| IODP Proceedings Volume contents Search | |||
|
|||
| Expedition reports Research results Supplementary material Drilling maps Expedition bibliography | |||
|
doi:10.2204/iodp.proc.313.102.2010 Downhole measurementsBorehole geophysical instrumentsOffshore downhole logging operations for Expedition 313 were provided by the University of Montpellier (Laboratoire de Géosciences-Montpellier; France), part of the EPC under ESO. The EPC is primarily responsible for the planning, acquisition, QA/QC, science support, and education outreach related to petrophysical measurements on MSPs. Petrophysics in relation to this consortium includes downhole measurements (i.e., borehole logging) and physical properties (continuous and discrete; see "Physical properties") measured on cores (e.g., magnetic susceptibility, acoustic properties, and density and porosity). The set of downhole geophysical instruments utilized during Expedition 313 was constrained by the scientific objectives, the coring technique (PQ and HQ diameter boreholes), and the hole conditions at the three sites. Slimline logging tools of several manufactures were used (Table T6). The suite of downhole geophysical methods was chosen to obtain high-resolution images of the borehole wall, to measure borehole size, and to measure or derive petrophysical or geochemical properties of the formation such as porosity, electrical resistivity, acoustic velocities, and natural gamma radioactivity. Deployment of nuclear tools was not allowed during Expedition 313. The slimline suite comprised the following tools:
Vertical seismic profiles (VSPs) were acquired by the University of Alberta (Department of Physics; Canada) in Holes M0027A, M0028A, and M0029A for two primary purposes: (1) to calibrate time to depth correlations of the seismic reflection profile and, upon additional processing, (2) to construct seismic traces from the upgoing wavefield that may be directly compared to the reflection profile. OperationsThe logging team consisted of two engineers supervised and assisted by the Petrophysics Staff Scientists. A total of three boreholes (Holes M0027A, M00028A, and M0029A) were prepared for downhole geophysical measurements. Where possible, measurements were performed in open borehole conditions (no casing). The majority of spectral gamma ray logs and VSP measurements were done through the casing. Despite difficult borehole conditions (nonconsolidated formations, risk of collapse, etc.), the recovery and overall quality of the downhole logging data are very good. In each hole, spectral gamma logs were acquired as the first measurements through the steel drill string. After completion, the pipe was pulled up in steps because of difficult borehole conditions, the development of bridges, and difficulties in pulling the pipe up (e.g., in sandy sections in the upper 200 m of each hole and in other sand-prone intervals in Holes M0027A and M0028A). After conditioning the hole with drill mud, open-hole measurements were logged at each step. When logged in sections, each slimline tool run was identified as (1) lower, (2) middle, and (3) upper. In each section, if some tools were logged in several runs, runs are numbered 1 (deepest) upward. Because of borehole conditions and time constraints it was not possible to log with all tools in every borehole section (Table T7). Logging procedure and data recordingEach slimline tool was logged on an individual string. A logging run commenced with zeroing the tool to a fixed point (drillers zero) corresponding to the top of the drill pipe (Fig. F15). The zeroing point was measured for each run because of depth discrepancies that may arise when moving from one open interval to another. Zeroing the tool in this way is important for achieving the most accurate postprocessing depth integration between tools for each borehole. Each logging run was recorded and stored digitally. During acquisition, data flow was monitored for quality and security in real time using tool-specific acquisition boxes (Advanced Logging Technology Ltd. [ALT] and Mount Sopris Instrument Company, Inc. [MSI] loggers) and software (ALTlogger and MSlog). Table T6 summarizes the acquisition system for each tool. Tools were raised at speeds that ranged from a minimum of 2 m/min for ABI40 image logs to a maximum of 10 m/min for induction conductivity logging. Recordings were taken both downhole and uphole. Because of the real-time data display, it was possible to stop uphole recovery once the casing/drill pipe was entered. At the end of each log run, the instrument was rezeroed at the top of the drill pipe. After each run, the WellCAD software package (Advanced Logic Technology) was used for visualization, QA/QC, processing, interpretation, and plotting the data. Except in specific cases, only uphole logs were used to process data. VSP setup for measurements is shown in Figure F16. The seismic source consisted of two air guns (20 in3 and 40 in3) (supplied by DWS International, Corpus Christi, Texas [USA]) linked to a high-pressure air reservoir of four bottles that were recharged by a high-pressure air compressor. The air gun was hung from the ship starboard crane at a depth of ~2 m below the sea surface. Compressed air (~1500 psi) and electrical firing signals were provided to the air gun via an umbilical that hung from the ship deck. A number of recording sensors were employed in the measurements, including
This last detector is now commonly used in deep-sea seismic exploration studies. The borehole geophone was lowered on a four-conductor armored wireline; unfortunately, only the vertical component could be measured, as the wireline carried insufficient conductors. Note that geophones detect particle velocities and hydrophones detect wave pressure. Each VSP shot was recorded and stored digitally. The signal received by the geophone was digitized using a Geometrics "Geode" and recorded using Geometrics Seismodule Controller software. Raw data were saved in SEG2 format. Table T8 provides details of the surveys carried out in terms of date, interval spacing, and the segments obtained with the downhole geophone in open hole or within the steel drill string. Logging tool description and acquisition parametersTechnical schemes on individual tools are shown in Figure F17. Additional information can be found on the manufacturers' Web sites (www.antares-geo.de/, www.mountsopris.com/, www.geovista.co.uk/, and www.alt.lu/). Detailed information on their geological applications can also found in Schlumberger (1989) and Serra (1984, 1986). Spectral natural gamma probeUnlike other slimline instruments that record total gamma ray emissions, the ASGR (ANTARES Datensysteme GmbH) allows identification of individual elements that emit gamma rays. Naturally occurring radioactive elements such as K, U, and Th emit gamma rays with a characteristic energy. K decays into two stable isotopes (Ar and Ca), and a characteristic energy of 1.46 MeV is released. U and Th decay into unstable daughter elements that produce characteristic energy at 1.76 and 2.62 MeV, respectively. The most prominent gamma rays in the U series originate from decay of 214Bi and in the Th series from decay of 208Tl. It is thus possible to compute the quantity (concentration) of parent 238U and 232Th in the decay series by counting gamma rays from 214Bi and 208Tl, respectively. The ASGR detector for gamma rays is a bismuth germanate (BGO) scintillation crystal optically coupled to a photomultiplier. The BGO detector has an absorption potential eight times greater than the more classic NaI crystal. As most of the spectral discrimination is performed in the high-energy range, only instruments equipped with BGO detectors prove to be reliable in the slimline tool domain. The instrument was master-calibrated by the manufacturer. On site, the stability of the sensor was checked using a volume of pure potassium. As the probe moves up the borehole, gamma rays are sorted according to their emitted energy spectrum (the tool has 512 reference spectra in memory) and the number of counts in each of three preselected energy intervals are recorded. These intervals are centered on the peak values of 40K, 214Bi, and 208Tl. Tool output comprises K, U, and Th in Becquerel per kilogram and total gamma ray counts in counts per second (cps). The vertical resolution of the tool is ~15 cm. The downhole measurement interval was 0.1 m. Induction resistivity probeThe ALT DIL45 measures electrical conductivity of the geological formation. Variations in electrical conductivity correspond primarily to variations in lithology (composition and texture), formation porosity and saturation, and interstitial fluid properties (salinity). An oscillator sends an alternating current of constant amplitude and frequency through an emitting coil. This current generates an alternating electromagnetic field that induces Foucault currents in the formation. These Foucault currents are proportional to the formation conductivity and generate their own electromagnetic fields. When passing through a receiving coil (solenoid), these secondary magnetic fields induce electromotive forces that are proportional to the flow running through the coil. The output of the DIL45 tool comprises two conductivity logs: ILM (0.57 m) and ILD (0.83 m). Measured conductivity can be converted into electrical resistivity. The instrument was calibrated against a Wenner array in a reference hole. The measurement window ranges from 5 to 5000 mmho, with a resolution of 0.5 mmho. The downhole measurement interval was 0.05 m. Magnetic susceptibilityThe Geovista EM51 probe measures formation electrical conductivity and formation magnetic susceptibility using electromagnetic induction. This measurement relies on a principle similar to the DIL45 induction probe, whereby a current is induced by an oscillating magnetic field in the probe within a toroidal zone of formation at some radial distance from the probe coils. The oscillating current produces a secondary field that is detected by the receiver coils. The "in-phase" signal is a measure of susceptibility in formations with magnetic properties. The probe is most effective in high-conductivity geological formations and low-conductivity borehole fluid, including air. The instrument was calibrated using a set of two coils for conductivity measurements and another set of two coils for magnetic susceptibility measurements. The measurement window ranges from 10–5 to 2 SI. The downhole measurement spacing interval selected was 0.05 m. Expedition 313 was the first time this specific slimline probe was used during IODP. Acoustic borehole televiewerThe ALT ABI40 produces millimeter scale, high-resolution images of the borehole surface and can be directly used for sedimentological and structural interpretation. A voltage is applied in a piezoelectric ceramic to produce acoustic waves (1.2 MHz) at regular intervals. On hitting a focalizing mirror, these waves are deflected perpendicularly to the wave source and toward the borehole wall. The focal point corresponds to the point of maximum energy, giving an ~4 mm diameter footprint on the wall in a 100 mm diameter borehole. To obtain a 360° image, the mirror pivots on a central axis. The resolution is user defined and depends on the number of measurements made in one rotation and the rate at which the tool is raised up the borehole while measurements are made. The highest quality images are obtained with a vertical sampling of 2 mm and a radial sampling with 288 shots per circumference. During Expedition 313, ABI40 acoustic images were acquired with various resolutions ranging from 72 samples × 4 mm to 288 samples × 2 mm, which results in a pixel size of <5 (horizontal) × 4 (vertical) mm to <4 (horizontal) × 2 (vertical) mm. The ABI40 produces two distinct images of the borehole wall. First, an acoustic impedance image (from the contrast between the borehole fluid and wall) is derived from the reflected wave amplitude obtained around the hole. The amplitude ratio between the emitted wave and the reflected wave provides information on the formation's capacity of absorption (low returned amplitude corresponds to a high capacity of absorption [i.e., soft formation]). Second, a traveltime image is derived from the reflected wave traveltime from the ceramic transducer to the borehole wall and back. The traveltime is directly proportional to the distance between the borehole wall and the probe. The tool is equipped with magnetometers and accelerometers for tool orientation with respect to north. The precision of the measured inclination is 0.5°, and the precision of the measured azimuth is 1.5°. For each of the two images, a set of false colors is assigned. From the measurements, a virtual image of the borehole wall depth is produced. This image is displayed as an unfolded representation of the 360° view. If mudcake is present, the images of those intervals would be expected to have a homogeneous appearance. As core features can be correlated to the images in many places, it is assumed that mudcake is absent over these intervals. Full waveform sonic probeThe Mount Sopris 2PSA-1000 sonic probe was used to measure compressional wave velocities of the formation. When bulk density is known (from core), elastic properties (bulk and shear moduli) and an estimate of porosity can be derived from sonic measurements. This downhole instrument is composed of an acoustic transmitter and two receivers. The transmitter transmits an acoustic signal that propagates through the borehole fluid to the rock interface, where some of the energy is critically refracted along the borehole wall. As a result of wavefront spreading (Huygens principle), some of the refracted energy is transmitted back into the borehole adjacent to a receiver. Each receiver picks up the signal, amplifies it, and digitizes it. Recorded waveforms are then examined, and wave arrival times are manually or automatically selected (picked). Arrival times are the transit times of the acoustic energy. By measuring the acoustic transit time and knowing the distance between the two receivers (1 ft), the velocity in the fluid, and the borehole diameter, the sonic velocity of the rock is calculated. Consequently, the interval velocity value is calculated at each sampling point. In the specific configuration used, P-waves (15 kHz monopole survey) were recorded. Calibration of the tool was performed either in water (1500 m/s for P-wave; freshwater at 28°C) or in a steel pipe (5440 m/s) while running downhole. The downhole measurement interval was 0.1 m. Vertical seismic profilingA VSP test consists of sending a seismic signal from a surface source down to a geophone at depth in the borehole. The downhole geophone converts wave particle motions into an electrical signal that is transferred to the surface via the wireline to be digitally recorded. The wave field is recorded at a number of positions along the borehole. The one-way traveltime from surface to depth is obtained by picking the times of arrival on the observed records; these are often called check shot times. Check shot times are used to assign true depths to the events seen in a seismic reflection profile. VSP profile times can be erroneous for a variety of reasons, including error-prone velocity analysis that may be contaminated by noise and seismic multiples, as well as seismic anisotropy that normally results in any layered sedimentary environment. Further processing of the reflection profile allows for the construction of a zero-offset seismic trace that may be further compared to the seismic reflection profile to assist in assigning depths to the reflectors and to allow for the removal of multiple reverberations that can sometimes appear as events within the surface seismic profile. During VSP operations, each VSP shot was recorded on seven channels (Table T9) to measure the vertical component on the downhole geophone, the vertical and two horizontal components on a seafloor geophone, a seafloor hydrophone (within the housing of the seafloor geophone), a suspended hydrophone, and the electrical signal caused by the movement of the air gun shuttle. Of the two air guns (20 and 40 in3), only the 20 in3 was fired during each shot. The record lengths were 2.5 s, covering the time period from –0.5 to 2 s. The sampling rate was 250 µs. The vertical measurement spacing in Holes M0027A, M0028A, and M0029A was, respectively, 0.91, 1.83, and 3.05 m. Depending on noise conditions (primarily weather dependent) and signal strength, up to five individual records were obtained at each depth. Data processingSlimline logging data were processed using the WellCAD software package. The processing procedure is described below for the logs: natural gamma radioactivity (ASGR), induction (DIL45), and magnetic susceptibility (EM51); image data (ABI40); and sonic data (2PSA). Depth adjustmentsOne main processing task involved evaluating the depth of each log run and referencing the data to the rig floor and seafloor. While deploying all the tools separately in the same section, a fixed zero depth position (loggers zero) was maintained at the top of the drill pipe. Typical reasons for depth corrections include ship heave and tides, but as the logging for Expedition 313 was performed from a platform resting directly on the seafloor, no such corrections were necessary. Using WellCAD, the original logs were depth adjusted to the rig floor (in meters wireline log depth below rig floor [m WRF]). This adjustment included several corrections:
Logs were subsequently shifted to the seafloor (in meters wireline log depth below seafloor [m WSF]) using the drillers depth to seafloor. Slight discrepancies (<0.5 m) may exist between the seafloor depths determined from the downhole logs (ASGR) and those determined by the drillers ("bottom felt" depth). When necessary, logs have been matched manually by the log analyst to a reference log using distinctive peaks. In such cases, gamma ray logs through pipe (or occasionally induction logs) are taken as reference logs (continuous). Generally, depth discrepancies between logs are <1.5 m. Matched log depths are referenced to seafloor and are referred to as meters wireline log matched depth below seafloor (m WMSF). The EM51 log from Hole M0028A and the EM51 and 2PSA logs from Hole M0029A have been depth matched. Invalid dataInvalid log values have been replaced by a null value of –999.25. Invalid log values removed can include the following:
Environmental correctionsEnvironmental corrections are designed to remove any effect from the borehole (size, roughness, temperature, and tool standoff) or the drilling fluids that may partially mask or disrupt the log response from the formation. Here, no postacquisition corrections of this type were applied. Log mergesWhere applicable for the processed logs, overlapping log runs have been merged to give one continuous log. In the overlapping regions, data have been checked by the log analyst to make sure distinctive peaks and troughs correlate. For all except three logs (EM51 log from Hole M0028A and EM51 and 2PSA logs from Holes M0029A), no match adjustments were required before merging. Quality controlData quality is assessed in terms of reasonable values for the logged formation, repeatability between different passes of the same tool, and correspondence between logs affected by the same formation property (e.g., the conductivity log should show inverse response features to the gamma log). Repeatability between data acquired on down and up acquisition of logs was checked by the log analyst. Considering the challenging borehole conditions, the overall quality of the downhole logging data is very good. The quality of the ASGR data is good even when logging through pipe. Negative values indicative of incorrect statistics were <12% of the total record. Gamma ray logs recorded through drill pipe should, however, be used only qualitatively because of attenuation of the incoming signal. Sections were acquired in open-hole conditions in two intervals of Hole M0027A and a small interval in Hole M0029A for through-pipe data calibration. ASGR data clearly correlate/anticorrelate with magnetic susceptibility and conductivity. ASGR data also correlate very well with NGR measurements on the unsplit cores (see "Physical properties"), allowing, when necessary, precise core depth positioning with respect to log depth. The quality of EM51 data is good. Magnetic susceptibility logs compare well with MSCL magnetic susceptibility data acquired on core sections. High values of magnetic susceptibility tie with levels where either glauconite or other possible magnetic minerals have been described by the sedimentologists. The quality of DIL45 data is good. Induction conductivity is a deep-investigation measurement and is least sensitive to borehole conditions. Conductivity log values are within the expected range. Logs correlate well with the EM51 logs and anticorrelate with the MSCL resistivity data acquired on core sections. The quality of ABI40 data varies from one hole to another. The resolution selected during image acquisition also varied between 72 and 288 shots per circumference. As the distance to the borehole wall also affects the quality of this imaging log, the probe should be run with centralizers. However, for technical reasons related to the size of the coring bit diameter relative to the diameter of the open hole below (larger), centralization was not possible in most cases, resulting in medium quality images containing dark lines oriented at 180°. For this reason, image data in these sections should be treated with great care. Only in the middle section of Hole M0028A were acoustic images acquired at high resolution and with centralizers, resulting in high-quality and high-resolution images. The presence of mudcake in some places in the hole possibly affects image quality by partly or totally masking borehole geology. The quality of the 2PSA data is good. Measurements of compressional wave velocity are highly dependent on borehole conditions. Larger cavities cause the induced wave to scatter, and acoustic energy is lost more rapidly. The picked P-wave arrivals show features very similar to the EM51 logs and the MSCL density signal from core sections. ABI40 image data processingSeveral corrections have been applied to the ABI40 data using the WellCAD software. Bad trace interpolation"No Data" or "NULL" traces in the images have been removed using an algorithm that replaces bad traces with the closest trace containing valid data points. Image centralizationWhere possible, traveltime images have been corrected for decentralization effects of the probe within the borehole. Assuming that the decentralization effect on the data can be approximately described through a sinusoid, the "Centralize Process" in the WellCAD software was used to remove this trend and correct the input data according to a best-fitted sinusoid. Caliper calculationABI40 images of the borehole wall were apparently unaffected by borehole fluid quality. Acoustic caliper values were obtained by assuming an acoustic wave velocity of 1530 m/s through borehole fluid. The calculation was checked using known internal pipe diameter. Oriented presentationImages available in any format have been oriented with respect to magnetic north. In some borehole intervals, several passes have been oriented and merged on the same presentation. Images provided in PDF format have been depth corrected to the seafloor (m WSF). Images are displayed as an unwrapped borehole cylinder. A dipping plane in the borehole appears as a sinusoid on the image, with the amplitude of this sinusoid proportional to the dip of the plane. Using the image module of WellCAD, a "static" normalization of the images has also been applied to enhance structure visualization. In the static normalizations, the amplitude and traveltime ranges of the entire interval of data have been computed and partitioned into 256 color levels. Because of the decentralization tool effect, a change in image color is observed at the contact between successive logged sections. Data delivery
ABI40 image files in PDF format ( ASCII raw data have been oriented with respect to magnetic north and have been depth corrected to the rig floor (m WRF). In some cases, several passes have been oriented and merged together on the same presentation. 2PSA acoustic dataThe 2PSA tool was run at a frequency of 15 kHz, and resultant logs can be used to calculate compressional wave velocities. Data were processed in the WellCAD logging package. For processing purposes, data were filtered (frequency filter) in such a way that only the energy around the induced frequency was analyzed. Waveform picking was done manually in the WellCAD logging package to ensure good quality data. Where no clear arrivals in the waveform were present in two receivers, a null value was entered in the database. Time picks were saved, and acoustic velocities were calculated. The precision of acoustic traveltime measurements is ~5%. Vertical seismic profilingData flow was monitored for quality in real time, and data quality was recorded on paper to be entered into an Excel spreadsheet postsurvey, along with shot and receiver coordinates (datum WGS84). Where available, borehole tilt and azimuth measured by borehole logging are used to calculate the horizontal location of the downhole receiver. Where this information is not available, the borehole is assumed to be vertical. The Excel quality and coordinate measurements were entered into the VSP data files using Mathworks MATLAB and Seismic Imaging Software VISTA software. Data were then processed using MATLAB and VISTA. Processing involved removing poor-quality shots and stacking the remaining shots at a given depth to improve the signal to noise ratio. 3-D offset information (Fig. F16) was calculated and entered. Where possible, first break arrival times were picked for the downhole data, the vertical seafloor geophone, the suspended hydrophone, and the air gun shuttle signal. Based on a combination of data from the vertical seafloor geophone, suspended hydrophone, and air gun shuttle signal, static corrections were calculated and applied to account for any time shift caused by the air gun firing past time zero. The wave form and arrival times of these data were assessed for each survey to determine the best technique of calculating static corrections. In most cases, obtaining records in open hole was preferred, as this eliminated any noise induced by the drill string. Indeed, "ringing" of the drill string as the seismic waves passed reduced data quality substantially unless the drill string was in good contact (i.e., stuck) with the borehole. This problem was particularly severe for Hole M0029A, where borehole stability conditions did not allow for any open-hole measurements and drill string reverberations were strong. For Holes M0027A and M0028A, no processing was conducted prior to first break picking of the seismic arrival through the sediment. For Hole M0029A, it was necessary to apply F-K and bandpass filters prior to picking seismic arrivals through the sediment for survey "a." An F-K filter and automatic gain control was applied prior to picking sediment arrival times on survey "b." The F-K filters were designed to remove all upgoing waves and downgoing waves traveling with velocities over ~3000 m/s. These filters were not saved in order to preserve the information within the data for further research. Interference from seismic arrivals through steel meant that, despite processing, the only consistent first break from survey "a" was a peak, whereas in survey "b" from Hole M0029A and on all surveys from Holes M0027A and M0028A, where this interference did not exist, the first break picked was a trough. Times derived from survey "a" are therefore likely to be time shifted by a half-wavelength. For Hole M0027A, the seafloor geophone in conjunction with the air gun shuttle signal was used to determine traveltime from the source to the seafloor. The seafloor geophone peak was picked, and the times the air gun shuttle signal amplitude began to be clipped were used. Variations in air gun elevation and firing time differences between the two air guns are accounted for using this time, and data are time shifted within the stacked data to reflect the source firing at time zero and at a constant elevation. The time shift reflects the source firing at time zero. The stacked data files are all time shifted, with the static shift used saved within the stacked data files. Because of the nature of SEG-Y formatting, the time used for the static shift has been truncated to be an integer; however, static shifts used reflected the 250 µs sampling rate and were calculated and applied with this 250 µs accuracy. For Hole M0028A, the seafloor geophone was used in conjunction with the air gun firing time data from Hole M0029A measurements to correct for variations in air gun firing times, and the stacked downhole data are shifted to reflect the air gun firing at time zero. For Hole M0028A uniquely, the seafloor geophone first arrival did not have a clear peak. Here, the first breaks for the seafloor geophone were picked when the signal came up from zero amplitude. The first breaks (peaks) of the hydrophone suspended 20 m above the seafloor geophone are assessed in conjunction with the seafloor arrivals to confirm a relatively vertical ray path for the seismic wave between the source and the seafloor receiver. The static shifts are applied and saved within the stacked data file, as above. For Hole M0029A, the suspended hydrophone and its known distance from the air gun source were used to determine the firing time of the air gun, and from this the stacked data were shifted to reflect the air gun firing at time zero. Peaks observed on the hydrophone records were picked for this purpose. First breaks (peaks) were also picked for the seafloor geophone and used to determine the traveltime through the water column. The remotely operated vehicle (ROV) was dispatched for both surveys on this hole to assess the location of the seafloor hydro-/geophone. In both cases, there was estimated to be a maximum 5 m horizontal offset between the top of the cable where it came onto the ship and the bottom of the cable where it connects to the seafloor hydro-/geophone. Despite the considerable effort and time devoted to the collection and processing of VSP data, their reliability as check shot information was limited in the time available. Further analysis may determine that these provide time-depth conversions superior to those derived from stacking velocities (see discussion in "Stratigraphic correlation"), but for this volume we have chosen the latter. Particular issues of concern regarding VSP data forcing this decision include the following:
Data delivery
Processed and raw data are available in the USIO online log database. Processed standard data are available in ASCII format. Processed depths are referred to as either WSF and/or WMSF scales. ABI40 image files are also available in PDF format ( |
|||
and 
scales) are available online in the U.S. Implementing Organization (USIO) log database. ABI40 raw and processed data are also available in ASCII format. ASCII-processed data are depth corrected to the seafloor (m WSF); bad traces have been interpolated, and images have been centralized. Images have not been normalized.