pTRM Checks Research Articles

Using Preisach Theory to Evaluate Chemical Remanent Magnetization and Its Behavior During Thellier‐Thellier‐Coe Paleointensity Experiments

AbstractThe behavior of grain‐growth chemical remanent magnetizations (gCRM) are investigated for different coercivity and magnetostatic‐interaction‐field distributions and acquisition conditions using a thermally activated Preisach model for assemblages of interacting single‐domain grains. A new growth‐rate dependent equation was derived, from which it was found that gCRM intensity is over 10% more sensitive to growth rate than previously modeled. We compare the gCRM results to the behavior of thermoremanences (TRM). gCRMs are two times more sensitive to changes in coercivity distribution, whereas TRMs are four times more sensitive to changes in magnetostatic interactions. The Thellier‐Thellier‐Coe paleointensity protocol was simulated in Preisach space, and gCRMs were found to produce concave‐up Arai plots with pTRM checks which plot to the left of the Arai plot and positive partial‐TRM tail checks that increase with magnetostatic interactions. This often leads to the failure of selection criteria, but high‐temperature segments can pass the criteria for weakly interacting gCRMs; these estimates can underestimate the field by up to 66%.

Do changes in geomagnetic secular variation, dipole moment and polarity reversal frequency correlate over the past 155 Myr?

SUMMARY Changes in palaeosecular variation, dipole moment and polarity reversal frequency are salient features of the Earth’s magnetic field over the geological past, yet how these changes are linked by the geodynamo remains controversial. To further understand this issue, we provide new absolute (API) and relative (RPI) palaeointensities from the ∼1-km-thick basaltic sequence of Waja (North Ethiopia) emplaced around 31 Ma, yielding an instantaneous virtual dipole moment of 57 ± 9 ZAm2 (1σ, N = 18) and a relative variability in intensity εF = 0.39 ± 0.07 (1σ, N = 19). Our analysis of the API database with strict selection criteria (inclusion of Thellier-style determinations with pTRM checks only, at least five determinations per cooling unit, and within-unit relative standard error lower than 10 per cent) fails to identify any robust correlation between changes in dipole moment and reversal frequency over the past 155 Myr. More convincingly, the available RPI results are consistent with an increase of the palaeosecular-variation proxy εF with reversal rate, as predicted by numerical dynamo simulations. We also find that the API-based estimate εF = 0.40 ± 0.03 (1σ, N = 104), computed from the filtered version of the World Palaeointensity Database for the 0.77–31 Ma interval, is consistent with the scaling rule, suggesting that the API record has been sufficiently sampled over the past 31 Ma. We thus speculate that the absence of negative correlation between changes in dipole moment and reversal frequency in the API database over the past 155 Myr may be the result of insufficient sampling prior to 31 Ma rather than the signature of an intrinsic geomagnetic feature.

RESET: A Method to Monitor Thermoremanent Alteration in Thellier‐Series Paleointensity Experiments

AbstractAccurately estimating the paleomagnetic field intensity recorded in terrestrial and planetary materials is the key to understanding dynamo processes. Thellier‐series stepwise‐heating methods with partial thermoremanent magnetization (pTRM) checks for alteration are considered the most reliable technique even though pTRM checks are shown to be incapable of detecting the entirety of thermal alteration. We utilize a recently developed multidomain‐correction experiment (or RESET) to monitor thermally induced magnetic property alteration that may happen to the very specimen used in Thellier‐series experiments. We also use rock magnetic property changes to track the physicochemical alteration of fresh companion specimens of Galapagos lavas that were previously used for paleointensity determinations. Our results show that laboratory heating induced thermal alteration of a typical Galapagos lava sample starts to occur at around 500°C. It escaped detection by pTRM checks but was caught by our RESET method.

Paleomagnetic study of basaltic rocks from Baengnyeong Island, Korea: efficiency of the Tsunakawa\u2013Shaw paleointensity determination on non-SD-bearing materials and implication for the early Pliocene geomagnetic field intensity

Finding the statistical intensity signatures of the Earth’s magnetic field over geologic time has helped understanding of the evolution of the Earth’s interior and its interactions with other integral parts of Earth systems. However, this has been often hampered by a paucity of absolute paleointensity (API) data, which are difficult to obtain primarily because of non-ideal magnetic behaviors of natural materials. Here, we present new API determination data with paleodirectional and rock magnetic analyses from basaltic rocks probably aged ~ 4‒5 Ma in Baengnyeong Island, Korea. Paleodirectional analysis obtained an overall mean direction of D = 347.3° and I = 38.3° (α95 = 4.9°, k = 113.4) corresponding to a virtual geomagnetic pole at 342.1° E and 70.2° N. Comprehensive rock magnetic analyses identified Ti-poor titanomagnetite with, in part, multi-domain (MD) particles as a main carrier of remanent magnetization. The Tsunakawa–Shaw (TS) method yielded 12 qualified API estimates with a high success rate, efficiently removing possible MD influences, and resulted in a mean value of 13.1 μT with good precision (1.7 μT, standard deviation). The Thellier method of the IZZI protocol with pTRM checks, coupled with the use of a bootstrap approach instead of the “conventional best-fitting” in API determination, gave 6.6‒19.7 μT as a 95% confidence interval of its mean API estimate, which supports the reliability of our TS-derived API mean estimate; but it is not considered in the final mean value because of the relatively large uncertainty. The virtual dipole moment corresponding to the TS-derived API mean, 2.9 (± 0.4) × 1022 Am2, is somewhat lower than the expectations of the past few Myr averages. Combined with a global API database, our new data implies a larger dispersion in the dipole moment during the early Pliocene than previously inferred. This also suggests that the issue of whether the early Pliocene average dipole strength was moderately high (> 5 × 1022 Am2) or consistent (4‒5 × 1022 Am2) should be discussed further.

Effects of titanomagnetite reordering processes on thermal demagnetization and paleointensity experiments

AbstractTitanomagnetite (Fe3‐xTixO4, 0 ≤ x ≤ 1) is a common, naturally occurring magnetic mineral critical to many paleomagnetic studies. Underlying most interpretations is the assumption that, lacking chemical alteration, Curie temperature (Tc) remains constant. However, recent work has demonstrated that Tc of many natural titanomagnetites varies strongly as a function of thermal history, independent of chemical alteration. This is inferred to arise from reordering of cations and/or vacancies in the crystal structure, and changes occur at temperatures and times relevant to standard paleomagnetic thermal treatments. Because changes take place at T < Tc, they have the potential to dramatically affect thermal remanence acquisition or demagnetization, impacting interpretation of paleomagnetic results. Here we have modeled the effects of reordering on standard thermal demagnetization and paleointensity experiments. Results suggest that Tc changes during laboratory heating make it impossible to accurately measure the unblocking temperature spectrum without modifying it. Samples with a starting Tc0 less than the closure temperature (Tclose) for the reordering process will develop a high‐temperature “tail” that did not exist prior to heating. Samples with a starting Tc0 > Tclose will have their original Tb spectrum truncated at T ≈ Tclose. Predicted behavior during Thellier‐type paleointensity experiments results in only modest deviations in NRM‐lost or pTRM*‐gained from the nonreordering case. Much larger deviations are predicted for pTRM checks. Compared to paleointensity results from titanomagnetite‐bearing pyroclastic deposits, modeled nonideal behavior occurs in the same temperature intervals, but is much more systematic. Reordering is likely one contributing factor to failure of paleointensity experiments.

Microwave paleointensities indicate a low paleomagnetic dipole moment at the Permo-Triassic boundary

The quantity of igneous material comprising the Siberian Traps provides a uniquely excellent opportunity to constrain Earth’s paleomagnetic field intensity at the Permo-Triassic boundary. There remains however, a contradiction about the strength of the magnetic field that is exacerbated by the limited number of measurement data. To clarify the geomagnetic field behavior during this time period, for the first time, a microwave paleointensity study has been carried out on the Permo-Triassic flood basalts in order to complement existing datasets obtained using conventional thermal techniques. Samples, which have been dated at ∼250Ma, of the Permo-Triassic trap basalts from the northern extrusive (Maymecha-Kotuy region) and the southeastern intrusive (areas of the Sytikanskaya and Yubileinaya kimberlite pipes) localities on the Siberian platform are investigated. These units have already demonstrated reliable paleomagnetic directions consistent with the retention of a primary remanence. Furthermore, Scanning Electron Microscope analysis confirms the presence of iron oxides likely of primary origin. Microwave Thellier-type paleointensity experiments (IZZI protocol with partial thermoremanent magnetization checks) are performed on 50 samples from 11 sites, of which, 28 samples from 7 sites provide satisfactory paleointensity data. The samples display corresponding distinct directional components, positive pTRM checks and little or no zig-zagging of the Arai or Zijderveld plot, providing evidence to support that the samples are not influenced by lab-induced alteration or multi-domain behavior. The accepted microwave paleointensity results from this study are combined with thermal Thellier-type results from previously published studies to obtain overall estimates for different regions of the Siberian Traps. The mean geomagnetic field intensity obtained from the samples of the northern part is 13.4±12.7μT (Maymecha-Kotuy region), whereas from the southeastern part is 17.3±16.5μT (Sytikanskaya kimberlite pipe) and 48.5±7.3μT (Yubileinaya kimberlite pipe), suggesting that the regional discrepancy is probably due to the insufficient sampling of geomagnetic secular variation, and thus, multiple localities need to be considered to obtain an accurate paleomagnetic dipole moment for this time period. It demonstrates that the overall mean paleointensity of the Siberian Traps is 19.5±13.0μT which corresponds to a mean virtual dipole moment of 3.2±1.8×1022Am2. Results indicate that the average magnetic field intensity during Permo-Triassic boundary is significantly lower (by approximately 50%) than the present geomagnetic field intensity, and thus, it implies that the Mesozoic dipole low might extend 50Myr further back in time than previously recognized.

Archeomagnetism of Jesuit Missions in South Brazil (1657–1706 AD) and assessment of the South American database

South America contributes only a small fraction of the global archeomagnetic data. Recent work in the region has expanded significantly the previous database with new data being generated from Chile, Argentina, Northeast and Southeast Brazil. We report here new results from Jesuit Missions in South Brazil, at the triple border with Argentina and Paraguay. Our archeological collection comprises a total of 24 fragments of baked clay construction materials from three Jesuit Missions, São Luiz Gonzaga 1657–1687 AD (3 fragments), São João Batista 1667–1697 AD (4 fragments) and Santo Ângelo 1676–1706 AD (17 fragments). Archeointensity determinations were performed with the double-heating technique in its modified form, with pTRM checks and pTRM tail checks. Measurements were complemented by anisotropy and cooling-rate corrections. A total of 24 specimens (11 fragments) passed strict quality selection, corresponding to a success rate of 45%. We also performed an experimental test for the 6-specimen average anisotropy correction technique and we show that it does not correct for the effects of TRM anisotropy. Results were similar within error for the three missions: São Luiz Gonzaga (40.2±2.4 μT), São João Batista (39.1±1.6 μT) and Santo Ângelo (41.1±2.0 μT). These data were then compared with the most reliable data from South America, after a critical assessment of the current database. According to our analysis, only 39 intensity data for the continent can be considered as high-quality, most within the last 700 years; only three data were retained for older periods (800–1100 AD). The filtered data match reasonably well the available models for the past five centuries. A combined curve for South and Southeast Brazil plus Argentina plot systematically below relocated data from NE Brazil and Chile. These differences are likely due to complexities in the geometry of the field in South America not appropriately accounted for by a simple axial dipole. Our analysis highlights the need for high-quality data for the continent.

The effects of secondary mineral formation on Coe-type paleointensity determinations: Theory and simulation

Thellier-type experiments are the most widely applied approaches for determining the absolute paleointensities of Earth's magnetic field. One major problem, however, is that specimens are prone to thermal alteration due to the intensive thermal treatment during experiments. Linear Arai plots with acceptable partial thermal remanent magnetization (pTRM) checks have been considered as evidence for the absence of or negligible effects of thermal alteration and as reliable indicators of high-quality paleointensity estimates. However, by simulating the Coe variant of the Thellier method on assemblages of single domain (SD) magnetite particles, it is demonstrated that new magnetic minerals, which form during thermal treatments, can result in linear, concave-up, or concave-down Arai plots depending on the magnetic properties of both the primary and secondary magnetic phases. Among this range of behavior, pseudoideal Arai plots, which are linear with acceptable pTRM check statistics, would lead to paleointensity underestimates. It is further demonstrated that pTRM checks are proportional to the degree of underestimate with a magnetic granulometry dependency for SD particles. Due to the complexity of this dependency, pTRM check statistics are only comparable when specimens have similar magnetic properties. This suggests that a universal threshold for pTRM check statistics is not likely to be effective. Since the criteria of linearity and low pTRM check statistics are insufficient to guarantee the fidelity of the estimates auxiliary rock magnetic methods such as temperature-dependent hysteresis parameters and anhysteretic remanent magnetization are highly recommended to identify the presence of alteration.

Palaeointensity and palaeomagnetic study of Cretaceous and Palaeocene rocks from Western Antarctica

SUMMARY A combined palaeodirectional and palaeointensity study of a representative collection of plutonic rocks from the Antarctic Peninsula batholith from the western part of the Antarctic Peninsula, near the Ukrainian Antarctic base 'Academik Vernadsky' were carried out. Petrographically, the collection includes gabbros, diorites and quartz diorites, tonalities, granodiorites and granites. The ages of igneous complex emplacement vary from 50 to 117 Ma with most of the rocks belonging to the Cretaceous Normal Superchron. The characteristic remanent magnetizations were isolated by stepwise thermal demagnetization over the temperature interval 440–590°C and their intensities amount to 95 per cent of the NRM. The geographic positions of palaeopoles do not contradict the ‘key poles’ of the Antarctic Peninsula between 90 and 60 Ma. A significant part of the collection was subjected to Coe-modified Thellier palaeointensity experiments with the pTRM checks, which yielded seven reliable palaeointensity determinations for seven different locations. The obtained VDMs are relatively low for all sites, being on average about half of the present day VDM. The analysis of available palaeointensity data for the Cretaceous, Miocene and Middle Jurassic indicates the existence of strong correlations between the mean VDM and VDM scatter versus the rate of reversals. However, due to the shortage of data, the correlations are not significant at the 5 per cent significance level.

Archaeomagnetic results on three Early Iron Age salt-kilns from Moyenvic (France)

Variations of the Earthfs magnetic field during the first millennium BC in western Europe remain poorly constrained, especially archaeointensity changes. Three salt-kilns (MOA, MOB and MOC) sampled in Moyenvic (Lorraine, eastern France) have been studied to provide new reference data. Each kiln has been dated by radiocarbon to originate from the Early Iron Age or Hallstatt period (between VIII and Vth Century BC). Rock magnetic experiments and hysteresis results suggest the predominance of pseudo-single domain (PSD) Ti-poor magnetite. Archaeomagnetic directions obtained by thermal and alternating field demagnetizations have high mean inclination (close to 70.) and declination (between 19 and 31.). A first set of classical Thellier.Thellier experiments was conducted on 46 samples with a laboratory field almost parallel to the direction of the characteristic remanent magnetization (ChRM). Only 24 of these specimens present a linear NRM.TRM plot. For other specimens, NRM.TRM plots are concave-up with positive pTRM checks. The very large dispersion observed between the determined palaeointensity values suggests some artefacts have not been fully recognized. A second set of Thellier experiments was conducted on 34 sister specimens with the laboratory field applied quasi-perpendicular to the ChRM. In these cases,mineralogical evolutions during heating and chemical remanent magnetization acquisitions have been clearly recognized, despite positive pTRM checks. The concave-up shapes of NRM.TRM plots appear mainly due to mineralogical alteration rather than to the presence of PSD.MD grains. For the entire set of samples the success rate of the palaeointensity determinations is very low with 80 per cent of the samples rejected. Nevertheless, reliable mean archaeointensities have been obtained for two of the three kilns (MOA, 80.1 } 14.5 ƒET and MOB, 86.6 } 6.9 ƒET at the latitude of Paris). The high field strength and the archaeomagnetic directions determined, combined with previous published data, provide further evidence for important changes of the Earth magnetic field in Europe during the first half of the first millennium BC. These large variations of the geomagnetic field during the Iron Ages indicate that archaeomagnetism is highly suitable for dating of structures from this period.

Geomagnetic field intensity from Kilauea 1955 and 1960 lava flows: Towards a better understanding of paleointensity

The record of the Earth’s magnetic field intensity during the past (paleointensity) carries important information about the geodynamo and the state of the Earth’s interior that is not contained in the record of its paleodirection. To determine what the critical factors in obtaining reliable estimate of paleointensity are, we present new results of a paleointensity study of the 1955 and 1960 Kilauea volcano lava flows, from the Big Island of Hawaii. Rock magnetic measurements on representative hand samples from each flow in conjunction with reflected light microscopy observations show the primary carriers of remanence to be pseudo-single domain titanomagnetite with various titanium contents. Paleointensity samples (small fragments previously embedded into salt pellets) were subjected to the Thellier-Coe experimental procedure. Fourteen temperature steps were distributed over the entire temperature range used (ambient temperature to 570°C). Control heating steps (commonly referred to as partial thermo-remanent magnetization — pTRM checks) were also conducted each third double heating step. Mean field intensity value (36.6 ± 0.7 μT) retrieved from 3 reliable site mean determinations reproduces the expected value within 1.1%. With the knowledge of the rock magnetic characteristics of the samples and the strength of the geomagnetic field during cooling of the lava, our investigation suggests that the Hawaiian lavas can faithfully record the local geomagnetic field and confirms that the Thellier-Coe type techniques are suitable on historical lava flows to yield reliable absolute paleointensity determinations. The variations in direct field measurements and in lab paleofield determinations may reflect local heterogeneities of the lava or influence of very local field anomalies due to the volcanic underlying terrain. These results underscore the importance that a better understanding of intensity results of historical lava flows is still required if reliable paleointensity determinations of older periods, for which we do not know the answer, are sought.

Testing determinations of absolute paleointensity from the 1955 and 1960 Hawaiian flows

A new exhaustive study of the flows erupted in 1955 and 1960 from the Kilauea volcano (Big Island of Hawaii) has been conducted in order to constrain the dispersion inherent to the data of absolute paleointensity reported so far and ultimately to determine what are the most critical parameters to obtain suitable determinations. The mean directions derived from both flows agree with the expected field within less than 2°. Paleointensity experiments have been performed using a double heating Thellier procedure completed by systematic pTRM checks. Almost all samples display a linear NRM–TRM straight line but only 20% of the samples have provided suitable results less than 10% away from the expected field. In the other cases, the deviation from the expected paleointensity is related to the quality of the pTRM checks, which thus provides a threshold value for acceptance of pTRM checks. All successful determinations less than 10% away from the expected value have a maximum DRAT ( Selkin and Tauxe, 2000) that never exceeds 5% while the mean value of all deviations is lower than 3.5%. Another interesting feature is the existence of a link between the unblocking temperature spectra and the accuracy of the paleofield, which points out the origin of difficulties in the presence of a wide distribution of grain sizes and/or the existence of several magnetic phases. The results also indicate that magnetite with a narrow range of high unblocking temperatures and little variability in grain size is a favorable situation to obtain suitable determinations with a high success rate. These observations may well explain the dispersion inherent to the previous studies of the Hawaiian flows.

Geomagnetic paleointensity between 1300 and 1750 A.D. derived from a bread oven floor sequence in Lübeck, Germany

Geomagnetic paleointensities have been determined from a single archaeological site in Lübeck, Germany, where a sequence of 25 bread oven floors has been preserved in a bakery from medieval times until today. Age dating confines the time interval from about 1300 A.D. to about 1750 A.D. Paleomagnetic directions have been published from each oven floor and are updated here. The specimens have very stable directions and no or only weak secondary components. The oven floor material was characterized rock magnetically using Thellier viscosity indices, median destructive field values, Curie point determinations, and hysteresis measurements. Magnetic carriers are mixtures of SD, PSD, and minor MD magnetite and/or maghemite together with small amounts of hematite. Paleointensity was measured from selected specimens with the double‐heating Thellier method including pTRM checks and determination of TRM anisotropy tensors. Corrections for anisotropy as well as for cooling rate turned out to be unnecessary. Ninety‐two percent of the Thellier experiments passed the assigned acceptance criteria and provided four to six reliable paleointensity estimates per oven floor. Mean paleointensity values derived from 22 oven floors show maxima in the 15th and early 17th centuries A.D., followed by a decrease of paleointensity of about 20% until 1750 A.D. Together with the directions the record represents about 450 years of full vector secular variation. The results compare well with historical models of the Earth's magnetic field as well as with a selected high‐quality paleointensity data set for western and central Europe.

A comparison of a quasi-perpendicular method of absolute palaeointensity determination with other thermal and microwave techniques

We describe, test, and explain the relative effectiveness of six thermal and microwave techniques of absolute palaeointensity determination applied to lava samples from Mount Etna (Sicily) that acquired their thermoremanence (TRM) in a known geomagnetic field. The application of a single alternating field demagnetisation treatment prior to each measurement in a standard Thellier experiment is demonstrated to be beneficial: increasing the quality of the results and making the analysis simpler. A new quasi-perpendicular (QP) method dramatically improves the linearity of Arai plots by reducing or altogether eliminating non-ideal effects due to multidomain (MD) grains. Two broad types of sample were encountered by this study and the QP method is the only one capable of producing accurate measurements from them both. Its ease of use, speed, and increased reliability should allow it to replace standard double heating methods in studies such as these where laboratory alteration can be detected by other means than pTRM checks. The results of experiments performed using a microwave palaeointensity system provide further proof that the microwave (de)magnetisation process is equivalent to conventional thermal excitation. They also demonstrate that microwave experiments are subject to the same risks as conventional experiments when performed on samples containing MD grains. The behaviour of the samples in both the thermal and microwave experiments is in good agreement with a recently proposed model of MD TRM which is based on first-order symmetry of the remagnetisation process.

First-order symmetry of weak-field partial thermoremanence in multi-domain (MD) ferromagnetic grains: 2. Implications for Thellier-type palaeointensity determination

Two models, derived in very different ways but both incorporating the symmetry constraints outlined in the previous paper, are developed for use in simulating Thellier-type palaeointensity experiments. These are then tested against experimental data which endorse them both as powerful tools. They predict that non-ideal behaviour due to MD grain content is sensitive not only to the protocol employed and the vector difference between the partial thermoremanent magnetisations (pTRMs) imparted in the laboratory and the natural remanent magnetisation (NRM), but also to the spacing of the temperature steps and the type of pTRM imparted during the remagnetisation stages. Consistent with experimental studies, the models predict that pTRM checks may fail despite no alteration occurring and that any double-stage experimental protocol will produce non-ideal results to some degree provided that MD grains are present. A single-heating (perpendicular or high-angle single heating) protocol is predicted to remove non-ideal effects provided that the grains obey the symmetry constraints perfectly. However, this remains to be experimentally tested and does not allow for the use of pTRM checks. Five straightforward recommendations for future palaeointensity studies to follow are outlined. These are designed to minimise non-ideal behaviour due to MD grains as far as possible while retaining the capacity to detect and reject those samples which will nonetheless produce significant errors.

Reliability of geomagnetic paleointensity data: the effects of the NRM fraction and concave-up behavior on paleointensity determinations by the Thellier method

To test the reliability of the Thellier method for paleointensity determinations, we studied six historic lavas from Hawaii and two Gauss-age lava flows from Raiatea Island (French Polynesia). Our aim is to investigate the effects of the NRM fraction and concave-up behavior of NRM–thermal remanent magnetization (TRM) diagrams on paleointensity determinations. For the Hawaiian samples, the paleointensity results were investigated at both sample and site levels. For consistency and confidence in the paleointensity results, it is important to measure multiple samples from each cooling unit. The results from the Raiatea Island samples confirm that reliable paleointensities can be obtained from NRM–TRM diagrams with concave-up curvature, provided the data are accompanied by successful partial TRM (pTRM) checks and no significant chemical remanent magnetization (CRM) production. We conclude that reliable determinations of the paleofield strength require analyses of linear segments representing at least 40–50% of the total NRM. This new criterion has to be considered for future studies and for evaluating published paleointensities for calculating average geomagnetic field models. Using this condition together with other commonly employed selection criteria, the observed mean site paleointensities are typically within 10% of the Definitive Geomagnetic Reference Field (DGRF). Our new results for the Hawaii 1960 lava flow are in excellent agreement with the expected value, in contrast to significant discrepancies observed in some earlier studies. Overestimates of paleointensity determinations can arise from cooling-rate dependence of TRM acquisition, viscous remanent magnetization (VRM) at elevated temperatures, and TRM properties of multidomain (MD) particles. These outcomes are exaggerated at lower temperature ranges. Therefore, we suggest that, provided the pTRM checks are successful and there is no significant CRM production, it is better to increase the NRM fraction used in paleointensity analyses rather than to maximize correlation coefficients of line segments on the NRM–TRM diagrams. We introduce the factor, Q = N〈 q〉, to assess the quality of the weighted mean paleointensity, H w, for each cooling unit.

Linear and nonlinear Thellier paleointensity behavior of natural minerals

We tested single‐domain (SD) and multidomain (MD) theories of the acquisition and thermal demagnetization of thermoremanent magnetization (TRM) and partial TRM (pTRM) in Thellier paleointensity determination, using as samples plagioclase and mafic minerals extracted from Matachewan diabase dikes. Susceptibility, hysteresis, and coercivity data indicate that the plagioclase contains mainly SD grains of magnetite while the mafic minerals (clinopyroxene, biotite, and actinolite) contain mainly MD magnetite. The extracted minerals were used in simulated Thellier experiments with laboratory TRM as a proxy for natural remanent magnetization (NRM). The plagioclase behaved ideally, as predicted by Néel SD theory, giving a linear NRM‐pTRM (Arai) plot with positive pTRM checks. The mafic minerals behaved nonideally, with strongly curved Arai plots sagging below the ideal SD line, in accord with observations for MD magnetites and with predictions that in heating to moderate temperatures T, more TRM is lost than can be regained as pTRM, causing convex‐down Arai plots. In experiments with pTRM parallel to NRM, all pTRM tail checks were zero, but pTRM checks increasingly deviated from the Arai curve toward the SD line as T increased. However, in repeat Thellier experiments, pTRMs at all T agreed with the original pTRMs. The negative pTRM checks were therefore not due to physicochemical alteration but rather to departure from initial states with different magnetic histories than those of the original pTRMs. The inability of pTRM checks to reliably distinguish between MD behavior and irreversible alteration could lead to overestimates of paleointensity from straight‐line fits to the low‐T “unaltered” part of the Arai curve or underestimates from fits to high‐T data. Alternating field cleaning to 15 mT before each Thellier step made the Arai plot only marginally less curved. Low‐temperature demagnetization (LTD, zero‐field cycling through the 120 K Verwey transition) hardly changed the curvature of the Arai plot for pTRM parallel to NRM. For pTRM perpendicular to NRM, LTD led to an inflected Arai curve straddling the ideal SD line. Data of either form would not give reliable paleointensity values. No presently available pretreatment renders the paleointensity data of MD grains usable. SD and nearly SD grains which behave ideally in the Thellier experiment remain the only useful recorders of absolute paleointensity.

Thellier paleointensity theory and experiments for multidomain grains

We extend theories of thermoremanent magnetization (TRM) and partial TRM (pTRM) in multidomain (MD) grains to model thermal demagnetization and pTRM acquisition steps in Thellier paleointensity determination. Because of the interleaving of zero‐field and in‐field heating‐cooling steps to increasing temperatures the initial state for any step is complex, and theoretical modeling is more intricate than for pTRM production or thermal demagnetization separately. At low to moderate temperature T, TRM lost exceeds pTRM regained, causing convex down Arai plots sagging below the ideal single‐domain (SD) line. At moderate to high T, pTRM acquisition outweighs TRM loss. As T approaches TCurie, pTRM gain exactly equals TRM loss, and the Arai plot becomes ideal. When pTRMs are produced perpendicular to the original TRM and measured directly rather than by differencing field‐on and field‐off results, there is less deviation from ideality at low to moderate T. Our theory agrees semiquantitatively with results for parallel and perpendicular pTRMs for large MD magnetites (135 μm). Smaller MD magnetites (6 and 20 μm) have less curved Arai plots, and the smallest magnetites (0.6 and 1 μm) have almost linear plots. Positive pTRM checks demonstrate that curved Arai plots of MD grains are reproducible, unlike curved plots for rocks that alter physicochemically in the Thellier experiment, while negative pTRM tail checks indicate undemagnetized pTRM residuals. Low‐temperature demagnetization improves linearity only slightly. Practical applications of this work include using the predicted threshold T below which no net pTRM is produced in a parallel Thellier experiment to screen data used for paleointensity fits. Straight line fits through low‐ and medium‐T points in Arai plots of MD (135 μm) grains overestimated the paleofield by as much as 100% and for small pseudo‐single‐domain (PSD) (0.6 and 1 μm) grains overestimated by about 25%. However, by using linear segments of medium‐ to high‐T data with f values ≥0.5 it may be possible to obtain reasonable paleointensity estimates even for larger PSD (6 and 20 μm) and MD grains. Middle‐ to high‐T fits for 0.6 μm grains gave paleointensities within 4% of the correct value, utilizing essentially the entire data set (f > 0.9). Perpendicular data always gave superior linear fits. Orienting samples with their natural remanent magnetizations perpendicular to the laboratory field is therefore recommended for rocks containing PSD and MD grains. However, double heatings are preferable to single heatings because they allow pTRM tail checks to be carried out.

The application of acceptance criteria to results of Thellier palaeointensity experiments performed on samples with pseudo-single-domain-like characteristics

The results of a simulated Thellier palaeointensity experiment show that the low temperature portion of Arai plots produced by samples with pseudo-single-domain (PSD)-like hysteresis properties can produce significant inaccuracies despite passing conventional acceptance criteria. Furthermore, full Arai plots which fail these criteria can still produce reasonably accurate results. The evidence presented suggests that during a palaeointensity experiment: pTRM checks can fail despite no alteration occurring, pTRM tail checks are not reliable as indicators of MD-like behaviour, room temperature susceptibility monitoring is unreliable for detecting alteration, and that the minimum requirement for the f (fraction)-value should be increased from 0.15 to at least 0.5.

Analysis of long-term variations in the geomagnetic poloidal field intensity and evaluation of their relationship with global geodynamics

SUMMARY The 1167 published cooling unit (CU) palaeointensity estimates contained in the 400‐10 Ma portion of the PINT global database were rigorously filtered according to accurate age determinations, palaeodirectional reliability, recognition of polarity and the method of palaeointensity acquisition. The remaining 865 estimates (group 1) were further filtered to ensure self-consistency, reducing the data set to 425 estimates (group 2). Group 1 and 2 data were clustered into temporally and/or spatially distinct rock suites (RS) enabling each part of the record to be assessed for potential biasing by overrepresentation of palaeosecular variation (PSV). The record was segmented according to the distribution of the data, rather than using arbitrary time windows, to ensure quasi-consistent behaviour within each segment. Differences between these segments clearly indicate that a significant long-timescale (10 7 and 10 8 yr) variation of the mean geomagnetic poloidal field intensity (GPFI) occurred during the 400‐10 Ma period and hence that changing lowermost mantle conditions affect the capacity of the geodynamo to generate a poloidal field. Both the mean dipole moment and its standard deviation appear to be a function of the range of values each CU may adopt at one particular time. This range is itself controlled by the variation of the maximum limit of dipole moment, while the value of the minimum limit remains relatively constant. Tentative support is provided for the recent suggestion that PSV may have been reduced during the Cretaceous normal superchron (CNS), though more data are needed in the range 120‐60 Ma to confirm this. No conclusive evidence was found to support the suggestion that the GPFI record may be biased towards low or high values by palaeointensity determinations obtained using methods that do not adopt pTRM checks. Indeed, offsets caused by unreliable data in well-represented parts of the record are likely to be random and cancel one another out. When GPFI variation is analysed at a sufficiently high resolution to allow comparisons with the geomagnetic polarity reversal frequency (RF), it is not possible to confirm whether the two parameters are anticorrelated, decoupled or related in some more complex way. However, it is clear that GPFI and RF are definitely not positively correlated as has been previously suggested. The present database documents sharp increases in GPFI around the onset times of the two recognized superchrons, itself implying an anticorrelation. The implications, for geodynamo and mantle modelling, of both an anticorrelation and a decoupling of the geomagnetic parameters are discussed briefly. A generic geodynamic model is proposed to explain the relationship between observed longterm changes in GPFI and global geodynamic processes. The model predicts that changes in GPFI result from a chain of geodynamic processes extending from crust to core, beginning with plate reorganizations at the surface and culminating in increases in the vigour of outer core convection. Supercontinents are transient surface expressions of such geodynamic processes and provide the potential to test the generic model. Four time stages are proposed to describe the major long-term changes in GPFI since the Early Devonian: 400‐350, 350‐250, 250‐175