Three archaeomagnetic applications of archaeological interest to the study of burnt anthropogenic cave sediments

1 Three archaeomagnetic applications of archaeological interest to the study of burnt anthropogenic cave sediments Á. Carrancho1*, Á. Herrejón2 and J.M. Vergès3, 4 1. Área de Prehistoria. Dpto. de Ciencias Históricas y Geografía. Universidad de Burgos. Edificio I+D+I. Plaza Misael Bañuelos s/n. 09001, Burgos, Spain. E-mail: acarranc[email protected] 2. Dpto. Física, Universidad de Burgos. Escuela Politécnica Superior, Avda. Cantabria S/N 09006, Burgos, Spain. 3. IPHES, Institut Català de Paleoecologia Humana i Evolució Social, C/Marcel.lí Domingo s/n, Campus Sescelades (Edifici W3), 43007 Tarragona, Spain 4. Área de Prehistoria, Universitat Rovira i Virgili (URV), Avinguda de Catalunya, 35, 43002 Tarragona, Spain Abstract Recent archaeomagnetic studies carried out on Midto Late Holocene burnt anthropogenic cave sediments have shown that under certain conditions, these materials are suitable geomagnetic field recorders. Archaeomagnetic analyses carried out on these contexts constitute a rich source of information not only for geophysical purposes -in terms of reconstructing the variation of Earth's magnetic field in the pastbut also from the archaeological point of view, for example by archaeomagnetic dating. Here, we report three different archaeomagnetic applications to the study of burnt cave sediments: (i) archaeomagnetic dating; (ii) determining palaeotemperatures and (iii) assessing postdepositional processes. The first case study is a dating attempt carried out on a Late Holocene (Bronze Age) burnt level from El Mirador Cave (Burgos, Spain). Using the directional European secular variation curve, several dating intervals were obtained for the last burning of this combustion feature. Considering the archaeological evidence and the independent radiometric (14C) dating available the possible ages obtained are discussed. This is the first archaeomagnetic dating obtained in these contexts so far. The second case study is an application of the method to determine the last heating temperatures reached by the carbonaceous facies of these fires. Stepwise thermal demagnetization of oriented samples can be used to quantitatively estimate heating temperatures. An intermediate normal polarity component interpreted as a partial *Manuscript Click here to view linked References 2 thermo-remanence (pTRM) with maximum unblocking temperatures of 400 – 450 ºC was systematically identified, revealing the last heating temperatures experienced by this facies. These temperatures were confirmed with partial thermomagnetic curve experiments. Finally, archaeomagnetic analyses on a partially bioturbated burning event were performed in order to evaluate until what spatial extent the burnt sediments were affected by post-depositional mechanical alteration processes. For each case study, the archaeological implications are discussed highlighting the potential of archaeomagnetic methods to retrieve archaeological information. Keywords: Fumiers, Holocene, Thermoremanent magnetization, Secular variation, Ashes, Bronze Age. 1. Introduction Since the pioneering work of Brochier (1983a,b), the study of Holocene burnt anthropogenic cave sediments has experienced considerable progress. A great number of archaeological excavations as well as the increasing amount of data provided by disciplines such as soil micromorphology (Angelucci et al 2009; Boschian 1997; Macphail et al. 1997), palaeobotany (Rasmussen 1993; Delhon et al. 2008; Cabanes et al. 2009) or zooarchaeology (Martín et al. 2014; Rowley-Conwy 1998) among others, is yielding valuable information about the formation and use of these deposits. Archaeomagnetism has emerged as one of these lines of research. Although it has a long tradition in Earth sciences its application in prehistoric archaeology is still sporadic and its potential to retrieve archaeological information remains underutilized. Broadly speaking, archaeomagnetism deals with the study of the record of the Earth´s magnetic field direction and/or intensity changes in the past in burnt archaeological materials. Most archaeological materials contain small amounts of ferromagnetic minerals (s.l.), such as magnetite or haematite. When heated to high temperatures (> 500 – 600 ºC) and subsequently cooled these minerals acquire a remanent (permanent) magnetization parallel to the ambient magnetic field. Under several conditions this information may be very stable over long periods of time and used in a wide variety of applications, among which dating is likely the most known. However, given their versatility, magnetic methods can provide valuable information ranging from determining palaeotemperatures (e.g., Brown et al. 2009), ash sourcing (Church et al. 2007) or assessing the degree of preservation in archaeological cave fires (e.g., 3 Carrancho et al. 2012). This paper provides a review of some of these applications specifically applied to anthropogenic cave sequences. These stratigraphic sequences usually contain multiple burning events generated by the periodic burning of organic material (e.g., vegetal remains and dung) produced by livestock penning (Angelucci et al. 2009). Their preservation state is usually good, are generally well-dated by independent methods (namely radiocarbon) and have a broad geographical distribution throughout the Mediterranean region (Angelucci et al. 2009). Therefore they constitute a great source of archaeomagnetic data and the information obtained has both geophysical and archaeological interest. The main goal of this article is to highlight the potential of magnetic methods to answer archaeological questions through three different applications. The first is a dating attempt of a firing event from El Mirador Cave (Spain) using the recently designed directional European Secular Variation (SV) curve for the Neolithic (Carrancho et al. 2013). The second is a methodological application to determine the last heating temperature undergone by these fires. The third consists on evaluating to what extent a burning event might be affected by post-depositional processes. The archaeological and archaeomagnetic implications of these cases studies will be discussed as well as the limits of each application. 2. Materials and methods 2.1 Sites The studied materials correspond to samples from Neolithic, Chalcolithic and Bronze Age burning events exposed in the Holocene stratigraphies of El Mirador and Portalón de Cueva Mayor caves (Sierra de Atapuerca, Burgos) and El Mirón Cave (Cantabria, Spain; Fig. 1a). For detailed information on the archaeology, stratigraphy and chronology of these sites the reader is referred to Straus and González Morales (2012), Carretero et al., (2008) and Vergès et al. (2008; this volume). These fires generally contain a grey/white ash facies of variable thickness (2-10 cm) over a thin (~ 2 cm) black carbonaceous subjacent facies. 2.2 Sampling Archaeomagnetic sampling was carried out with the aid of a non-ferromagnetic cylindrical tube which incorporates a built-in orientation system specifically designed 4 for soft (unlithified) lithologies (Carrancho et al. 2013). Its main advantage is that it allows a precise geographical orientation of the samples besides being minimally invasive. The tube is pressed against vertical profiles where the burnt facies outcrop. After the azimuthal reading, the sediment is carefully inserted in cylindrical plastic boxes (Ø 16.5 mm, 17 mm length; volume of about 3.6 cm3) and stored in cold conditions (3-4 ºC) until measurement to avoid chemical alterations. Samples for thermal (TH) demagnetization of the natural remanent magnetization (NRM) were oriented by the same means and introduced into home-made plaster cubes (Carrancho 2010). These contain a cylindrical hole with the same dimensions and volume as the plastic capsules in order to keep the sample in fixed position. The NRM of the plaster cubes is at least two orders of magnitude less than the sample´s magnetization. Details of the number and type of samples collected for each case study are given below. 2.2.1 Case study 1 (archaeomagnetic dating) A burning event (Ci1) from El Mirador Cave (42º 20´ 58´´ N, 03º 30´ 33´´ W; Sierra de Atapuerca, Burgos, Spain) was intensively sampled for archaeomagnetic dating purposes (Fig.1a-b). The archaeostratigraphic unit where Ci1 is located (MIR103 – Sector 100) has a 14C (AMS) dating (sample code: Beta – 339094) obtained from a charcoal fragment with a 2σ dating interval of 1510 to 1410 cal. BC (3190 +/- 30 BP). Archaeological evidence is limited to few pottery remains suggesting a possible Bronze Age for the MIR103 unit. The objective here was to obtain an archaeomagnetic date of the last heating of this event using the directional European SV curve (Carrancho et al. 2013). The Ci1 burning event is composed of an ash and a carbonaceous facies. The ashes are white on top and reddish brown on the bottom with a total thickness of about 15 cm. Just beneath, a dark carbonaceous (~ 2 cm) facies is preserved delimiting the surface where burning occurred (Fig. 1). At the top of the lower level, just at the base of the burning event, a burrow can be observed that may have partially affected the structure. A total of 29 oriented samples (22 ashes and 7 carbonaceous samples) were collected following the sampling procedure described in section 2.2. 2.2.2 Case study 2 (estimating palaeotemperatures) The samples analysed in this case study are representative carbonaceous samples from 6 different Holocene burning events from El Mirador, El Portalón and El Mirón Cave (Spain). They were previously studied along with hundreds of burnt samples in the 5 design of the first directional European PSV curve for the Neolithic (Carrancho et al. 2013). The objective is to show how the identification of partial thermal remanent magnetizations (pTRMs) permits the quantitative estimation of the last heating temperature in the carbonaceous facies. The validity of this approach was verified carrying out thermomagnetic curve analyses on bulk (unoriented) sample from this facies and studying their degree of reversibility (section 4.2). The sampling procedure was the same as described in section 2.2. 2.2.3 Case study 3 (assessing post-depositional processes) In order to test the reliability of the palaeomagnetic method to determine to what extent the mechanical reworking might have affected an anthropic cave fire, an archaeomagnetic study of a Late Holocene burning event from El Portalón Cave (Burgos, Spain; map of Fig. 1a) is reported. This burning event contains a white ash facies (~ 10 cm) over a ~ 2 cm dark carbonaceous facies both partially altered by an ancient burrow (Fig. 8). The colour and texture of ashes on the right side of the burning event are somewhat mixed, suggesting that some kind of mechanical reorganization might have occurred. In contrast, the ashes of the central and left part are pure white ashes seemingly in situ. This event was intensively sampled collecting 24 oriented samples of both facies (18 ashes and 6 carbonaceous samples). The archaeomagnetic mean direction obtained has been reported by Carrancho et al. (2013). Nevertheless, the objective here is to describe what magnetic features display in situ samples compared to those that are reworked. These results will allow testing the criteria established in a similar case study (Carrancho et al. 2012) as well as evaluating the degree of alteration that the structure might have suffered. 2.3 Laboratory methods All analyses were performed in the laboratory of palaeomagnetism of Burgos University (Spain). The measurement of the natural remanent magnetization (NRM) was carried out with a 2G SQUID magnetometer (noise level 5 × 10−12 Am2). Low-field magnetic susceptibility at room temperature was measured with a KLY-4 susceptometer (AGICO, noise level 3×10−8 S.I.). The NRM directional stability was analysed by stepwise progressive alternating field (AF) and thermal (TH) demagnetization. AF demagnetization was carried out in 18–20 steps up to maximum fields of 100–120 mT with the 2G magnetometer AF demagnetization unit. TH demagnetization was 6 performed using a TD48-SC (ASC) thermal demagnetizer in 15-17 steps up to 660 ºC. The Characteristic remanent magnetization (ChRM) direction of every specimen was determined by principle component analysis (PCA; Kirschvink, 1980) including at least four demagnetization steps (usually five or more). In order to study further the ferromagnetic mineralogy present, different rock-magnetic experiments were carried out with a variable field translation balance (MM_VFTB). These comprised progressive isothermal remanent magnetization (IRM) acquisition curves, hysteresis loops (± 1 T), backfield curves and thermomagnetic curves up to 700 ºC in air. These analyses were undertaken on representative bulk sample (~ 400 mg) both on ash and carbonaceous samples. Curie temperatures of Js-T curves were determined using the two-tangent method of Grommé et al. (1969). Saturation magnetization (Ms), remanence saturation magnetisation (Mrs) and coercive field (Bc) were calculated from hysteresis loops after subtracting the paramagnetic contribution. In combination with the coercivity of remanence (Bcr) determined from the backfield curves, the domain state distribution was analysed in the Day diagram (Day et al. 1977; Dunlop 2002). 3. Case 1: Archaeomagnetic dating 3.1. Background Archaeomagnetic dating is based on two fundamental phenomena. First, the ability of ferromagnetic minerals (s.l.) to acquire a remanent magnetization when heated and subsequently cooled from high temperatures parallel with and proportional to the geomagnetic field. This mechanism of magnetization is known as thermoremanent magnetization or TRM and is characteristic of structures such as ovens, kilns and hearths. Second, the Earth´s magnetic field undergoes subtle variations in direction and intensity on a timescale of 102-103 years on a regional scale. These fluctuations are known as secular variation (SV) and are reproducible for regions no bigger than 500600 km of radius (Lanos 2004). Over recent years great efforts have been undertaken to derive regional SV curves for different regions, particularly in Europe. These master curves are composed of directional and/or intensity data of the Earth´s magnetic field obtained from previously well-dated burnt archaeological materials (and occasionally also from lava flows). With some exceptions in Eastern Europe (Tema and Kondopolou 2011; Kovacheva et al. 2014), most European SV curves cover the last 2-3 millennia 7 (Gallet et al., 2002; Gómez-Paccard et al., 2006; Marton and Ferencz, 2006; Schnepp and Lanos 2005, 2006; Tema et al., 2006; Zananiri et al., 2007). Standard archaeomagnetic dating works on the basis of comparing the mean direction and/or intensity determined from a site with the SV curve available for the region and period concerned. Many archaeomagnetic dating examples are reported in the literature using directional, intensity data or both combined (e.g., Casas et al. 2007; EchChakrouni et al. 2013). The more archaeomagnetic data added to these regional SV curves the better defined they will be, thus improving the dating technique. More recently, archaeomagnetic dating using geomagnetic field models has become feasible. For instance, the SCHA.DIF.3K European regional model (Pavón-Carrasco et al. 2009) based exclusively on archaeomagnetic directional and intensity data for the last 3 millennia, directly predicts the geomagnetic field at the site of interest even for regions where no SV curve is available. This avoids any eventual relocation error which has been proved to introduce significant errors (Casas and Incoronato 2007). There are also global models for longer periods (e.g., Pavón-Carrasco et al. 2010; Korte and Constable, 2005; Korte et al., 2011) but not suited for archaeomagnetic dating because they include sedimentary data that smooth the geomagnetic field variations through time. Also, new software has been developed to carry out archaeomagnetic dating using various SV models (Pavón-Carrasco et al. 2011). Archaeomagnetic dating has a typical range of error of a few centuries although there are good examples reaching dating resolution of a few tens of years as the one reported from an early 18th century brick kiln by Casas et al. (2007). This depends on several factors such as sampling or analytical errors, inconsistent behaviour of the material or the rate of variation of the Earth´s magnetic field. However, dating applicability of the method depends on the length and completeness of the SV curve for the region concerned. The longest and systematic archaeomagnetic records for the last 8 ky exist for Eastern Europe (Tema and Kondopolou 2011; Kovacheva et al. 2014) but that is not the case for Western Europe as mentioned before. Current efforts aim to temporally and geographically extend SV records using well dated, in situ archaeomagnetic materials. Recent studies carried out on Mid to Late Holocene burnt anthropogenic cave sediments from the Iberian Peninsula (Carrancho et al. 2009, 2012, 2013) and Central Europe 8 (Kapper et al. 2014a,b) have allowed the extension to mid-Holocene times of the archaeomagnetic database and the dating technique. These authors showed how under certain conditions reliable archaeomagnetic directions can be obtained from these materials. As multiple burning events are usually present in these archaeological sequences, various archeomagnetic data (spanning a time period in the range of hundreds to thousands of years) can be obtained from a single site. Combining 26 new directions obtained from Neolithic, Chalcolithic and Bronze Age burnt levels from three caves in Spain with the existing archaeomagnetic database for Eastern Europe (Korte et al. 2011; Kovacheva et al., 2009), a directional European SV curve for the Neolithic exclusively based on archaeomagnetic (TRM) data was published (Carrancho et al. 2013). Although new results are being reported (e.g., Hervè et al. 2013a,b), archaeomagnetic data for times prior to around 1000 BC in Western Europe are rather scarce. Burnt anthropogenic cave sediments emerge thus as a new geomagnetic field recorder with a great potential both for geophysical and archaeological purposes. 3.2 Results and discussion 3.2.1 Magnetic properties Natural remanent magnetization values are between 4.08 x 10-5 and 8.27 x 10-4 Am2kg-1 whereas low-field magnetic susceptibility values oscillate between 6.42 x 10-7 and 4.78 x 10-6 m3kg-1. The highest values for both parameters correspond to the ashes indicating a major concentration of ferromagnetic minerals in this facies. The Koenigsberger ratio ([Qn = NRM/(χH) (cf . Stacey 1967)]) where χ is the magnetic susceptibility and H is the local geomagnetic field strength, yielded values between 1.6 and 19.6. These values agree well with others reported for similar materials (Carrancho et al. 2009, 2012; Kapper et al. 2014a,b) and indicates that the NRM is of thermal origin. The rock magnetic experiments carried out allowed characterizing the magnetic mineralogy, domain state and thermal stability. The IRM acquisition curves are almost saturated at fields of 150 – 200 mT indicating that they are dominated by a lowcoercivity mineral (Fig. 2). A small fraction of a high-coercivity mineral (up to 5-10 % of the SIRM or Saturation of IRM at 1T), most probably haematite, seems also to be present. However, its contribution to the magnetization is not significant. The Curie temperatures (TC) determined from thermomagnetic curves performed on selected samples are around 580 ºC indicating the dominance of magnetite in both facies (Fig. 9 3). Occasionally, TCS of up 615 ºC have been observed in some reddish brown ashes pointing out that stable maghaemite might also be present (Fig. 3b). The occasional presence of maghaemite has been already observed in this type of fire (e.g., Carrancho et al. 2009; 2013) and it would imply a thermochemical remanent magnetization (TCRM), making such specimens unsuitable for absolute archaeointensity determinations. The high thermomagnetic reversibility of ashes is noteworthy, particularly the white one (Fig. 3a). Conversely, carbonaceous specimens exhibit much lower thermomagnetic reversibility producing secondary magnetite on cooling (Fig. 3c). This indicates that they underwent lower heating temperatures as is explained in more detail in case study 2 (section 4). 3.2.2. NRM directional stability and archaeomagnetic dating Fig. 4 (a-f) illustrates representative NRM orthogonal demagnetization diagrams of both facies and the stereographic projection with all the individual Characteristic remanent magnetization (ChRM) directions determined. All specimens show a secondary viscous component of normal polarity easily removable in the first steps of the magnetic cleaning (< 10 – 15 mT or < 200 – 250 ºC) particularly evident in carbonaceous specimens (Fig. 4d-e). The NRM stability of the ashes is defined by a stable, high intensity normal polarity component almost demagnetized at 80–100 mT decaying univectorially towards the origin (Fig. 4a-b). AF demagnetized carbonaceous specimens exhibit also a single component (Fig. 4d) or occasionally two-component magnetizations partially overlapping. In the latter case, these specimens were not considered to calculate the ChRM direction. Three out of 5 specimens sampled for TH demagnetization of the NRM broke during laboratory analyses. The two remaining specimens (Fig. 4c and e) correspond to an ash and a carbonaceous specimen, respectively. The ChRM direction in the ash was determined between 250 ºC to 580-600 ºC. The ChRM direction in the carbonaceous specimen was defined between 250 ºC and 450 ºC, reflecting a partial thermo-remanent magnetization (pTRM) likely caused by moderate heating that this facies underwent. This is consistent with the irreversible thermomagnetic behaviour of this facies (e.g., Fig. 3c) as is more detailed in section 4 (case study 2). AF demagnetization is adequate to determine successfully the ChRM direction because the main remanence carrier is a low-coercivity mineral. 16 remains, mixing of burnt and natural sedimentary components and in the most extreme cases, the complete homogenization of the sediment. The implications of these processes are not only cultural but also chronological. Some authors have noted the importance of collecting samples for thermoluminiscence (TL), optical stimulated luminescence (OSL) and electron spin resonance (ESR) dating from undisturbed areas showing the least evidence of mineralogical change (e.g., Mercier et al. 1995; Bateman et al. 2007). The measurements of the radiation dose-rates can be seriously affected and not accurately reflect the dose-rates prevailing in the past. It is easy to understand the significant consequences derived from the correct assessment of the degree of alteration caused by these processes in terms of establishing a reliable age determination. Regardless of whether the responsible agent is anthropogenic, biogenic or geogenic (see Goldberg and Sherwood 2006 for a good synthesis), syn/post-depositional processes in cave fires can be generally grouped as physical and/or chemical. The latter imply mineralogical changes and diagenesis in general. Particularly, ash diagenesis from archaeological cave fires has been extensively studied over recent years with diverse techniques such as soil micromorphology, Fourier transform Infrared spectrometry (FTIR), geochemistry or scanning electron microscopy, among others (e.g., Weiner et al. 1993; 2002; Karkanas 2010; Bull and Goldberg 1985). Particularly interesting are some studies carried out on Middle Palaeolithic sites establishing a diachronic sequence of diagenetic alteration of calcite, the major component of wood ashes (e.g., Schiegl et al. 1996; Weiner et al. 1993, 2002). However, burnt anthropogenic cave sediments (and combustion features in general) are susceptible not only to diagenesis but also to reworking. That is, mechanical disturbances of the burnt sedimentary facies. Mechanical reworking of cave fires has been traditionally addressed through simple macroscopic or field observations. The absence of some of the facies composing these fires (rubefied sediment, charcoal and ashes), absence of their lateral continuity or mixing of burnt and unburnt material are the main criteria used. Recently, Mentzer (2014) detailed a comprehensive description of the main features characteristic of reworked combustion structures both at macro and microscale. The palaeomagnetic technique has been recently proposed to evaluate mechanical post-depositional processes in archaeological cave fires (Carrancho et al. 2012). This case study aims to test the reliability of the method determining to what extent the mechanical reworking 17 might have affected a partially bioturbated Late Holocene burning event from El Portalón Cave (Burgos, Spain; Fig. 8). 5.2. Results and discussion Representative examples of NRM demagnetization diagrams corresponding to ashes from different parts of the structure are shown in Fig. 8. Thermal demagnetization of a carbonaceous specimen from this event is shown in Fig. 6a (P3-16; Fig. 8) and whose characteristics are reported in section 3.2.2 (case study 2). The NRM demagnetization diagrams of specimens to the right side of the burrow (Fig. 8a-b) exhibit an anomalous and unstable directional behaviour. Qn ratio values are not greater than 1 and initial magnetization intensities (NRM0) are one order of magnitude lower than those from pure white ashes. On the contrary, NRM demagnetization plots to the left of the burrow (Fig. 8c-d) are defined by a stable single palaeomagnetic component, around 10 times more magnetic than carbonaceous samples, displaying high Qn ratio values and reproducible directions among them. The main magnetic carrier is a low-coercivity mineral as the normalized decay intensity plots indicate. According to thermomagnetic curves this mineral is low-Ti titanomagnetite or partially maghaemitized magnetite with Curie temperatures of around 580 ºC – 600 ºC (Fig. 9ac). Maghaemite might be responsible of the inflection observed at about 310 ºC in Fig. 9b, although it could also be due to change of grid structure. Even when these structures were partially affected by bioturbation, it is still possible to evaluate whether mechanical reworking extends beyond the visual alteration originally observed in the field in order to exclude those samples for calculating the mean archaeomagnetic direction. The quality selection criteria established by Carrancho et al. (2013) to obtain a reliable mean direction in these fires are related to the following factors: (i) a good preservation of the structure (presence of all the sedimentary facies for each burning event, meaning ashes over underlying carbonaceous facies), (ii) the intensity of the burning with regard to the quantity of fuel employed (ash thickness) and (iii) an efficient record of the magnetization (Koenigsberger ratio values greater than 1 and a majority of demagnetization diagrams with univectorial NRM among the ashes). 18 The results in this study are very similar to those reported by Carrancho et al. (2012) where the magnetic behaviour of two different burning events from El Mirador cave (one strongly bioturbated and other apparently in situ) was analysed and compared. It is evident that samples showing anomalous magnetic behaviour were reworked by the effect of bioturbation. However, the interesting fact as this case shows is that adjoining areas to the bioturbation may also suffer from reworking and in many cases this effect cannot be easily distinguished in the field. Although in this case it did not imply movement of archaeological remains in the stratigraphy (fumiers are usually not rich in archaeological materials), special care must be taken during the excavation of these fires. Archaeostratigraphic 3D projections of coordinated artefacts (e.g., pottery, lithic remains) can be particularly useful for a proper archaeological interpretation. From the magnetic point of view, a useful parameter with regard to TRM preservation is the Qn ratio. Koenigsberger values for this collection are between 1 and 7.3 (Fig. 10) whereas two out of three samples with values < 1 correspond to ashes from the reworked side (e.g., Fig. 8b). The other is a carbonaceous sample. On the basis of these results, the relationship between the in situ nature of the structure and the preservation of the TRM is obvious. Mechanical reworking promotes the disorganization of the magnetic moments of the ferromagnetic grains reducing the remanence but maintaining the bulk magnetic susceptibility. As this parameter does not depend on the orientation of the magnetic grains (excluding the anisotropy), the direct consequence is that the TRM is lost and Qn values become considerably reduced. Moreover, the multicomponent NRM structure of reworked samples is also indicative of alteration along with lower magnetization values. Carrancho et al. (2012) have described the importance of combining these analyses with macroscopic field observations such as determining the lateral continuity of the facies, absence of sedimentary mixtures, etc. No significant differences in terms of magnetic composition or domain state variation are observed between in situ and reworked ash samples from the rock magnetic experiments carried out. The backfield ratios obtained oscillates between 15.79 and 22.94 mT without distinctive differences between both types of samples. The hysteresis ratios obtained range from 0.116 < Mrs/Ms < 0.170 and 2.645 < Bcr/Bc < 4.380 (Fig. 11a), indicating a pseudo-single domain (PSD) state for the magnetite grains, which suggests that the granulometric distribution of both the in situ and reworked ashes is 19 quite similar. This homogeneity in magnetic properties can also be observed in the representative hysteresis loops shown in Fig. 11(b-c) and similar results were reported in analogous studies (Carrancho et al 2009, 2012; Kapper et al. 2014a,b). Summarizing, it is of primary importance for archaeomagnetic dating purposes to determine the in situ nature of a cave fire if only directional analyses are carried out. Magnetic orientation for archaeointensity determinations is not indispensable, although the material cannot be disaggregated. For archaeologists, the concept of “in situ” does not necessary mean the same as for archaeomagnetists. The latter look for burnt materials that preserve exactly the same position as they had when cooled. Any postdepositional movement, no matter how minimal, may have significant effects in the archaeomagnetic results. Archaeologists usually consider that a combustion feature remains in situ as long as artefacts or sediments do not experience significant stratigraphic movements which may compromise the cultural interpretation of the record. Using the above guidelines and when possible combining this information with that provided by other disciplines (e.g., micromorphology and FTIR) is the best way to infer the primary or secondary position of an archaeological combustion feature. 6. Conclusions Three applications of archaeoand rock magnetism to the study of burnt anthropogenic cave sediments have been reported in the following case studies: (i) archaeomagnetic dating; (ii) estimating palaeotemperatures and (iii) evaluating post-depositional processes. Case study 1: A mean archaeomagnetic direction was obtained from a burning event at El Mirador Cave. Its comparison with the directional European SV curve yielded several dating intervals. According to archaeological evidence, the most likely date of the last burning was 1651 – 1520 yr BC (95 % of confidence), slightly older than an independent radiocarbon date from this unit but both are archaeologically consistent. The agreement of the two dating methods reveals the potential of anthropogenic burnt cave sediments as geomagnetic field recorders as well as the possibility to be dated by archaeomagnetism. These data represent the first archaeomagnetic dating obtained in this type of materials. 20 Case study 2: Stepwise thermal demagnetization of the NRM of oriented carbocaneous samples is a useful method to estimate the last heating temperature. These samples show an intermediate palaeomagnetic component of normal polarity that we interpret as a pTRM with maximum unblocking temperatures of 400 – 450 ºC, representing the last heating temperature. These temperatures agree well with those obtained from partial thermomagnetic analyses. Case study 3: The archaeomagnetic analysis of a burning event partially bioturbated allowed to obtain a comparative characterization of the magnetic behaviour of in situ samples against reworked samples. The latter showed low NRM intensities (at least one order of magnitude), Qn ratios < 1 and multicomponent nature of NRM along with anomalous directions. Mechanical reworking extends beyond the deformation which one can visually identify in the field. Therefore, special care must be taken when excavating these features in order to interpret correctly the primary position of the materials. As a concluding remark, archaeomagnetic analyses on burnt anthropogenic cave sediments have a great potential not only from the geophysical point of view (reconstructing directional and/or intensity changes of geomagnetic field in the past) but also for archaeological purposes. We encourage our colleagues to work on this type of materials promoting multidisciplinary collaboration. Acknowledgments This work was funded by the Spanish Ministry of Economy and Competitiveness (MINECO projects CGL2012-32149 and CGL2012-38481). Special gratitude is devoted to the archaeological teams involved in the excavation of these sites by their efforts and much help in field work. References Angelucci, Diego E., Boschian, G., Fontanals, M., Pedrotti, A., Vergès, J. Mª., 2009. Shepherds and karst: the use of caves and rock-shelters in the Mediterranean region during the Neolithic. World Archaeology 41:2, 191 – 214. 21 Bardot, L., McClelland, E., 2000. The reliability of emplacement temperature estimates using palaeomagnetic methods: a case study from Santorini, Greece. Geophysical Journal International 143:39–51. Bateman, M.D., Boulter, C.H., Carr, A.S., Frederick, C.D, Peter, D., Wilder, M., 2007. Detecting post-depositional sediment disturbance in sandy deposits using optical luminescence. Quaternary Geochronology 2, 57-64. Boschian,G.,1997.Sedimentology and soil micromorphology of the Late Pleistocene and Early Holocene deposits of Grotta dell’ Edera (Trieste Karst, NE Italy). Geoarchaeology 12, 227–249. Brochier, J.E., 1983a. Combustion et parcage des herbivores domestiques. Le point de vue du sédimentologue. Bulletin de la Societé Préhistorique Française, 80 (5), 143–145. Brochier J.E. 1983b. Bergeries et feux de bois néolithiques dans le Midi de la France. Caractérisation et incidence sur le raisonnement sédimentologique. Quartär, 119-135. Brown K.S., Marean C.W., Herries A.I.R., Jacobs Z., Tribolo C., Braun D., Roberts D.L., Meyer M.C., Bernatchez J., 2009. Fire as an Engineering Tool of Early Modern Humans. Science 325, 859-862. Bull P.A, Goldberg P., 1985. Scanning electron microscope analysis of sediments from Tabun Cave, Mount Carmel, Israel. Journal of Archaeological Science 12: 177–185. Cabanes, D., Burjachs, F., Expósito, I., Rodríguez, A., Allué, E., Euba, I., Vergès, J.M., 2009. Formation processes through archaeobotanical remains: the case of the Bronze Age levels in El Mirador cave, Sierra de Atapuerca, Spain. Quaternary International, 193, 160–173. Calvo-Rathert, M., Carrancho, Á., Stark, F., Villalaín, J.J., Hill, M., 2012. Are burnt sediments reliable recorders of geomagnetic field strength? Quaternary Research 77, 326-330. 22 Canti, M. G., Linford, N., 2000. The effects of fire on archaeological soils and sediments: Temperature and colour relationships. Proceedings of the Prehistoric Society 66, 385–395. Carrancho, Á., 2010. Arqueomagnetismo y magnetismo de las rocas en registros de fuegos arqueológicos holocenos. PhD thesis. Universidad de Burgos, Spain. 282 pp. Carrancho, Á., Villalaín, J. J., Angelucci, D. E., Dekkers, M. J., Vallverdú, J., Vergès, J. M., 2009. Rock-magnetic analyses as a tool to investigate archaeological fired sediments: a case study of Mirador cave (Sierra de Atapuerca, Spain). Geophysical Journal International 179, 79–96. Carrancho Á., Villalaín J.J., 2011. Different mechanisms of magnetisation recorded in experimental fires: Archaeomagnetic implications. Earth Planetary Science Letters 312, 176–187. Carrancho, Á., Villalaín, J. J., Vergès, J. M., Vallverdú, J., 2012. Assessing postdepositional processes in archaeological cave fires through the analysis of archaeomagnetic vectors. Quaternary International 275, 14–22. Carrancho, Á., Villalaín, J. J., Pavón-Carrasco, F. J., Osete, M. L., Straus, L. G., Vergès, J. M., Carretero, J. M., Angelucci, D. E., González Morales, M. R., Arsuaga, J. L., Bermúdez de Castro, J.M., Carbonell, E., 2013. First directional European palaeosecular variation curve for the Neolithic based on archaeomagnetic data. Earth Planetary Science Letters, 380, 124–137. Carretero, J.M., Ortega, A.I., Juez, L., Pérez-González, A., Arsuaga, J.L., PérezMartínez, R., Ortega, M.C., 2008. A Late Pleistocene-Early Holocene archaeological sequence of Portalón de Cueva Mayor (Sierra de Atapuerca, Burgos, Spain). Munibe 59, 67-80. Casas, L., Incoronato, A., 2007. Distribution analysis of errors due to relocation of geomagnetic data using the Conversion via Pole (CVP) method: implications on archaeomagnetic data. Geophysical Journal International, 169 (2), 448–454. 23 Casas, Ll., Linford, P., Shaw, J., 2007. Archaeomagnetic dating of Dogmersfield Park brickkiln (Southern England). Journal of Archaeological Science 34, 205–213. Church, M.J., Peters, C., Batt, C.M., 2007. Sourcing fire ash on archaeological sites in the Western and Northern Isles of Scotland, using mineral magnetism, Geoarchaeology, 22 (7), 747–774. Cioni R, Gurioli L, Lanza R, Zanella E., 2004. Temperatures of the AD 79 pyroclastic density current deposits (Vesuvius, Italy). Journal of Geophysical Research 109:B02207. doi:10.1029/2002JB002251 Delhon, C., Martin, L., Argant, J., Thiébault, S., 2008. Shepherds and plants in the Alps: multiproxy archaeobotanical analysis of neolithic dung from ‘La Grande Rivoire’ (Isère, France). Journal of Archaeological Science, 35: 2937–52 Dunlop D.J., Özdemir Ö., 1997. Rock Magnetism. Fundamentals and Frontiers. Cambridge University Press, New York. 573 pp. Dunlop D.J., 2002, Theory and application of the Day plot (Mrs/Ms versus Hcr/Hc) 2. Application to data for rocks, sediments, and soils. Journal of Geophysical Research, 107, doi:10.1029/2001JB000487 Ech-Chakrouni, S., Hus, J., Spassov, S., 2013. Constraints of archaeomagnetic dating and field intensity determinations in three ancient tile kilns in Belgium. Studia Geophysica et Geodaetica 57 (4), 585-604 Fisher, R.A., 1953. Dispersion on a sphere. Proceedings, Royal Society of London A 217: 295–305. Gallet, Y., Genevey, A., Le Goff, M., 2002. Three millennia of directional variation of the Earth’s magnetic field in Western Europe as revealed by archaeological artefacts. Physics of the Earth and Planetary Interiors 131, 81–89. 24 Goldberg, P., Sherwood, S. C., 2006. Deciphering human prehistory through the geoarcheological study of cave sediments. Evolutionary Anthropology 15(1), 20–36. Gómez-Paccard, M., A. Chauvin, P. Lanos, G. McIntosh, M. L. Osete, G. Catanzariti, V. C. Ruiz-Martínez, J. I. Núñez, 2006. First archaeomagnetic secular variation curve for the Iberian Peninsula: Comparison with other data from Western Europe and with global geomagnetic field models. Geochemistry Geophysics Geosystems, 7, Q12001, doi :10.1029/ 2006GC001476. Gose, W.A., 2000. Palaeomagnetic Studies of Burned Rocks. Journal of Archaeological Science 27, 409–421 Grommé C.S., Wright T.L, Peck D.L, 1969. Magnetic properties and oxidation of irontitanium oxide minerals in Alae and Makaopuhi lava lakes, Hawaii. Journal of Geophysical Research, 74, 5277-5294. Herries A.I.R., 2009. New approaches for integrating palaeomagnetic and mineral magnetic methods to answer archaeological and geological questions on Stone Age sites. In: Fairbrain, A., O’Conner, S., Marwick, B. (Eds.), Terra Australis 28 – New Directions in Archaeological Science. The Australian National University Press, Canberra, Australia, pp. 235–253. Hervé, G., Chauvin, A., Lanos, P., 2013a. Geomagnetic field variations in Western Europe from 1500 BC to 2000 AD. Part I: Directional secular variation curve. Physics of the Earth and Planetary Interiors 218, 1–13. Hervé, G., Chauvin, A., Lanos, P., 2013b. Geomagnetic field variations in Western Europe from 1500 BC to 200 AD. Part II: New intensity secular variation curve. Physics of the Earth and Planetary Interiors 218, 51-65. Hrouda, F., Müller, P., Hanák, J., 2003. Repeated progressive heating in susceptibility vs. temperature investigation: a new palaeotemperature indicator? Physics and Chemistry of the Earth 28, 653–657. 25 Kapper, K. L., Anesin, D., Donadini, F., Angelucci, D. E., Cavulli, F., Pedrotti, A., Hirt, A. M., 2014a. Linking site formation processes to magnetic properties. Rockand archeomagnetic analysis of the combustion levels at Riparo Gaban (Italy). J Journal of Archaeological Science 41, 836–855. Kapper, K. L., Donadini, F., Mauvilly, M., Panovska, S., Hirt, A. M., 2014b. New directional archeomagnetic data of burned cave sediments from Switzerland and geomagnetic field variations in Central Europe. Geophysical Journal International 198, 1208–1221. Karkanas, P., 2010. Preservation of anthropogenic materials under different geochemical processes: a mineralogical approach. Quaternary International 214, 63-69. Kent, D.V., Ninkovitch, D., Pescatore, T., Sparks, R.S.J., 1981. Palaeomagnetic determination of emplacement temperature of the Vesuvius AD 79 pyroclastic deposits. Nature 290:393–396. Kirschvink, J., 1980. The least-squares line and plane and the analysis of paleomagnetic data. Geophysical Journal, Royal Astronomical Society 62, 699–718 Korte M., Constable C.G., 2005. Continuous geomagnetic field models for the past 7 millennia: 2.CALS7K. Geochemistry Geophysics Geosystems 6, Q02H16, DOI: 10.1029/2004GC000801 Korte, M., Constable, C., Donadini, F., Holmes, R., 2011. Reconstructing the Holocene geomagnetic field. Earth and Planetary Sciences Letters 312, 497–505. Kovacheva, M., Boyadziev, Y., Kostadinova, M., Jordanova, N., Donadini, F., 2009. Updated archeomagnetic data set of the past 8 millennia from the Sofia laboratory, Bulgaria. Geochemistry Geophysics Geosystems 10, Q05002, http://dx.doi.org/10.1029/2008GC002347. Kovacheva, M., Kostadinova-Avramova, M., Jordanova, N., Lanos, P., Boyadzhiev, Y., 2014. Extended and revised archaeomagnetic database and secular variation curves Figure 1 Click here to download high resolution image 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0 100 200 300 400 500 600 700 800 900 1000 1100 Ci1-6 carbonaceous Ci1-7 carbonaceous Ci1-16 ash Ci1-19 ASH NormalizedIRM Appliedfield(mT) Figure 2 0 100 200 300 400 500 600 700 0.0 Mag[x10 Am²/kg] -2 Sample:Ci1-19 Whiteash 2.0 4.0 6.0 Temp[°C] a) 700 Sample:Ci1-16 reddishbrownash 0 100 200 300 400 500 600 Temp[°C] 0.0 Mag[x10 Am²/kg] -2 1.0 2.0 3.0 b) Sample:Ci1-7 carbonaceous 7000 100 200 300 400 500 600 Temp[°C] 0.0 4.0 6.0 8.0 2.0 Mag[x10 Am²/kg] -2 c) T =613ºC C T =585ºC CT =575ºC C Figure 3 Ci1-24 ( =4.3) ASHQn NRM =1.68x10 Am /kg 0 -4 2 A.F.(mT) M/M0 0 20 40 60 80 100 0.0 0.5 1.0 N up/W NRM 10mT 15mT 20mT 100mT a) A.F.(mT) 0 20 40 60 80 100 M/M0 0.0 0.5 1.0 N up/W 10mT 15mT 20mT 100mT NRM Ci1-25 ( =4.8) ASHQn NRM =5.40x10 Am /kg 0 -4 2 b) d) A.F.(mT) 0 20 40 60 80 100 M/M0 0.0 0.5 1.0 N up/W Ci1-5 ( =2.4)CARBONACEOUSQn NRM =1.06x10 Am /kg 0 -4 2 100mT 10mT 15mT NRM N=18Dec=20.1ºInc=56.5º =63.3 95=4.4ºka N f) c) M/M0 0.0 0.5 1.0 2000 T(ºC) 600 400 N up/W 100º 200º 250º 300º 400º 450º NRM 580º Ci1-8 ( =3.3) ASHQn NRM =9.24x10 Am /kg 0 -5 2 e) N up/W Ci1-1 ( =1.6)CARBONACEOUSQn NRM =4.08x10 Am /kg 0 -5 2 580º 200 0 T(ºC) 600 400 0.0 NRM 200º 350º M/M0 0.5 1.0 250º 400º 150º TH TH Figure 4 Figure 5 Click here to download high resolution image 0 100 200 300 400 500 600 700 T (ºC) 0.0 0.5 1.0 M/M0 TH 0 100 200 300 400 500 600 700 T (ºC) 0.0 0.5 1.0 M/M0 TH 0 100 200 300 400 500 600 700 T (ºC) 0.0 0.5 1.0 M/M0 TH 0 100 200 300 400 500 600 700 T (ºC) 0.0 0.5 1.0 M/M0 TH 0 100 200 300 400 500 600 700 T (ºC) 0.0 0.5 1.0 M/M0 N up/W NRM 100º 260º 300º 200º 150º 325º N up/W 400º 450º 500º 600º 357º 475º 357º 400º P3-16 ( =1.18)QnNRM =5.46x10 Am /kg 0 -5 2 a) TH RM2-9 (Q =1.2) nNRM =1.73x10 Am /kg 0 -5 2 NRM 150º 250º 300º 200º c) 300º 400º 450º 550º 600º 300º 400º E up/N P2-01 ( =1,03)QnNRM =3,21x10 Am /kg 0 -5 2 E up/N 500º 600º 542º 325º 300º 357º 400º RM1-11 ( =1.71)QnNRM =5,97x10 Am /kg 0 -5 2 b) d) p-TRM E up/N E up/N 450º 530º 600º 350º 300º p-TRM p-TRM p-TRM 0 100 200 300 400 500 600 700 T (ºC) TH 0.0 0.5 1.0 M/M0 N11-21 ( =0.77)QnNRM =1,54x10 Am /kg 0 -5 2 f) N up/W NRM 100º 200º 250º 300º 300º 350º N up/W 450º 580º N9-29 (Q =1.3) nNRM =9.01x10 Am /kg 0 -6 2 e) N up/W N up/W 580º 250º 300º NRM 200º 100º 350º 500º 400º 350º p-TRM NRM 125º 200º 300º 250º p-TRM NRM 150º 200º 260º Dec/Inc(pTRM)=330.9º/53.6º Dec/Inc(pTRM)=25.4º/34.4º Dec/Inc(pTRM)=339.4º/58.7º Dec/Inc(pTRM)=18.6º/71.9º Dec/Inc(pTRM)=24.6º/57.6º Dec/Inc(pTRM)=29.0º/62.1º Figure 6 0 100 200 300 0.0 Temp[°C] Mag[x10 Am²/kg] -2 0 100 200 300 400 350ºC Temp[°C] 0.0 0.0 0100 200 300 400 Temp[°C] HeatingJ =1.28x10 Am kg 30 -2 2 -1 400ºC 450ºC d) Sample:P2-22_350 HeatingJ =1.26x10 Am kg 30 -2 2 -1 Sample:P2-22_400 1.21.2 1.2 HeatingJ =1.28x10 Am kg 30 -2 2 -1 Sample:P2-22_450 e) f) 500ºC HeatingJ =1.16x10 Am kg 30 -2 2 -1 550ºC HeatingJ =1.26x10 Am kg 30 -2 2 -1 0100 200 300 400 Temp[°C] 500 0.0 1.2 h) 0.0 Mag[x10 Am²/kg] -2 g) 1.2 0100 200 300 400 Temp[°C] 500 Sample:P2-22_500 Sample:P2-22_550 200 250 300 400 Temp[°C] 500 350 450 550 -10 0 10 20 30 40 50 60 Heatingalterationindex accordingtoHroudaetal.(2003) A (%) 30 i) a) 0.0 1.2 Sample:P2-22_250 0100 200 Temp[°C] 250ºC HeatingJ =1.28x10 Am kg 30 -2 2 -1 b) 0 100 200 300 Temp[°C] Sample:P2-22_300 0.0 1.2 c) 300ºC HeatingJ =1.30x10 Am kg 30 -2 2 -1 0 100 200 300 400 500 600 700 Temp[°C] Sample:P2-22 (upto700ºC) 0.0 Mag[x10 Am²/kg] -2 1.0 2.0 3.0 4.0 5.0 Figure 7 Figure 8 Click here to download high resolution image Sample:P3-3 whiteash 2.0 0.0 Mag[x10 Am²/kg] -2 1.0 700 0100 200 300 400 500 600 Temp[°C] a) 1.0 0.0 7000 100 200 300 400 500 600 Sample:P3-6 whiteash Mag[x10 Am²/kg] -2 Temp[°C] b) 0.0 2.0 3.0 4.0 1.0 Mag[x10 Am²/kg] -2 5.0 Sample:P3-19 carbonaceous 7000 100 200 300 400 500 600 Temp[°C] c) T =575ºC C T =310ºC C1 T =585ºC C2 T =593ºC C Figure 9 101 NRM(A/m) 100 10-1 10-2 10-3 Susceptibility(S.I.) 10-2 10-3 10-4 10-5 10-1 Qn= 0.1 Qn=1 Qn= 10 Koenigsbergerratio,Qn 100 Qn= Ashes Carbonaceous Figure 10 6 This has been answered before (1st question, section 3.2.1). Burnt anthropogenic cave sediments are novel materials for archaeomagnetism and the few studies available report Qn ratio values mostly comprised between 1 and 10, with the highest values in ashes and the lowest in carbonaceous samples (see Carrancho et al. 2009, 2012, 2013; Kapper et al. 2014a,b). See also Fig. 4 and 10 of this manuscript. We claim that carbonaceous samples recorded a pTRM in the thermal demagnetization diagrams and Qn ratios > 1 might be an indication of it, so the sentence has been modified accordingly. Results from partial thermomagnetic curves and reversibility experiments (Fig. 7) are also a proof of it, as we justify in the next paragraph of the main text.  Page 12, last paragraph (section 4.2): “…partial thermomagnetic runs were carried out on a sister powered sample…” Ref.1: What does sister powered mean? You mean another, powdered sample? Yes, another powdered sister sample. It is now indicated.  Page 12, end of last paragraph (section 4.2): “This alteration can be quantitatively estimated (see Hrouda et al 2003) and starts at 450 – 500 ºC… Ref.1: do you want to make a quantitative estimate? It´s done and shown in fig. 6i. In addition, more details are added following indications of ref. 2.  Page 13, middle of 1st paragraph (section 4.2): “Canti and Linford (2000) also reported temperatures of around 400 ºC beneath fires exceeding 800 ºC and…”. Ref.1: 400° beneath 800°? What do you mean? The sentence has been modified to make it clearer. 400 ºC refers to the substrate and 800 ºC to the ashes.  Page 13, end of last paragraph (section 4.2): “The usefulness of the palaeomagnetic method for determining heating temperatures in burnt anthropogenic cave sediments is certainly of high value for the archaeologists”. Ref.1: Why is it? You haven't clarified what these temperatures represent. This is now better explained in the main text (1st paragraph section 4.1). 5. Case study 3: Assessment of post-depositional processes 5.1 Background  Page 14, middle of last paragraph: “…establishing a diachronic sequence of diagenetic alteration of calcite, the major component of wood ashes”. Ref.1: is calcite the major component of wood ash?? Yes, it is. This is well known and there are many papers published (e.g., Schiegl et al. 1996, Weiner et al. 1993, 2000). These are cited. 5.2 Results and discussion  Page 15, end 2nd paragraph (section 5.2): “…given the inflection at intermediate temperatures of Fig. 8b…”Ref.1: More specific temperature range Ok, it is included. This sentence was modified also following requirements of Ref. 2.  Page 15, end 3rd paragraph (section 5.2): “…(iii) an efficient record of the magnetization…”. Ref.1: what does "efficient" mean? It means that the magnetization was recorded efficiently, in a quick and trustworthy way. It does not require further explanation.  Page 16, 2nd paragraph: “…and two out of the three samples with values < 1 correspond to ashes from the reworked side (e.g.: Fig. 7b). The other is a carbonaceous sample. On the basis of these results, the relationship between the in situ nature of the structure and the preservation 7 of the TRM is obvious.”. Ref.1: I would not say that 2 out of 3 makes for an obvious relationship 2 out from 3 samples may not be a particular statistically robust result. However, what is particularly interesting is the relationship between their location (right side of the burrow, in the bioturbated zone) and their low Qn ratio values, always < than 1. That´s not a coincidence and we observed the same behaviour in the bioturbated event studied by Carrancho et al. (2012). We really think that there is relationship between low Qn ratio values (< than 1) and reworked samples. Furthermore, that relationship is complemented by the other features described (e.g., high intensity, univectorial NRM diagrams, reproducible directions among specimens). It is explained with enough detail in that paragraph.  Page 16, 1st lines last paragraph (section 5.2): “This is critical for directional analyses but not so much for absolute archaeointensity determinations since magnetic orientation is not indispensable”. Ref. 1: but if a material is disaggregated, it will not give a valid paleointensity That´s right. For directional analyses orientation is critical, not so for archaeointensity analyses. Archaeointensity can only be performed on compact (not disaggregated) samples, mainly because of the numerous heatings steps required. The sentence has been modified.  Last paragraph page 16 / 1st paragraph page 17: “For archaeologists, the concept of “in situ” does not necessary mean the same as for archaeomagnetists”. Ref. 1: so what does it mean for archaeologists? A sentence explaining it has been included. 6. Conclusions  Page 17: “As a concluding remark, archaeomagnetic analyses on burnt anthropogenic cave sediments have a great potential ... but also for archaeological purposes” Ref. 1: Where in this paper are those archaeological purposes made explicit? We have reported three different applications of archaeological interest (archaeomagnetic dating, estimating palaeotemperatures and assessing post-depositional alterations). For example, they are explicitly mentioned in the abstract and in the last paragraph of the introduction: “The main goal of this paper is to highlight the potential of magnetic methods to answer archaeological questions through three different applications…, etc.”. Certainly, they provide valuable information for “archaeological purposes”. Specific comments to the PDF file (Reviewer #2): Minor changes suggested by reviewer 2 complementary to those from reviewer 1 have been introduced. Please, find below detailed answer to the most important questions. Abstract (Ref. 2): “This is the first archaeomagnetic dating obtained in these contexts so far”. Reviewer 2 suggests to remove this sentence. We prefer to maintain it because is true and highlights the relevance of the dating attempt carried out in the case study 1. 1. Introduction  Page 2, last line 1st paragraph: “…its application in prehistoric archaeology is still sporadic and its potential to retrieve archaeological information remains underutilized.” Referee 2 suggests changing “remains underutilized” by “is mainly focused on archaeomagnetic dating”. Archaeomagnetic dating in prehistoric materials has been barely used because available secular variation curves only reach the last 2-3 millennia. Furthermore, there isn´t any archaeomagnetic dating specifically carried out on these materials yet. For these reasons we leave the comment. 2. Materials and methods 8 2.1 Sampling  Pages 3-4: “…As three different case studies with different applications are reported, specific details of sampling and laboratory analysis will be given in each one of them”. Ref. 2: I would add this information here in this sub chapter rather than in 3 separate subchapters, in order not to distract from the case studies Following suggestions of both reviewers, section 2.2 now includes sampling details for each case study. Sampling subsections in the previous version (3.1.2 and 3.3.2) are now removed. 2.2.1 Case study 1 (previous section 3.1.2)  Page 4, section 2.2.1: “…The ashes are white on top and reddish brown on the bottom with a total thickness of about 15 cm”. Is it also ash if it is reddish brown? Or could it also be a thermally altered part? It is certainly a thermally altered facies but we considered it as ashes (distinguishing the colour) because they are directly above the underlying carbonaceous facies, which represents the substrate upon which the heating took place. The similarity in the magnetic properties between white and brown ashes from Ci1 event in terms of magnetic carrier, mineral magnetic concentration as well as domain state is a clear indication that they underwent high temperature heating as expected in ashes. It can be observed in Fig. 2 (IRM curves), Fig. 3a-b (thermomagnetic curves) and Fig. 4 (NRM demagnetization diagrams). 2.3Laboratory methods This information previously given in other sections in the first version is now reported here. 3.1 Background  Page 6, end 2nd paragraph (section 3.1): “…but not suited for archaeomagnetic dating because they include sedimentary data that smooth the geomagnetic field variations through time” Ref. 2: This depends on the time period. Besides, a record of lake sediments might not be wrong, but only smoothed. Yes, the reviewer is right but if the record is smoothed (and is well known that sedimentary data from lakes or marine sequences produces that effect), it is not suited for archaeomagnetic dating. It can be used for correlating sequences, but not for dating. The consensus within the archaeomagnetic community is that the design of secular variation curves must be done with materials carrying a thermoremanence (TRM). For this reason we leave the statement.  Page 7, first line: “…that is not the case for Western Europe as mentioned before”. Ref. 2 suggest: whereas for Western Europe the longest record reaches back only XXXX years. It is said at the end of the 1st paragraph of this section 3.1: “…most European SV curves cover the last 2-3 millennia…”. We leave it to avoid repetitions. 3.2.1 Magnetic properties Representative examples of IRM acquisition curves are now in the new Fig.2.  End page 7 / beginning page 8: “The Curie temperatures (TC) determined from thermomagnetic curves performed on selected samples are around 580 ºC indicating…”. Ref. 2: please add error range. Curie temperatures were calculated with the two-tangent method of Grommé et al. (1969). It´s now included in the main text (2nd paragraph, section 2.3). The Curie point is determined projecting onto the abscissa axis (X-axis) the cross point of the two tangents. So it is a visual estimate. However we estimate that the error range is of ± 10 ºC in the worst case, but that depends on every curve, its quality signal or the slope. This analysis is used to infer the ferromagnetic mineralogy and in practical terms, these facies are all dominated by low-Ti titanomagnetite so adding this information is not particularly useful. 9 3.2.2. NRM directional stability and archaeomagnetic dating  Page 8, 1st paragraph (section 3.2.2): “All samples show…” Ref.2: are they from samples or specimens? It is specimens. We have checked it along the manuscript.  Page 8, 2nd paragraph (section 3.2.2): “The ChRM direction in the carbonaceous sample was defined between 250 ºC and 450 ºC, reflecting a partial thermo-remanent magnetization (p-TRM) likely caused by moderate heating that this facies underwent” Ref.2: This sample is demagnetized by about 450 degrees reflecting moderate heating that this facies underwent Please, see answer to this question in answers to Ref. 1 (page 4, this document). It is specifically explained in reference to Reviewer 2.  Page 8, last paragraph: “Following the quality selection criteria established by Carrancho et al. (2013), …” Ref.2: please summarize the selection criteria here shortly They are now included in the main text (3rd paragraph, section 3.2.2)  Page 9, end 1st paragraph (section 3.2.2): “As is discussed further in the case study 3 (section 3.3), all these features are indicative of some type of post-depositional reworking”. Ref.2:.or heating to low temperatures? That is not likely because in situ ashes from this burning event show very high intensities, univectorial NRM demagnetization diagrams or high values of Qn ratio. All of them features related with their in situ and well-heated nature. If the samples with “anomalous” magnetic behaviour (e.g., anomalous directions or multicomponent demagnetization diagrams, etc) come from the bioturbated area, post-depositional reworking is most likely the cause of such results.  Page 9, end of 2nd paragraph: (about the archaeomagnetic dating of case study 1) “The last one is within the bounds of possibility but is out of the radiocarbon date range (1510 - 1410 yr BC) by more than three centuries. Ref.2:.but this age has the largest probability according to fig. 4. Can you explain this discrepancy? It is true that this age range has the largest probability from the statistical point of view. However, it does not seem to be archaeologically consistent and is out of the radiocarbon date range (1510 - 1410 yr BC) by more than three centuries. This was pointed out but a brief comment is now included. Archaeomagnetism is a relative dating method and irrespective of the possible ages obtained, they must be coherent with the archaeological context to be reliable.  Page 9, last paragraph of section 3.2.2: “Beyond that, the important fact is that it is already possible to date with archaeomagnetism burnt archaeological features from Western Europe…” Ref.2: Please reform, not so clear to me. Ok, this paragraph has been modified to improve its understanding. 4. Case study 2: estimating heating temperatures 4.1 Background  Page 10, end of 2nd paragraph section 4.1: Ref.2:.Here you could also cite Rada Torres et el. (2011) Ok, a brief comment about this reference is now included. 4.2 Results and discussion  Page 11, 1st paragraph section 4.2: “…from El Mirador, Portalón and El Mirón Cave (Spain).” Ref.2:.please add reference to Fig. 1 after adding location Map of Fig. 1 now includes location of sites. The reference to Fig. 1 is added in the text here. 10  Page 11, 2nd paragraph section 4.2: “Finally, a high temperature component…”. The acronym “HT” (high temperature) is added after its first time cited. Thus we avoid repetitive text hereinafter.  Page 11, 2nd paragraph section 4.2: “After removing a low temperature component probably of viscous origin (< 150 – 200 ºC),…” Ref.2: I see maximum temperature of the viscous component of 125 degrees Looking carefully, max TUB of the secondary low-temperature (viscous) component for some diagrams reaches 200 ºC (e.g., Fig. 6c or d). It is observable looking the demagnetization vectors for each diagram combining both the horizontal and the vertical projection. For example, in Fig. 6d the max TUB of this viscous component is not 125 ºC, but clearly 200 ºC (see solid dots of NRM demagnetization diagram; the horizontal plane). Underestimating the max TUB of the viscous component implies errors determining the ChRM direction.  Page 11, end of 2nd paragraph section 4.2: Ref.2: (replace the previous sentence for this one): This is at about 400-450 degrees where the intermediate magnetization component switches the direction ... Ok, it has been changed (end of 2nd paragraph, section 4.2).  Page 11, 3rd paragraph section 4.2: “The fact that the intermediate magnetization component lies along the Earth´s magnetic field direction is the basic principle of the technique in these materials” Ref.2:.What are D and I of the present geomagnetic field at this location? Declination and inclination for the three sites studied are shown below. They were calculated for the 2015 September 28th, with the WMM2015 model. However, we would like to give an explanation to this reviewer´s comment. These materials are Holocene so is obvious that, if they are in situ, they all should show normal polarity as is the case. It has no sense to perform any comparison of the pTRM directions obtained with the present geomagnetic field at the studied sites because of the secular variation (SV), since they do not necessary have to coincide. Precisely because of the SV, some directional variation with respect to the present field is expected for mid latitudes as the Iberian Peninsula (e.g., Gómez-Paccard et al. 2006): ± 20 º in declination and between about 45º to 70º in inclination. So, indicating the present field direction for each site will not give any useful information to the reader and will introduce confusion. Please, see next answer. The present geomagnetic field at every location is (Fig. 6 includes examples from 3 sites, not only one): -El Mirón Cave: -1° 4' 7" (W) / 58º 27´ 19´´ -Mirador Cave (Sierra de Atapuerca): -1° 1' 53" (W) / 57° 26' 50" -Portalón Cave (Sierra de Atapuerca): -1° 2' 4" (W) / 57° 26' 49" Following this reasoning, why not to calculate the field direction for year 2000 or 1950 or 1900? It is a way of saying that this information is not useful to the reader. Please add D and I of the pTRM in Fig 5. It is now included in the new figure 6. The important point is that the pTRM directions are northward as we argue in the text and is now shown in Fig. 6. This information is really helpful to the reader.  Page 12, end 2nd paragraph (section 4.2): “Nonetheless, some results suggest a TRM origin of the magnetization…”. Ref.2: which results? please specify! We referred to the Qn ratio explained in the next sentence and also to the J-T curves explained in the next paragraph. Following also suggestions of Ref. 1, the sentence has been modified. 11  Page 12, end 2nd paragraph (section 4.2): Ref. 2: Figure reference is wrong, should be 6 to keep the order. Please change the subsequent figure numbers. Sorry, but the reference to that figure was correct. We referred to sample N11-21. As a new figure 2 has been included, this is now figure 6f.  Page 12, end of last paragraph: “This alteration can be quantitatively estimated…” Ref.2: Please put one sentence here, which explains how it is done. Ok, a brief explanation is now included.  Page 12, end of last paragraph: “This alteration can be quantitatively estimated and starts at 450 – 500 ºC, reaching a maximum at 550 ºC (Fig. 6i)” Ref.2: How do you know that is is not at more than 550 degrees? You do not have partial thermomagnetic curves up to 700 each 50 degree steps. Why is A30(%) for 700 degrees missing in the figure? PLease add it in Fig. 6i. We don´t know empirically because the maximum heating temperature applied in this experiment was 550 ºC, as it is said a few lines before in the main text. However, from 600 ºC to 700 ºC the alteration index progressively will reduce because magnetite neoformation is no possible. These temperatures are over the Curie temperature of magnetite (Tc ~ 580 ºC), so the sample loose its ferromagnetism. Unfortunately, this sample cannot be analyzed again because our Balance is currently not working due to a breakdown. However, to demonstrate our argument, we show below results from other carbonaceous sample from El Mirador Cave (sample FU1-28; see figure below) on which this experiment was performed from 250 ºC to 700 ºC. Please, note how the maximum alteration occurs between 400 and 550 ºC (exactly the same as the example shown in Fig. 7) and from 550 ºC to 700 ºC the alteration index is reduced for the reason given above. This example cannot be incorporated into the main text because we don´t have specifically a TH demagnetization diagram of the NRM of this sample and the idea is to compare the “pTRM method” with this partial thermomagnetic curve experiments on carbonaceous samples from the same burning event. Anyway, this result confirms that the range of temperatures at which carbonaceous facies were heated is comprised between 400 – 550 ºC. See graph below. 12  Page 13, 1st lines 1st paragraph: “As far as the ashes are concerned, these most likely reached temperatures over 600 – 700 ºC…”. Ref.2: Please reference here, Figure not shown, or explain from which experiment you got this results. The sentence has been rewritten following indications of Ref. 2. The references to the studies where this information come from were already in the text as well as a description of their behaviour.  Page 13, middle of 1st paragraph: “…from a series of actualistic fire experiments.” Ref.2: I do not understand... actual? Actualistic is correct. “Actualistic study”: a detailed observation of the actual use of archaeological artifacts, ecofacts, and features, used to produce general analogies for archaeological interpretation 13 5. Case study 3: Assessment of post-depositional processes 5.1 Background  Page 14, 1st paragraph (section 5.1): “…samples for TL, OSL and ESR dating…”. Ref.2: please explain abbreviations. Ok, they are now expanded.  Ref. 2: Please move also this part to chapter 2 (about the previous 3.3.2 section: Sampling and laboratory analyses). It has been moved and described in 2.2.3 subsection (sampling case study 3). 5.2. Results and discussion.  Page 15, 1st paragraph section 5.2: Ref.2: Fig. 7: please add location of P3-16 in figure. Ok, it has been added. It is new figure 8.  Page 15, 1st paragraph section 5.2: “…NRM demagnetization plots from the central-left part of the burning event …”. Ref.2: burrow?. The sentence has been rewritten to be better understood.  Page 15, 1st paragraph section 5.2: “…displaying high values of the Qn ratio…”. Ref.2: in the range of xxx. It is specifically said on page 16 (2nd line, 5th paragraph of section 5.2), when taking about the Qn ratio. It is also visible in Fig. 10.  Page 15, 2nd paragraph section 5.2: According to thermomagnetic curves it is low-Ti Titanomagnetite with Curie temperatures of around 580 ºC (Fig. 8a-c) and possibly also maghaemite given the inflection at intermediate temperatures of Fig. 8b”. Ref. 2: Fig. 8a seems to have a Tc at about 600 degrees. PLease clarify.. Yes, for Fig. 9a the Tc is more 600 ºC than 580 ºC. It is better explained in the text now and Curie temperatures indicated in Fig. 9(a-c)  Page 15, end of 2nd paragraph section 5.2: “…possibly also maghaemite given the inflection at intermediate temperatures of Fig. 8b”. Ref. 2: the inflection might also be due to change of grid structure. Ok, it has been included.  Page 15, 3rd paragraph section 5.2:: “…(ii) with the intensity of the burning (ash thickness)…” Ref. 2: a lot of ash might be produced by a lot of fuel, but does not mean that burning took long. Yes, that´s true. It is now better indicated in the main text.  Page 16, end of 2nd paragraph: “It has been claimed the importance of combining these analyses with macroscopic field observations (Carrancho et al. 2012).” Ref. 2: This sentence is not clear to be, please reform. The sentence has been modified to make it clearer.  Page 16, 3rd paragraph: Ref. 2: What about the other rock magnetic experiments: IRM, backfield, hystereses.... do they show differences between disturbed and undisturbed parts? Please mention here too. A paragraph has been included with an appropriate explanation. No significant differences were observed between the in situ and the reworked ashes. A new figure 11 was included. 14 References: The following references have been added: Calvo-Rathert et al. (2012) Carrancho (2010) Dunlop (2002) Grommé et al. (1969) Kirschvink (1980) Vergès et al. (this issue). Figures:  Figure 1: Site locations of all sites were added in the map  Figure 2: A new Figure 2 was created showing representative IRM curves from Ci1 event.  Figure 3: The corresponding TCS for each curve were added in the figure (Ref. 2).  Figure 5: The last line of the legend was eliminated (Ref. 1).  Figure 6: Declination and Inclination of the pTRM component was indicated for each panel (Ref. 2)  Figure 8: Location of specimen P3-16 (carbonaceous) was inserted in the photo (Ref. 2).  Figure 9: The corresponding Tcs for each curve was added in the figures (Ref. 2).  Figure 10: “(S.I.)” was not eliminated as Ref. 2 suggested. It refers to “Système Internationale” and is necessary to indicate it to differentiate from the “cgs” system (centimetre, gram, second).  Figure 11: A new figure 11 was included with a Day plot and two representative hysteresis loops of an in situ and a reworked ash, respectively. (Ref. 2). Caption figures: They were revised following reviewers´ suggestions.