 and 370C Fractional remanence remaining a) SD o o C 1 µm Temperature ( o C) 9

10 only true for SD grains Fractional remanence remaining SD o o C 1 µm Temperature ( o C) 10

11 consequences in paleointensity experiments: saggy Arai plots Fractional remanence remaining SD 0.6 µm ptrm only true slope is red line 11

12 Blocking Temperature check with IZZI or ptrm tail checks IZZI ZI IZ ZI IZ T 4 T 3 T 2 T 1 zero field in field in field zero field zero field ptrm check in field ptrm tail check in field zero field NRM demagnetized Lab ptrm 12

13 C Fraction NRM remaining ptrm gained 13

14 problem of alteration during experiment 100 o C δ can be checked with ptrm checks 14

15 ideal - accurate requires subjective judgment a Ideal b High β, high DRATS c Low f; f vds NRM / NRM C mk107101a TRM / NRM C mk109301a TRM / NRM C 200 NRM 0 =9.09e-08 Am 2 NRM 0 =3.65e-07 Am 2 NRM 0 =8.39e-08 Am mk100702g TRM / NRM 0 Need a magic mix of selection criteria: exclude the bad data select the good data

16 Arai plot: ancient : field b B lab where b is the slope standard error: b scatter: FRAC: fraction of NRM used Zijderveld plot: MAD: directional scatter Krv = b b : curvature DANG: deviation from origin a Ideal b High β, high DRATS c Low f; f vds NRM / NRM C mk107101a TRM / NRM C mk109301a TRM / NRM C 200 NRM 0 =9.09e-08 Am 2 NRM 0 =3.65e-07 Am 2 NRM 0 =8.39e-08 Am mk100702g TRM / NRM 0

17 How things go wrong failure of single domain assumption failure of linearity assumption anisotropy of TRM cooling rate

18 Whatever method you use Must build in checks of fundamental assumptions Must have some tests for quality assurance (statistics, acceptance criteria) 18

19 State of the database: 19

20 Over 20,2000 sites with absolute paleointensity in database (method codes: LP-PI and NOT LP-PI-REL) 20

21 Changes through time? CNS 21

22 2013 MagIC database: 0-5 Ma N=2180 OOPS!

23 Worse news! Data from a single lava flow span the entire range of intensities on Earth! Paleointensity (μt) This Study (IZZI) de Groot et al., 2013 (IZZI, MS, PT) Böhnel et al., 2011 (MS) Morales et al., 2010 (TC) Herrero-Bervera and Valet, 2009 (TC) Chauvin et al., 2005 (TT) Yamamoto et al., 2003 (TC, S) Hill and Shaw, 2000 (MT) Valet and Herrero-Bervera, 2000 (AF) Tanaka et al., 1995b (TT, KT, KU) Tsunakawa and Shaw, 1994 (S) Tanaka et al., 1991 (TC) Abokodair, 1977 (TT) Kilauea Lava Flow (samples ordered by publication) Cromwell et al. (submitted)

24 Without measurements No way to quantify alteration tests because there is no standard ptrm check statistic No way to tell if non-linear Arai plot or even what fraction of data are used No way to tell if MD sag is present No way to make sure in linear TRM range No way to know if cooling rate or anisotropy are problems 24

25 With measurements Can use SCAT to filter out altered, scattered data Can use Krv to quantify sagging. Can use IZZI method to quantify ptrm tails Can repeat total TRM as a function of field to test for non-linearity of TRM acquisition Can test for cooling rate dependence and correct for anisotropy 25

26 Comparing two sets of selection criteria on results of known answer CCRIT: Cromwell et al TTA: Leonhardt et al. 2004

27 Paleointensity with DRMs: same two key assumptions The proportionality function between remanence and field is known (usually assumed to be linear) The proportionality constant can be approximated in the laboratory 27

28 In graphical form: 28

29 Usual procedure: Establish (or assume) DRM origin of ChRM Establish magnetic carrier (should be magnetite!) Establish homogeneousness of bulk properties (linear relation between ARM, IRM, susceptibility) to rule out changes in magnetic grain size Establish maximum bounds for changes in concentration (no more than ~10x) Choose bulk normalizer (Rarely) compare nearby records 29

30 State of the database: First the good news 30

31 Over 100 relative paleointensity papers in database 31

32 Examples of long records from deep sea sediments 32 Ziegler et al. (2011)

33 RPI works? stack of long RPI records (Ziegler et al., 2011) Relative paleointensity magnetic anomaly inversion (Gee et al., 2000) Age (ka) Roberts et al., 2013

34 How things go wrong failure of linearity assumption sensitivity of DRM to environmental conditions (salinity, particle size, mineralogy)

35 Where things can go wrong: (a) DRM/sIRM μm (b) DRM/sIRM (%) ± ± ± ± ppt 5 ppt μ B ( T) B (μt) Intensity (ma/m) Roberts et al. Figure Salinity (ppt) Inclination ( )

36 In database No way to know effect of flocculation on relative paleointensity No way to make sure in linear TRM range No consistent approach to data quality 36

37 In summary Theory is getting better and better Need to apply theoretical understanding to data selection Need access to all the experimental data - not just what gets published. 37