GP43C-01 INVITED
Sample Inhomogeneity and its Effects on Multispecimen Paleointensity Methods
There are various mechanisms that cause errors in absolute paleointensity determinations, most importantly, remanence carried by multidomain particles and chemical alteration during the experiment. In order to reduce both these errors, the multispecimen parallel differential pTRM method has been put forward. The use of multiple specimens of one given sample, however, can introduce a substantial degree of data scatter due to small specimen sizes and therefore non-representative statistical sampling of the characteristic blocking temperature spectrum. Although typical specimen volumes of around 1 cc are large enough in the case of randomly dispersed magnetic single domain particles to form a representative sample of the particle population, this is not necessarily true for assemblages of larger magnetic mineral particles or in the case of non-randomly dispersed, i.e., clustered particles. A review of typical textures of volcanic rocks shows that such clusters of remanence carriers are quite common due to various reasons such as titanomagnetite oxy- exsolution or the presence of fine-grained magnetic mineral inclusions in larger silicate mineral grains. A statistical model will be presented to assess the effect of small specimen volumes on the results of multispecimen paleointensity methods.
GP43C-02
Paleointensities of young lava flows obtained with the multispecimen parallel differential pTRM method
Paleointensity experiments have been carried out with the multispecimen parallel differential pTRM method on rocks from historic and recent lava flows, on a total of more than 25 flows. The results are compared with observatory measurements of the geomagnetic field intensity and data obtained using other paleointensity methods like Thellier-type, microwave, and other methods. The purpose of these comparisons is to evaluate how exact the multispecimen methods reproduces the known paleointensity, and how it compares to other paleointensity methods in terms of the success rate and obtained paleointensity.
GP43C-03
Multispecimen and temper archeomagnetic studies: Application to Iron Age sites from southern Africa
Here we compare methods of paleointensity analysis as applied to pottery shards. We prefer the Thellier technique because of its foundation in basic rock physics and its associated reliability checks. Nevertheless, we recognize that additional techniques may be useful on problematic samples. We start with multiple specimen techniques. These are attractive because they can potentially address the presence of multidomain magnetic carriers. In addition, pottery shards can often be cut into multiple samples with nearly identical physical properties, and the magnetic component structure is often simple. Nevertheless, grain size effects appear to be present in synthetic samples used to test multi-specimen methods, and selection/rejection criteria are not fully developed. Irrespective of these issues, we find that results from multi- specimen experiments which yield linear results over a broad range of applied fields tend to agree well with results from Thellier runs. The results of other samples are more difficult to interpret, with apparent non- linear behavior at high and low-field values. We also introduce a new paleointensity approach: analysis of temper from pottery shards. This method is an extension of single crystal techniques, and relies on the addition of grains to the pot matrix prior to firing. As is the case in the analysis of lavas, careful separation and selection of grains is needed as temper is diverse between archeological areas. We apply these methods to a new collection of pottery shards from southern Africa, in hopes of constructing a Southern Hemisphere paleointensity record.
GP43C-04
Testing Determinations of Absolute Paleointensity from the 1955 and 1960 Hawaiian Flows.
We have performed a new exhaustive paleointensity study of the flows erupted in 1955 and 1960 from the Kilauea (Big Island of Hawaii. We have used the Coe-Thellier procedure in order to constrain the origin of the dispersion that has been reported in most studies so far. We have observed a direct relationship between the unblocking temperature spectra and the suitability of the paleofield estimate, which points out the importance of problems generated in presence of more than one magnetic phase, like titanomagnetite and magnetite, as is the case in most samples from the 1955 flow and several from the 1960 flow. It is striking that, despite good quality linear Arai plots with a unique slope, determinations of paleointensity can be heavily biased, and significantly deviate from the expected field by up to 40 per cent. The results indicate that only the samples containing magnetite with a narrow range of high unblocking temperatures provided accurate determinations within less than 10 per cent from the expected field value. This implies that no less than 60 to 70 per cent of the magnetization should be included in the determination of paleointensity. We observed that such conditions are met for the samples taken in the middle of the flow which led us to propose that this part of lava flows is the most appropriate for paleointensity studies. We also reiterate the importance of performing pTRM checks at all steps of the experiment. These conditions impose more restrictions but this selection would certainly reduce the uncertainties and improve our knowledge of the past geomagnetic field.
GP43C-05
Toward a new non-heating method of determining absolute paleointensities
Most methods of determining absolute palaeointensities from rocks, require either multiple heating steps, e.g., the many modifications of the Thellier method, or require a single high-temperature heating step as part of the normalisation process. Due to chemical instabilities of the magnetic minerals in rocks, this heating commonly alters the minerals under investigation giving rise to incorrect palaeointensity estimates. Various checks have been introduced to correct for this, but the experiments are time consuming and inefficient. Other lower temperature heating methods have been proposed, e.g., the multispecimen parallel differential pTRM method, which has met with some success, as has the microwave palaeointensity determination protocol. We have developed a new non-heating method based on Preisach theory for determining absolute palaeointensities. From a measured Preisach distribution it is possible to predict thermoremanence (TRM) acquisition versus field for a given sample. Simply comparing the measured natural remanence magnetisation (NRM) with the TRM versus field curves yields the ancient geomagnetic field intensity. We have tested this method on a suite of historical lavas from various locations.
GP43C-06
Volcanic Glass - an Ideal Recording Material?
It has been suggested that volcanic glass is a perfect material for paleomagnetic research, especially for paleointensity studies, as it is often pristine and contains a magnetic fraction in the SD range. Furthermore, all biasing effects which are usually hard or impossible to detect in other recording media, are either absent or can be corrected for by using mineralogical and rock magnetic constraints. To verify this supposed ideal character, silicic volcanic glass from Pantelleria, Italy, from Tenerife, Spain, and from Mayor Island, New Zealand, was analyzed. Paleointensity measurements and rock magnetic experiments were conducted, including determinations of the anisotropy of thermoremanence. Structural properties of volcanic glass, particularly the glass transition and the natural cooling rate across this transition were investigated by relaxation geospeedometry. By additionally determining the magnetic cooling rate dependency of the thermoremanence, a correction of paleointensity data for cooling rate effects could be applied. Samples from Pantelleria and Mayor Island show very low NRM intensities, hampering rock magnetic measurements. Nevertheless, paleointensity results (~40~μT) were obtained for 50% of the Pantelleria samples though the scatter is remarkable (30 to 50~μT). Intensity measurements of almost all Mayor Island samples agree very well and yield a paleointensity of ~55~μT. Anisotropy and cooling rate corrections for Pantelleria and Mayor Island could not be applied. The glass transition of these samples is close to their blocking temperature spectra. Thus, during repeated heating to higher temperature the samples enter the transition interval and alter. The reasons and consequences of this alteration are investigated. NRM intensities are much larger for Tenerife samples. Rock magnetic experiments indicate low Ti- or even pure Magnetite in the SD range. Very high success rates for paleointensity experiments are observed. An application of anisotropy and cooling rate corrections is possible as glass transition temperatures are well above the magnetic blocking temperature. High success rates and well defined paleointensity data, accompanied by the possibility to apply reliability tests and corrections, support the high suitability of silicic volcanic glass for paleomagnetic measurements.
GP43C-07
Vibrating Sample Magnetometer (VSM) Paleointensity Determinations
Determining the absolute intensity of Earth's magnetic field from rocks has proven to be fraught with pitfalls
that paleomagnetists have sought to overcome with carefully crafted but often tediously time-consuming
measurement protocols. Without at least some form of protocol, little confidence is given to a paleointensity
determination, and even with rigorous protocols, doubt can remain as to the validity of the determination.
Most protocols involve heating, cooling, and then measuring the remanence of a sample in and out of an
applied magnetic field and then repeating this process at progressively higher temperatures. Completion of
an experiment takes considerable time for a reasonable number of temperature steps and requires a large
amount of manual sample manipulation. We have investigated whether it is possible to improve the speed
while increasing the number of thermal steps by about an order of magnitude by using a vibrating sample
magnetometer (VSM). In our test, we use a MicroMag 3900 VSM produced by Princeton Measurements
Corporation. This VSM includes a furnace capable of heating samples to 800°C and an automated
rotating head that allows the magnetic moment of a sample to be measured in multiple orientations within a
plane. Although this instrument was not designed for paleointensity determinations, it has several features
that facilitate them. The instrument is sensitive enough to measure basalt samples that weigh only a few
milligrams to a few tens of milligrams, measurements can be made while the sample is being heated, the
sample does not need to be manually manipulated once the experiment starts, heating or cooling of the small
samples can be accomplished quite rapidly (from about a minute to a few minutes), and the sample is bathed
in helium gas while being heated, which helps reduce alteration. So far, we have obtained highly accurate
and precise VSM paleointensity determinations on basalt samples in experiments that were completed in only
a few hours. We will discuss some of the advantages and disadvantages of the VSM paleointensity method
and show results obtained for the 1960 Hawaii basalt flow and for oceanic basalt samples from ODP Leg 200
(Site 1224) and IODP Expedition 206 (Site 1256).
http://paleomag.ucdavis.edu/paleoint-test.html
GP43C-08
Rock magnetic properties and palaeointensity determinations on Santorini historical lava flows
A detailed rock magnetic study of eight historical dacitic lava flows from Santorini, covering the time interval between 46 AD and 1950 AD, is presented along with palaeointensity determinations encompassing different experimental approaches such as: the classical Thellier-Thellier method (Thellier and Thellier 1959), Coe's version of the Thellier-Thellier method (Coe 1967), Shaw's alternating field method (Shaw 1974), the Microwave Thellier-Thellier approach (Walton et al. 1992, 1993) and the continuous high temperature magnetisation measurement method recently developed by Le Goff and Gallet (2004). The obtained palaeointensity results are compared with the Greek data base and the International Geomagnetic Reference Field (IGRF). Eventually, non-dipole field variations at Athens during the past 2000 years are discussed and compared with those at Paris. Thermomagnetic and isothermal magnetisation curve measurements reveal the presence of two physically distinct magnetic mineral phases with two different Curie-temperatures at around 280 and 500 °C. A rather low degree of magnetostatic interaction is observed and does not change considerably during heating. The dacitic lavas show excellent thermo-chemical stability as reported from backfield curve spectra analysis, and noticeable changes in the remanent coercive force spectrum occur only when heating above 450 °C. Magnetisation decay experiments confirm that superparamagnetic grains are practically absent. The magnetic grain size is assumed to be in pseudo-single- domain, because of considerable transient magnetisation contributions and typical PSD Day-plot parameters. Palaeointensity analyses give rather inconsistent results when comparing different methods. Results from Shaw's alternating field method seem to be biased due to considerable differences in coercive force spectra between NRM and TRM, and microwave results are the most scattered. In contrast, Thellier-Thellier and Coe- Thellier results are more consistent and are comparable with the IGRF estimates and Greek data obtained from archaeological samples. Dipole field intensity calculations and experimentally obtained dipole moments for Athens are fairly in agreement except for the time interval between 500 and 1000 AD. The discrepancy between theoretical and experimental results during this interval is explained by a lack of experimental data. The non-dipole fields at Athens and at Paris resemble well each other and the amplitude of variations is comparable with historical field records.
GP43C-09
The Implications of Varying Oxidation Parameters on Microwave Derived Palaeointensity Estimates.
Establishing whether the remanence recorded in a sample is the original thermo-remanent magnetisation is paramount to any palaeointensity investigation. Detailed rock magnetic, microscopy observations and palaeointensity experiments are discussed, in an attempt to establish a correlation between oxidation parameters and derived palaeointensities from sister samples. The effect of low temperature oxidation on iron-oxide crystals is relatively easy to identify with the aid of an SEM as brittle deformation results in distinct cracks along maghaemised regions. Establishing the effect of this oxidation on the remanece recorded within a sample however is not so clear. Pronounced oxidation is evident on thermo-magnetic plots by the inversion of titanomaghaemite, resulting in a more strongly magnetic phase with a Curie temperature nearer that of magnetite, yet less distinct oxidation can often go undetected using standard rock magnetic techniques. Subsequent palaeointensity estimates may fail to account for this chemical remanence contamination, resulting in an inaccurate estimate. Low temperature oxidation is widely regarded to lower the remanence recorded in a sample, possibly leading to a low bias in the palaeointensity record. High temperature oxidation, evident by the development of ilmenite lamellae within a titanomagnetite crystal host is also relatively easy to identify with SEM observations. Taking Haggerty's 1991 (Haggerty, S.E., Oxide textures - a mini atlas, Oxide Minerals: petrologic and magnetic significance, Reviews in Mineralogy, 25, Mineralogical Society of America, 129 - 219, 1991) classification of high temperature oxidation as a reference, the degree of oxidation can be determined and hence attempts made to determine the individual samples cooling history. Identifying high temperature oxidation using standard rock magnetic techniques however is more difficult. High temperature oxidation has conventionally been assumed to make a sample more suitable for palaeointensity investigation, with ilmenite lamellae subdividing larger multi domain grains into smaller, single domain behaving sections. The implication of this oxidation on palaeointensity estimates has recently been questioned however, with some proposing that samples that have undergone high temperature oxidation yield over estimates of the ancient field strength. With the overall aim of this research to determine field strength during the Cretaceous Normal Superchron, here we present a microwave based palaeointensity study of Madagascan and Chinese samples of Cretaceous age. Investigation of a single cooling unit produces consistent palaeointensity estimates for samples that have undergone high temperature oxidation. Samples taken from the same unit that have undergone low temperature oxidation (and no apparent high temperature oxidation) however yield higher, less consistent results. Further evidence supporting these observations is considered, and the implications for future palaeointensity experiments discussed.
GP43C-10
Deciphering rapid geomagnetic field intensity variations in Western Europe from new French archeointensity data
The number of reliable archeointensity determinations obtained from Western Europe is still limited. As a consequence, the occurrence of several features in intensity behavior suggested in the literature is under debate. In particular, the scatter between different datasets has led some authors to conclude that the uncertainties attached to the intensity results are in fact too large to allow the safe identification of rapid intensity fluctuations, such as those previously proposed in France over the past millennium. In this debate, which addresses both the reality of rapid regional intensity variations and the reliability of the archeointensity determinations, the past millennium in Western Europe appears as an ideal target. Numerous sampling opportunities make it possible to select very well-dated groups of fragments. We will present new archeointensity results from France precisely dated, sometimes to within a few years, between the 12th and the 19th centuries. This data were obtained using two different experimental protocols derived from the Thellier and Thellier method (IZZI procedure and high-temperature magnetization measurements) with both TRM anisotropy and cooling rate effects taken into account, making possible a direct comparison between the two related datasets. Special care was also paid to the coherence between archeointensity values of the same age. We will show that the tests above provide strong support for the reliability of our archeointensity determinations. They confirm that the geomagnetic field intensity in Western Europe significantly and continuously varied over the past 1000 years.
GP43C-11
Geomagnetic Field Models for 0-3kA Derived Using Various Data Types and Selection Criteria
To understand the content and the causes of the changes in the Earth's magnetic field before the era of direct field observations one has to rely on archeomagnetic and lake sediment paleomagnetic data. Recently, we have combined such records into an updated compilation, and used them to derive a series of time- varying geomagnetic field models covering the past 3000 years. Distinct models were obtained by regularized inversion of various selections of (i) data from archeological artifacts, (ii) lake sediment data, or (iii) both types. Additionally, we created models (iv) from selected subsets with the most reliable data, or (v) by rejecting outliers from the datasets. Initially we built distinct models from each of the five data sets. Then outliers from each model were rejected, and the remaining data were used to derive a second version of each model. We will show the features of these models and explain which model best suits a particular application. The archeological models appear to represent the field with better resolution in the Northern hemisphere, particularly in Europe and Asia. The other models are smoother, but yield a better global representation of the field. We will also show that the outlier rejection technique seems to be as effective as the data selection technique, without requiring so many data to be rejected.
GP43C-12
Remanence Acquisition in Marine Carbonates: a Lesson from the K-T Boundary Interval
An apparently complete carbonate-rich Cretaceous-Tertiary boundary interval in ODP section 119-738C- 20R-5 from the southern Kerguelen Plateau provides a unique insight into processes of magnetization acquisition in marine carbonates. The boundary interval is characterized by a 1-m-thick clay-rich zone. The basal 15 cm of this zone is finely laminated, the upper part is bioturbated. It has been inferred that the clay- rich zone formed over a long time interval, and the bulk of the clay in this zone has a local provenance. Although the elevated Ir concentration and the evolutionary change in the nannofossil assemblage are spread over the laminated interval, there is no recognizable change in the composition of the clay mineral assemblage between the laminated and bioturbated zones. No faunal, mineralogical, or chemical evidence for anoxic/sulfate-reducing conditions within the clay-rich zone was found. The total iron content of the clay-rich zone co-varies with the alumosilicate content, indicating detrital source for iron. Normalized by the alumosilicate content, the laminated and bioturbated intervals have comparable total iron values, yet strikingly different magnetic properties. The initial susceptibility and NRM intensities are approximately an order of magnitude higher in the bioturbated interval compared to the laminated one. Our detailed rock magnetic study indicates that PSD magnetite grains likely of biogenic origin are the dominant iron-bearing phase in the bioturbated interval. In the laminated interval, apart from a small ferromagnetic fraction with MD-like behavior, non-silicate-bound iron is mainly sequestered in paramagnetic phases, probably (poorly crystalline) oxyhydroxides. It appears that a shut-down of biological productivity after the K-T event allowed preservation of the initial detrital/early authigenic iron phases that are dominated by reactive iron oxyhydroxides. With the recovery of normal biological activity as evidenced by the resumption of bioturbation, the oxyhydroxides had been transformed into biogenic magnetite. This transformation led to a several-fold increase in the NRM intensity of the rocks. Our results suggest that in deep marine environments distant from clastic sources, the depositional remanent magnetization plays a subordinate role to the biogeochemical magnetization. Potential delay in magnetization acquisition in such carbonate rocks should be considered in high-resolution paleomagnetic studies.