-
Defence R&D Canada – Atlantic
DEFENCE DÉFENSE&
Preliminary Results from the Sable IslandGeomagnetic Coherence
Experiment
J. Bradley Nelson
Technical Memorandum
DRDC Atlantic TM 2006-004
January 2006
Copy No.________
Defence Research andDevelopment Canada
Recherche et développementpour la défense Canada
-
This page intentionally left blank.
-
Preliminary Results from the Sable Island Geomagnetic Coherence
Experiment
J. Bradley Nelson
Defence R&D Canada – Atlantic Technical Memorandum DRDC
Atlantic TM 2006-004 January 2006
-
Abstract Previous experiments suggested that the geomagnetic
field is coherent over distances of hundreds of kilometres at
high-altitude over the ocean, but not at low-altitude where most
magnetic anomaly detection (MAD) flights are conducted. The reason
for this is not understood but computer simulations suggest that
excess magnetic noise due to ocean dynamics or variations in the
seafloor conductivity along the flight path could cause the effect.
Detailed models are required to predict the magnitude of the effect
in specific areas. Flights were conducted near Sable Island in
August 2005 to gather a data set on which detailed computer
simulations can be performed. The flights were conducted in areas
where bathymetric and conductivity data exists, the magnetic
properties of the seafloor are well known, a basestation could be
set up relatively close by, and ocean dynamic processes can be
modelled. This report describes the processing steps required to
obtain the data set, and the preliminary results which will be
compared to NRL Stennis computer simulations of ocean dynamic noise
and conductivity anomaly effects.
Résumé Des expériences antérieures ont suggéré que le champ
géomagnétique est cohérent sur des distances de centaines de
kilomètres à haute altitude au-dessus de l’océan, mais non aux
faibles altitudes auxquelles s’effectuent la plupart des vols de
détection des anomalies magnétiques (MAD). Les causes du phénomène
ne sont pas comprises, mais des simulations sur ordinateur
suggèrent qu’il pourrait être engendré par un bruit magnétique en
excès attribuable à la dynamique des océans ou des variations de la
conductivité du fond marin le long des lignes de vol. Des modèles
détaillés sont nécessaires pour prévoir l’ordre de grandeur de cet
effet dans des régions spécifiques. Des vols ont été effectués près
de l’île de Sable en août 2005 afin de recueillir un jeu de données
avec lequel mener des simulations par ordinateur détaillées. Les
vols ont été exécutés dans des régions où une station de base
pouvait être installée relativement proche, pour lesquelles on
dispose de données sur la bathymétrie et la conductivité, dont on
connaît bien les propriétés magnétiques du fond marin et pour
lesquelles on pouvait modéliser les processus dynamiques de
l’océan. Dans le présent rapport on décrit les étapes du traitement
nécessaire pour l’obtention du jeu de données et les résultats
préliminaires qui seront comparés aux simulations par ordinateur du
bruit engendré par la dynamique de l’océan et des effets des
anomalies de conductivité menées au centre Stennis du NRL.
DRDC Atlantic TM 2006-004 i
-
This page intentionally left blank.
DRDC Atlantic TM 2006-004 ii
-
Executive summary Background: Next-generation MAD systems are
will employ geomagnetic noise reduction using reference sensors.
Previous experiments suggested that the geomagnetic field is
coherent over hundreds of kilometres at high-altitude over the
ocean, but not at low-altitude where most MAD flights are
conducted. The reason for this is not understood but computer
simulations suggest that magnetic noise due to ocean dynamics or
variations in the seafloor conductivity along the flight path could
cause the effect. Detailed models are required to predict the
magnitude of the effect in specific areas. Flights were conducted
near Sable Island in August 2005 to gather a data set on which
detailed computer simulations can be performed. The flights were
conducted in areas where bathymetric and conductivity data exists,
the magnetic properties of the seafloor are well known, a
basestation could be set up relatively close by, and ocean dynamic
processes can be modelled. Results: The lack of coherence between
magnetic signals measured at a basestation and in a low-altitude
aircraft was not immediately apparent in this experiment. However,
the geomagnetic signals were not very large for the low-altitude
portions of this experiment so it is difficult to say conclusively
whether the poor coherence was caused by excess noise or conductive
effects. However, the low-frequency portion of the airborne data
looks very much like the Sable Basestation total-field in the cases
where a significant geomagnetic signal was present. When the
geomagnetic signals were large, the coherence between the Sable
Island basestation and the airborne measurements was often >0.9,
no matter what the altitude was. When the geomagnetic signals were
small, and the coherence was smaller, it was probably due to excess
noise from imperfect geological noise cancellation or ocean
dynamics because conductive effects would not yield better
coherence when the geomagnetic signals were larger. There may be
some small phase and amplitude mismatches between the basestation
and airborne data collected over the Gully. These could be due to
imperfect geological noise removal, ocean dynamics, or, although
less likely for the reasons described above, conductive effects.
Forward modelling by NRL should determine if the latter two effects
are of the correct magnitude to account for these results.
Significance: When the geomagnetic noise and coherence are large,
some of the geomagnetic noise can be cancelled. The remaining noise
level is ~ 50-100 pT/√Hz at 0.05 Hz. In cases where the geomagnetic
noise is smaller than the other magnetic noise, there is little
cancellation. Regardless of the source of the excess noise, if it
cannot be modelled and removed, then it is unlikely that either
magnetic basestation or using a high-altitude magnetometer-equipped
aircraft as a geomagnetic reference station will result in
significant increases in MAD detection ranges. Future Work:
Conductivity and ocean dynamics effects will be modelled by NRL
Stennis and compared to the results found here. Future experiments
will address the issues of imperfect geological noise reduction and
additional noise due to ocean dynamics.
Nelson, JB. 2006. Preliminary Results from the Sable Island
Geomagnetic Coherence Experiment. DRDC Atlantic TM 2006-004.
Defence R&D Canada – Atlantic.
DRDC Atlantic TM 2006-004 iii
-
Sommaire Contexte Les systèmes de détection des anomalies
magnétiques (MAD) de la prochaine génération devraient comporter un
type ou un autre de capteur de référence pour la réduction du bruit
géomagnétique. Des expériences antérieures ont montré que le champ
géomagnétique est cohérent sur des distances de centaines de
kilomètres à haute altitude au-dessus de l’océan, mais non aux
faibles altitudes auxquelles s’effectuent la plupart des vols de
détection des anomalies magnétiques (MAD). Les causes du phénomène
ne sont pas comprises, mais des simulations sur ordinateur
suggèrent qu’il pourrait être engendré par un bruit magnétique en
excès attribuable à la dynamique des océans ou des variations de la
conductivité du fond marin le long des lignes de vol. Des modèles
détaillés sont nécessaires pour prévoir l’ordre de grandeur de cet
effet dans des régions spécifiques. Des vols ont été effectués près
de l’île de Sable en août 2005 afin de recueillir un jeu de données
avec lequel mener des simulations par ordinateur détaillées. Les
vols ont été exécutés dans des régions où une station de base
pouvait être installée relativement proche, pour lesquelles on
dispose de données sur la bathymétrie et la conductivité, dont on
connaît bien les propriétés magnétiques du fond marin et pour
lesquelles on pouvait modéliser les processus dynamiques de
l’océan. Résultats L’absence de cohérence entre les signaux
magnétiques mesurées à l’emplacement d’une station de base et à
bord d’un aéronef volant à basse altitude n’était pas immédiatement
apparente lors de cette expérience et il est ainsi difficile de
dire sans l’ombre d’un doute que la mauvaise cohérence était
attribuable à un bruit en excès plutôt qu’aux effets de
conductivité. Cependant, dans les basses fréquences, les données
aériennes ressemblent beaucoup à celles sur le champ total mesuré à
la station de base de l’île de Sable où est mesuré un important
signal géomagnétique. Lorsque l’intensité des signaux
géomagnétiques était grande, la cohérence entre les mesures
effectuées par la station de base à l’île de Sable et celles
effectuées à bord de l’aéronef était souvent > 0,9, quelle que
soit l’altitude à laquelle les données étaient acquises. Lorsque
l’intensité des signaux géomagnétiques était faible, et que la
cohérence était moindre, cela était probablement attribuable à un
bruit en excès résultant d’une annulation imparfaite du bruit
géologique ou du bruit causé parla dynamique de l’océan parce que
les effets de conductivité n’engendreraient pas une meilleure
cohérence pour des signaux géomagnétiques de plus grande intensité.
Il peut exister de faibles écarts de correspondance pour la phase
et l’amplitude entre les données de la station de base et les
données aériennes recueillies au-dessus du Goulet. Ils pourraient
être attribuables à une suppression imparfaite du bruit géologique
et du bruit engendré par la dynamique de l’océan ou, ce qui est
moins probable pour les raisons exposées ci-haut, aux effets de
conductivité. Une modélisation prédictive par le NRL devrait
permettre de déterminer si ces deux derniers effets sont du bon
ordre de grandeur pour expliquer ces résultats. Importance Quelle
que soit la source du bruit en excès, s’il ne peut être modélisé et
supprimé il est peu vraisemblable que l’utilisation des données
d’une station magnétique de base ou d’un aéronef volant à haute
altitude avec un magnétomètre permettront une réduction adéquate du
bruit géomagnétique.
DRDC Atlantic TM 2006-004 iv
-
Travaux futurs Les effets de conductivité et de la dynamique des
océans seront modélisés par le centre Stennis du NRL et comparés
aux résultats ici exposés.
Nelson, JB. 2006. Résultats préliminaires de l’expérience sur la
cohérence du champ géomagnétique à l’île de Sable. RDDC Atlantique
TM 2006-004. R et D pour la défense Canada – Atlantique.
DRDC Atlantic TM 2006-004 v
-
Table of contents
Abstract………………………………………………………………………………………......... i Executive
summary……………………………………………………………………...……...... iii
Sommaire………………………………………………….…………………………….……...... iv Table of
contents…………………………………………………………………...………..…..... vi List of
figures………………………………………………………………...……………..…..... vii 1.
Introduction…………………………………………………………………………..………... 1 2. The
Experiment…………………....................................................................…......................
2 2.1 Magnetic basestations………………………………………………………………….. 2 2.2
Flights near Sable Island……………………………………………………………….. 3 2.3 Flights
near the Gully…………………………………………………………………... 5 3. Data
Pre-Processing……...........................................................................................................
6 4. Basestation Analysis…………………………………………………………………............... 7
4.1 Sable Island magnetic field components…………..………………………………….... 7
4.2 Correlation between Sable Island and Greenwood
basestations…………..………….. 10 5. Separating Geological and
Geomagnetic Noise..........………………………...……………... 14 5.1 Read data,
define desired track, correct for offset from desired
track………………... 14 5.2 Basestation
correction……………………………………………………………….... 16 5.3 Combining several
lines……………………………………………………………..... 17 5.4 Re-sampling and
reverse-gradient correction back to the original sampling
position... 18 5.5 High-pass filter the original TF and geological
noise estimate……………………..... 18 5.6 Re-compensate the filtered
residual…………………………………………………... 19 5.7 Plotting the
results………………………………………………….…………………. 19 6.
Results…………………………………………………………………………………….….. 21 6.1 Sable Island
North/South lines………………………………………………………... 21 6.2 Sable Island
East/West lines………………………………………………………….. 31 6.3 Gully North/South
lines……………………………………………………………..... 39 6.4 Gully East/West
lines………………………………………………………………..... 46 7.
Conclusions…………………………………………………………………………………... 53 8. Future
Work.....................……………………..........………………………...…………….... 55
9.
References........................……………………..........………………………...……………....
56 10. Distribution list……………………………………………………………………………..... 57
DRDC Atlantic TM 2006-004 vi
-
List of figures Figure 1. Locations for the
experiment…………………….………………………………………………………….2 Figure 2. Flight lines
near Sable Island………………..……………………………………………………………..4 Figure 3. Sable
Island photographed from the Convair 580 while flying on one of the
North/South flight lines……………………………………………………………………………………..…………..…4
Figure 4. Actual lines flown near the Gully at altitudes of 1000,
2000, 5000, and 10,000 ft (~300,
600, 1500, and 3000 m).…………………………………………………………………………..………..5
Figure 5. Flow chart describing the data pre-processing
steps.……………………………………….………….6 Figure 6. Geomagnetic field components
(Black = North, Blue = West, Red = Up) recorded at Sable
Island on August 6, 2005……………………………………………………………………………..........7
Figure 7. PSD of geomagnetic field components (Black = North, Blue
= West, Red = Up) recorded
at Sable Island on August 6………………………………………………………………………………..8
Figure 8. TFS (Black), TFvector (Blue), and TFvector’ (Red)
measured at Sable Island on August 6,
2005. DC values have been removed and TFS is offset by 2 nT for
display purposes.……............9 Figure 9. PSD of TFS (Black),
TFvector (Blue), TFS-TFvector (Green, just barely visible beneath
the Red
trace), and TFS-TFvector’ (Red) of the time series data shown in
Figure 8…..……………...……….10 Figure 10. Total-field measured at
Greenwood (Black) and Sable Island (Blue) on August 10.
The DC values have been removed for display
purposes.………………………………………..…11 Figure 11. PSD’s of time series data
shown in Figure 10: Sable Island (Blue) and Greenwood
(Black).………………………………………………………………………………………………........12 Figure 12.
Coherence of the Sable Island and Greenwood basestation TF data on
August 10.
Data were high-pass filtered with a 2nd-order high-pass
Butterworth digital filter with a 3-dB point at 0.004 Hz
……………………………………………………………………………....13
Figure 13. Flow chart of geological noise modelling and removal.
Detailed description of various
processes are given in the Sections shown in
red.………………………………..………………...15 Figure 14a. Comparison of the raw
total-field (TF) and the gradient- and basestation-corrected
total-field (TF”) along the three North/South lines near Sable
Island at each altitude (Black vs,
Blue)………………………………………………………………………………………...21
Figure 14b. Horizontal gradient correction applied to each
North/South line at 500’ near Sable
Island, based on IGRF gradients. Blue=ΔNorth x GNorth; Red=ΔEast
x GEast…………………..22 Figure 14c. Horizontal gradient correction
applied to each North/South line at 1000’ near Sable
Island, based on IGRF gradients. Blue=ΔNorth x GNorth; Red=ΔEast
x GEast………………..…22 Figure 14d. Horizontal gradient correction
applied to each North/South line at 2000’ near Sable
Island, based on IGRF gradients. Blue=ΔNorth x GNorth; Red=ΔEast
x GEast…………………..22 Figure 14e. Horizontal gradient correction
applied to each North/South line at 5000’ near Sable
Island, based on IGRF gradients. Blue=ΔNorth x GNorth; Red=ΔEast
x GEast…………………..22 Figure 14f. Comparison of geology estimates
along the three North/South lines near Sable Island
at each altitude (Black, Blue, Green) vs the average (Red).
TFgeo is set to the average………23 Figure 14g. Effect of extra
compensation along the 500’ North/South lines near Sable
Island:
DRDC Atlantic TM 2006-004 vii
-
ResidHP (Black) vs ResidHPC (Blue)…………………………………………………………….…24
Figure 14h. Effect of extra compensation along the 1000’
North/South lines near Sable Island:
ResidHP (Black) vs ResidHPC (Blue)………………………………………………………………..24
Figure 14i. Effect of extra compensation along the 2000’
North/South lines near Sable Island:
ResidHP (Black) vs ResidHPC (Blue)………………………………………………………………..24
Figure 14j. Effect of extra compensation along the 5000’
North/South lines near Sable Island:
ResidHP (Black) vs ResidHPC (Blue)………………………………………………………………..24
Figure 14k. Upper Trace: Comparison of signals measured along the
three North/South lines near Sable Island at 500’: TFCHP vs
TFestHP, (Black vs Red at the top of each upper trace); ResidHPC vs
TFSHP (Blue vs Green in the middle of each upper trace); DIF1 vs
DIF2 (Black vs Orange at bottom of each upper trace). Middle Trace:
Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……26 Figure 14l. Upper Trace: Comparison of
signals measured along the three North/South lines near Sable
Island at 1000’: TFCHP vs TFestHP, (Black vs Red at the top of each
upper trace); ResidHPC vs TFSHP (Blue vs Green in the middle of
each upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each
upper trace). Middle Trace: Coherence between ResidHPC and
TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……27 Figure 14m. Upper Trace: Comparison of
signals measured along the three North/South lines near Sable
Island at 2000’: TFCHP vs TFestHP, (Black vs Red at the top of each
upper trace); ResidHPC vs TFSHP (Blue vs Green in the middle of
each upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each
upper trace). Middle Trace: Coherence between ResidHPC and
TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……28 Figure 14n. Upper Trace: Comparison of
signals measured along the three North/South lines near Sable
Island at 5000’: TFCHP vs TFestHP, (Black vs Red at the top of each
upper trace); ResidHPC vs TFSHP (Blue vs Green in the middle of
each upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each
upper trace). Middle Trace: Coherence between ResidHPC and
TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……29 Figure 14o. DIF1 vs Latitude for
North/South lines near Sable Island………………………………………….30 Figure 15a.
Comparison of the raw total-field (TF) and the gradient- and
basestation-corrected
total-field (TF”) along the three East/West lines near Sable
Island at each altitude (Black vs,
Blue)…………………………………………………………………………………………31
Figure 15b. Comparison of geology estimates along the three
East/West lines near Sable Island at each altitude (Black, Blue,
Green) vs the average (Red). TFgeo is set to the average…………32
Figure 15c. Upper Trace: Comparison of signals measured along
the three East/West lines near Sable Island at 500’: TFCHP vs
TFestHP, (Black vs Red at the top of each upper trace); ResidHPC vs
TFSHP (Blue vs Green in the middle of each upper trace); DIF1 vs
DIF2 (Black vs Orange at bottom of each upper trace). Middle Trace:
Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……34 Figure 15d. Upper Trace: Comparison of
signals measured along the three East/West lines near Sable Island
at 1000’: TFCHP vs TFestHP, (Black vs Red at the top of each upper
trace); ResidHPC vs TFSHP (Blue vs Green in the middle of each
upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each upper
trace). Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……35
DRDC Atlantic TM 2006-004 viii
-
Figure 15e. Upper Trace: Comparison of signals measured along
the three East/West lines near Sable Island at 2000’: TFCHP vs
TFestHP, (Black vs Red at the top of each upper trace); ResidHPC vs
TFSHP (Blue vs Green in the middle of each upper trace); DIF1 vs
DIF2 (Black vs Orange at bottom of each upper trace). Middle Trace:
Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……36 Figure 15f. Upper Trace: Comparison of
signals measured along the three East/West lines near Sable Island
at 5000’: TFCHP vs TFestHP, (Black vs Red at the top of each upper
trace); ResidHPC vs TFSHP (Blue vs Green in the middle of each
upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each upper
trace). Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red).…...37 Figure 15g. DIF1 vs Longitude for
East/West lines near Sable Island………………………………………….38 Figure 16a.
Comparison of the raw total-field (TF) and the gradient- and
basestation-corrected
total-field (TF”) along the three North/South lines near the
Gully at each altitude (Black vs,
Blue)………………………………………………………………………………………………….39
Figure 16b. Comparison of geology estimates along the three
North/South lines near the Gully at each altitude (Black, Blue,
Green) vs the average (Red). TFgeo is set to the
average………………..40
Figure 16c. Upper Trace: Comparison of signals measured along
the three North/South lines near the Gully at 1000’: TFCHP vs
TFestHP, (Black vs Red at the top of each upper trace); ResidHPC vs
TFSHP (Blue vs Green in the middle of each upper trace); DIF1 vs
DIF2 (Black vs Orange at bottom of each upper trace). Middle Trace:
Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……41 Figure 16d. Upper Trace: Comparison of
signals measured along the three North/South lines near the Gully
at 2000’: TFCHP vs TFestHP, (Black vs Red at the top of each upper
trace); ResidHPC vs TFSHP (Blue vs Green in the middle of each
upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each upper
trace). Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……42
Figure 16e. Upper Trace: Comparison of signals measured along
the three North/South lines near the Gully at 5000’: TFCHP vs
TFestHP, (Black vs Red at the top of each upper trace); ResidHPC vs
TFSHP (Blue vs Green in the middle of each upper trace); DIF1 vs
DIF2 (Black vs Orange at bottom of each upper trace). Middle Trace:
Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……43 Figure 16f. Upper Trace: Comparison of
signals measured along the three North/South lines near the Gully
at 10,000’: TFCHP vs TFestHP, (Black vs Red at the top of each
upper trace); ResidHPC vs TFSHP (Blue vs Green in the middle of
each upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each
upper trace). Middle Trace: Coherence between ResidHPC and
TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……44 Figure 16g. DIF1 vs Latitude for
North/South lines near the Gully……………………………………………...45 Figure 17a.
Comparison of the raw total-field (TF) and the gradient- and
basestation-corrected
total-field (TF”) along the three East/West lines near the Gully
at each altitude (Black vs,
Blue)………………………………………………………………………………………………….46
Figure 17b. Comparison of geology estimates along the three
East/West lines near the Gully at each altitude (Black, Blue,
Green) vs the average (Red). TFgeo is set to the
average………………..47
Figure 17c. Upper Trace: Comparison of signals measured along
the three East/West lines near the Gully at 1000’: TFCHP vs
TFestHP, (Black vs Red at the top of each upper trace);
DRDC Atlantic TM 2006-004 ix
-
ResidHPC vs TFSHP (Blue vs Green in the middle of each upper
trace); DIF1 vs DIF2 (Black vs Orange at bottom of each upper
trace). Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……48 Figure 17d. Upper Trace: Comparison of
signals measured along the three East/West lines near the Gully at
2000’: TFCHP vs TFestHP, (Black vs Red at the top of each upper
trace); ResidHPC vs TFSHP (Blue vs Green in the middle of each
upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each upper
trace). Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……49 Figure 17e. Upper Trace: Comparison of
signals measured along the three East/West lines near the Gully at
5000’: TFCHP vs TFestHP, (Black vs Red at the top of each upper
trace); ResidHPC vs TFSHP (Blue vs Green in the middle of each
upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each upper
trace). Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red)……50 Figure 17f. Upper Trace: Comparison of
signals measured along the three East/West lines near the Gully at
10,000’: TFCHP vs TFestHP, (Black vs Red at the top of each upper
trace); ResidHPC vs TFSHP (Blue vs Green in the middle of each
upper trace); DIF1 vs DIF2 (Black vs Orange at bottom of each upper
trace). Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2 (red).…...51 Figure 17g. DIF1 vs Latitude for
East/West lines near the Gully………………………………………………...52
List of tables Table 1. Basestation
Information……………………………………………………………………………………...3
DRDC Atlantic TM 2006-004 x
-
1. Introduction Previous experiments (Refs 1, 2) have suggested
that the geomagnetic field is coherent over distances of hundreds
of kilometres at high-altitude over the ocean, but not at
low-altitude where most magnetic anomaly detection (MAD) flights
are conducted. The reason for this loss of coherence is not
understood, but two possibilities have been suggested:
1) excess magnetic noise due to ocean dynamics, or 2) variations
in the seafloor conductivity along the flight path will affect
the
local phase and amplitude of the geomagnetic variations.
Computer simulations performed by Will Avera at NRL-Stennis (Ref 3)
have suggested that both of these possibilities could be the source
of the loss of coherence, although detailed models are required to
predict the magnitude of the effect in specific areas. A series of
flight tests were conducted near Sable Island in August 2005 to
gather a data set on which detailed computer simulations could be
performed. The flights were conducted in areas where bathymetric
and conductivity data exists, the magnetic properties of the
seafloor are well known, a basestation could be set up relatively
close-by, and ocean dynamic processes can be modelled. This report
details the experimental setup, processing steps, and preliminary
results which NRL Stennis computer simulations will be compared
to.
DRDC Atlantic TM 2006-004 1
-
2. The Experiment The flights were conducted near Sable Island
and the Gully, as shown in Figure 1. Two magnetic ground stations
were set up – one on Sable Island and one at CFB Greenwood on
mainland Nova Scotia (separation ~ 400 km). A GPS basestation was
set up in Dartmouth Nova Scotia to provide differential GPS
positioning for the aircraft. The NRC Convair 580 research aircraft
was used to gather the flight data. The Convair’s aeromagnetic
instrumentation has been described elsewhere (Ref 4) so it will not
be repeated here. Sections 2.1-2.3 describe the basestation
instrumentation and the series of flights conducted.
CFB Greenwood
Figure 1. Locations for the experiment.
2.1 Magnetic basestations The magnetic basestation at CFB
Greenwood consisted of a Geometrics G822 Caesium-vapour total-field
magnetometer and associated electronics, a GT200 time interval
analyser board to convert the Larmor signal from the magnetometer
into a measurement of the magnetic field, a GPS receiver to tag the
raw measurements with UTC time, and a desktop computer running
Labview data acquisition software. Unfortunately the CFB Greenwood
basestation stopped collecting data after only 3 days, but this was
enough data to base conclusions regarding the coherence of the
geomagnetic field between the two basestations. The magnetic
basestation at Sable Island consisted of a Geometrics G822
Caesium-vapour total-field magnetometer and associated electronics,
a GT658 time interval analyser board to convert the Larmor signal
from the magnetometer into a measurement of the magnetic field, a
Billingsley DFM100GX vector magnetometer and NRC-built 5 Hz
low-pass anti-alias filter, a
DRDC Atlantic TM 2006-004 2
-
National Instruments PCI-4472 24-bit A/D board, a GPS receiver
to tag the raw measurements with UTC time, and a desktop computer
running Labview data acquisition software. Unfortunately the
anti-alias filter malfunctioned after 2 days and only the
total-field and the northward component of the geomagnetic field
were recorded during the flights. However, there were 2 days of
data where all three components of the geomagnetic field were
recorded and a correlation analysis was performed on these data.
Table 1 gives a detailed description of the sensors and conditions
at each basestation site.
Table 1. Basestation Information.
PARAMETER CFB GREENWOOD (Gr) SABLE ISLAND (S) Location 44°57.8’
N; 64°55.2’ W 43°55.9’ N; 60°00.4’ W
Magnetic Sensors Geometrics G822 (TF) Geometrics G822 (TF),
Billingsley DFM100GX (3
components, 24 bit) Sample Rate 8 8
Time-tags UTC from GPS UTC from GPS Distance from Man-Made
Noise
Sources 500 m, but located at airbase ~ 70 m from buildings
Soil 1 m above non-magnetic sediments
0.5 m above sandy non-magnetic soil
Distance from Ocean 20 km 500 m
2.2 Flights near Sable Island One flight consisted of a series
of North/South lines flown off the western tip of Sable Island. All
lines went from 43° 42’N to 44° 09’N along 60° 09’W and there were
3 lines at each of 500, 1000, 2000, and 5000 feet ASL
(approximately 150, 300, 600, and 1500 m). Each of these lines was
approximately 50 km in length the separation between the aircraft
and the Sable Island basestation varied from ~ 16 to 31 km. The
water depth was only 25-100 m along these flight lines near Sable
Island. The second flight consisted of a series of East/West lines
flown off the southern edge of Sable Island. All lines went from
59° 40’W to 60° 18’W along 43° 55.25’N, and there were 3 lines at
each of 500, 1000, 2000, and 5000 feet ASL (approximately 150, 300,
600, and 1500 m). Each of these lines was approximately 50 km in
length the separation between the aircraft and the Sable Island
basestation varied from ~ 1 to 25 km. In both cases the flight
lines were chosen to avoid the infrastructure associated with the
gas fields and drilling platforms that are situated near Sable
Island. Appendix A gives information on the platform locations and
the pipelines that run between them and the Nova Scotia mainland.
Figure 2 shows the actual flight lines flown superimposed on a map
of Sable Island. Figure 3 shows a photograph of Sable Island taken
from the NRC Convair during the experiment.
DRDC Atlantic TM 2006-004 3
-
Magnetic Basestation
Figure 2. Flight lines near Sable Island.
Magnetic Basestation
East
North
Figure 3. Sable Island photographed from the Convair 580 while
flying on one of the North/South flight lines.
DRDC Atlantic TM 2006-004 4
-
2.3 Flights near the Gully One flight consisted of a series of
roughly North/South lines flown down the axis of the Gully. All
lines went from 43° 41.1’N; 58° 51.6’W to 44° 07.5’N; 59° 07.1’W
and there were 3 lines at each of 1000, 2000, 5000 and 10,000 feet
ASL (approximately 300, 600, 1500, and 3000 m). Each of these 12
lines was approximately 50 km in length the separation between the
aircraft and the Sable Island basestation roughly was 80 km for all
the lines. The water depth varies from 100-1500 m along the flight
lines near the Gully. The second flight consisted of a series of
roughly East/West lines flown across the axis of the Gully. All
lines went from 44° 0.8’N; 58° 42.4’W to 43° 50.6’N; 59° 9.6’W, and
there were 3 lines at each of 1000, 2000, 5000, and 10000 feet ASL
(approximately 300, 600, 1500 and 3000 m). Each of these 12 lines
was approximately 50 km in length the separation between the
aircraft and the Sable Island basestation varied from ~ 68 to 104
km. Figure 4 shows the actual flight lines flown near the
Gully.
Figure 4. Actual lines flown near the Gully at altitudes of
1000,
2000, 5000, and 10,000 ft (~300, 600, 1500, and 3000 m).
DRDC Atlantic TM 2006-004 5
-
3. Data Pre-Processing The details of the DRDC Atlantic aircraft
noise removal process has been described elsewhere and will not be
repeated here. Because the basestation data were sampled at a
different rate (8 Hz vs 32 Hz for the aircraft), they had to be
re-sampled up to 32 Hz prior to any filtering to ensure that there
were no phase delays introduced between the airborne and
basestation data sets. Figure 5 shows the flow chart for the data
pre-processing. The aeromagnetic flights were performed in
conjunction with DRDC Val Cartier hyperspectral infra-red
experiments. Analysis of previous flight data had indicated that
there was no magnetic effect from running that equipment. However,
DRDC Val Cartier installed a cryo-cooler on the system prior to
these flight trials and the cryo-cooler introduced very small DC
steps into the magnetic data. These DC steps were < 0.1 nT in
height, lasted for only a fraction of a second, and occurred
roughly every 5 seconds. These DC steps had to be removed manually
as a reliable automated algorithm could not be found. In addition
to the cryo-cooler steps, DC steps were also introduced by VHF
radio transmissions. These were removed manually. The East-West
flight lines near Sable Island went directly over 4 wrecks that
produced significant signals that caused problems for the next
stage of processing. These signatures were modelled by magnetic
dipoles and, using the aircraft track data, the best-fit dipole
signatures were removed from the profile data. Although there was
some residual signature from the wrecks, it was much smaller in
amplitude than the original wreck signatures and it had little
effect on the subsequent processing.
Greenwood Base
TFGr 8 Hz
Sable Base TFS, North, 8 Hz
Aircraft Raw DataGPS 4 Hz
Others 32 Hz
GPS Base Raw Data4 Hz
Inspect & Correct
Inspect & Correct
Inspect & Correct
Inspect & Correct
DGPS Processing, Interpolate to 32 Hz,
Low-Pass filter at 1 Hz, sub-sample
to 4 Hz
Standard Compensation,Adaptive Compensation,Low-Pass filter at 1
Hz
Sub-sample to 4 Hz
Interpolate to 32 Hz, Low-Pass filter at 1 Hz, sub-sample
To 4 Hz
Time-align, write a single output file for each flight, 4 Hz
Interpolate to 32 Hz, Low-Pass filter at 1 Hz, sub-sample
To 4 Hz
Figure 5. Flow chart describing the data pre-processing
steps.
DRDC Atlantic TM 2006-004 6
-
4. Basestation Analysis
4.1 Sable Island magnetic field components Figure 6 shows the
time series of the three components of the magnetic field (North,
West, Up) as recorded at the Sable Island basestation site on
August 6, and Figure 7 shows the corresponding power spectral
density (PSD) plots. The DC value has been removed for display
purposes. It is clear that the vertical component has much less
high-frequency activity than either the North or the West
component. This suggests that Sable Island is a very good
electrical conductor, in agreement with Ref 5 which indicates that
the conductivity of Sable Island is about the same conductivity as
the surrounding seawater. There is essentially no power in the
frequency band of interest (0.01-0.5 Hz) in the vertical component
of the geomagnetic field, so only the North and West components are
relevant to this analysis.
Figure 6. Geomagnetic field components (Black = North, Blue =
West, Red = Up) recorded at Sable Island on August 6, 2005.
DRDC Atlantic TM 2006-004 7
-
Figure 7. PSD of geomagnetic field components (Black = North,
Blue = West, Red = Up) recorded at Sable Island on August 6,
2005.
We can also compare the total-field measurements made with the
Caesium magnetometer to the vector magnetometer measurements to
determine which vector component contributes the most to the
variations in the total-field. The vector total-field (TFvector) is
given by: TFvector = √(North2 + West2 + Up2) . (1) Alternatively,
we can estimate the total-field from just North, Up, and mean of
West. Let us denote this TFvector’. TFvector’ = √(North2 + 2 + Up2)
. (2) Figure 8 shows the time series of the total-field as measured
by the Caesium magnetometer (TFS), TFvector, and TFvector’. The DC
values have been removed and TFS is offset by 2 nT for display
purposes. Figure 9 shows the PSD of (TFS- TFvector) and (TFS-
TFvector’). Clearly the total-field constructed from the measured
components is much noisier than the total-field measured with the
Caesium magnetometer. This may be due to residual wind or vibration
of the vector sensors which were mounted on a light wooden
platform. However, sensor testing at DRDC Atlantic suggests (Ref 6)
that the 1/f noise for the National Instruments A/D board starts
near 6 Hz and it is possible that some of the noise may be due this
as well.
DRDC Atlantic TM 2006-004 8
-
Figure 8. TFS (Black), TFvector (Blue), and TFvector’ (Red)
measured at Sable Island on August 6, 2005. DC values have been
removed and TFS is offset by 2 nT for display purposes.
Figures 8 and 9 indicate that the geomagnetic activity seen in
the West component actually contributes very little to the
geomagnetic activity measured by a Caesium total-field magnetometer
at Sable Island. Since it was previously shown that the Up
component has no geomagnetic activity in the frequency band of
interest, this implies that only the North component of the
geomagnetic field variation contributes significantly to the
total-field variations seen at Sable Island. Thus it is fortuitous
that the North component of the geomagnetic field was recorded at
Sable Island for the entire experiment, but the Up and West
components were not. It also suggests that there will be very
little difference between using just the total-field sensor, or
both the total-field and North component sensors, for geological
noise reduction.
DRDC Atlantic TM 2006-004 9
-
Figure 9. PSD of TFS (Black), TFvector (Blue), TFS-TFvector
(Green, just barely visible beneath the Red trace), and
TFS-TFvector’ (Red) of the time series data shown in Figure 8.
4.2 Correlation between Sable Island and Greenwood basestations
Figure 10 shows the time series of the total-fields recorded at the
Sable Island and Greenwood (TFGr) basestations on August 10, and
Figure 11 shows the corresponding power spectral densities. Clearly
most of the low-frequency geomagnetic activity is seen by both
basestations and track each other quite closely. The very-low
frequency (periods of several hours) show considerable differences
at the two basestations. Closer examination shows that there are
considerable time lags between the two basestations for some of the
low-frequency geomagnetic signals. These lags can be 20-70 seconds
and are frequency-dependent. This suggests that they may not be
propagating directly from the ionosphere through the air. It is
possible that some of these signals are travelling through the
seawater or seafloor where the propagation speeds are much slower.
Although both basestations see activity in the 0.01-0.02 Hz region,
usually denoted “PC-4 pulsations”, the amplitude of that
geomagnetic signal is significantly greater at Sable Island. If
Sable Island is a good electrical conductor, then the vertical
component of the geomagnetic field will be reduced (as shown in
Figures 6 and 7) and the horizontal components will be amplified.
Depending on the dip angle of the Earth’s field, this can lead to
an increase in total-field anomaly. Ref 1 also found that the
amplitude of the geomagnetic signals at the Greenwood and Keji
basestations were smaller than at basestations near the ocean, so
this result is not unexpected.
DRDC Atlantic TM 2006-004 10
-
There is also evidence of 0.03-0.06 Hz pulsations in the Sable
Island basestation data which are just barely discernible in the
Greenwood data. Figure 12 shows the frequency-domain coherence
between the Greenwood and Sable Island total-field measurements on
August 10. One hundred and one averages were used in calculating
the coherence. There is excellent coherence of the PC-4 pulsations
(~0.85) and somewhat poorer coherence for the pulsations in the
0.03-0.06 Hz band. It should be noted immediately that the
coherence of the geomagnetic field measured between the Greenwood
or Keji basestation, and low-flying magnetometer-equipped aircraft
at similar separations, has been much lower in the previous trials
(Refs 1,2) than is seen in Figure 12. However, the important
distinction here is that the Sable Island basestation is
stationary, and the hypothesis for this experiment is that either
the aircraft was flying through local conductivity anomalies, or
through magnetic fields generated by ocean dynamics. As long as the
conductivity of Sable Island is not changing on the timescale of
10-100 seconds (which is clearly unlikely), and any magnetic noise
caused by ocean dynamics is smaller than the geomagnetic signals
(which is probably true considering the basestation is ~ 500 m from
the ocean), or on time scales greater than 100 seconds (which is
probably true for large-scale processes), then one should expect a
high degree of correlation between the two basestations in this
experiment.
Figure 10. Total-field measured at Greenwood (Black) and Sable
Island (Blue) on August 10. The DC values have been
removed for display purposes.
DRDC Atlantic TM 2006-004 11
-
Figure 11. PSD’s of time series data shown in Figure 10: Sable
Island (Blue) and
Greenwood (Black).
DRDC Atlantic TM 2006-004 12
-
Figure 12. Coherence of the Sable Island and Greenwood
basestation TF data on August 10. Data were high-pass filtered
with a 2nd-order high-pass Butterworth digital filter with a
3-dB
point at 0.004 Hz.
DRDC Atlantic TM 2006-004 13
-
5. Separating Geological and Geomagnetic Noise Geological noise
is a spatial noise source whereas geomagnetic noise is a temporal
noise source. However, because the aircraft is moving, these two
effects are mixed. The technique used here for separating these
noise sources involves:
1) average the measurements along flight lines at nearby
altitudes to obtain an estimate of the geology,
2) subtract this geology estimate from the original airborne
measurements to yield a residual,
3) compare this residual to the geomagnetic total-field measured
at a basestation. Figure 13 shows a flowchart of the processing
involved in this processing. Sections 5.1-5.7 deal with each the
processes shown in red in Figure 13.
5.1 Read data, define desired track, correct for offset from
desired track Appendix B contains the analysis code for the August
12 flight where the North/South lines near Sable Island were flown.
Similar IDL procedures were written for each flight described in
Sections 2.2-2.3. A 4-Hz data file obtained from the pre-processing
described in Figure 5 was read into IDL and the relevant variables
were extracted. Although we intended to fly from the starting
waypoint directly to the ending waypoint, the aircraft can never be
flown exactly on a straight line between these two points. Thus the
first step in the analysis was to define the desired track in
3-dimensional space, and the vector (ΔX, ΔY, ΔZ) from the actual
track to the desired track. It is important to have the same number
of data points along the desired track as there were in the
original flight data. In order to estimate what the total field
would be at each point along the desired track (TF’), the following
simple equation is used:
TF’i = TFi + (ΔXi, ΔYi, ΔZi)·(Gxi, Gyi, Gzi) (3) where i denotes
the ith point along the flight line, TFi is the measured
total-field at that point, and (Gxi, Gyi, Gzi) is the 3-dimensional
total-field gradient vector at each point along the flight line.
For simplicity, the total-field gradient vector will be referred to
simply as the gradient vector. There are several methods for
estimating the gradient vector. The simplest method is to use the
International Geophysical Reference Field (IGRF) model (Ref 7). The
IGRF gradients vary little over the distances flown in this
experiment so constant values of Gx = 0.0038 nT/m (North) Gy = -
0.0014 nT/m (East) Gz = 0.026 nT/m (Down) . (4) This technique
makes sense if the geological gradients are small in comparison to
the IGRF gradients.
DRDC Atlantic TM 2006-004 14
-
Define “Desired track” and deviation Δx,Δy,Δz
from that track
Read 4 Hz file for each flight
Correct TF with gradients × (Δx,Δy,Δz)
yielding TF’ along desired track
Subtract Basestation TF”=TF’-Base
IGRF Gradients?
Sable TF, filtering?
Geocode & average several lines to
obtain estimate of geology TFgeo
Just 3 lines at each altitude?
5.1
5.2
5.3
5.1
5.1
5.6
Subtract TFest from TF & avoid discontinuities
= Resid
0.02 Hz High-Pass filter= ResidHP
(TFHP & TFestHP
Re-compensate ResidHP along
individual lines = ResidHPC &TFC
IGRF Gradients? Measured Gx, Gy & Estimated Gz?
18 vector mag terms Accelerometers, INS, control surfaces,
etc
Plot Results Dif=ResidHPC-Sable TF
Dif vs Position
5.4
5.5
5.6
5.4
5.7
Re-sample geology along desired
track with correct # dps
Correct re-sampled geology with
gradients × (Δx,Δy,Δz)yielding TFest
along actual track
Measured Gx, Gy & Estimated Gz?
Greenwood TF, filtering?
Upward/downward continue nearby lines?
5.4
Use coherent part of geology estimate?
Figure 13. Flow chart of geological noise modelling and removal.
Detailed description of various processes are given in the Sections
shown in red.
DRDC Atlantic TM 2006-004 15
-
The second technique for estimating the gradients is to use
measured values for Galong and Gacross for the horizontal
gradients. The only difficulty with this is that the actual DC
values that come from the Convair’s lateral gradient measurements
are highly dependent on the aircraft compensation. Small
heading-dependent shifts can sometimes remain unless a very careful
analysis of cross-over errors is done for proper surveys. Since
these lines were only flown in two directions (either North/South
or East/West on a given flight), there may be some heading error in
the gradient measurements. If the geological gradients are much
larger than any residual heading error (in other words there is a
large geological gradient), then this technique should work much
better that simply using the IGRF gradient values. However, for the
areas near Sable Island and the Gully, the geological gradients
should be small compared to the IGRF. The technique was tried, but
it was determined that simply using the IGRF horizontal gradients
yielded the best results. The Convair was not equipped with a tail
magnetometer for these experiments so there was no measurement of
the vertical gradient. However, a common technique for estimating
the vertical gradient from measured total-field data along a
profile is to assume that the sources are two-dimensional and
oriented perpendicular to the flight line. Under this assumption,
the vertical gradient is simply the Hilbert Transform of Galong.
(The Hilbert Transform simply phase shifts a time series by 90
degrees.) While this assumption is rarely valid, the vertical
gradient estimates that result from the technique are usually
acceptable. This technique was tried, but it was determined that
simply using the IGRF vertical gradient yielded the best results. A
final technique for estimating both the horizontal and vertical
gradients is to use existing aeromagnetic maps of the area. This
technique hasn’t been tried as yet, but it may be included in
follow-on work on the project if high-quality aeromagnetic maps of
the area can be obtained. The gradient-corrected TF is denoted
TF’.
5.2 Basestation correction If there are enough flight lines that
are being averaged, then one can assume that the geomagnetic noise
measured along the flight lines will average out to zero. In this
experiment however, there were only three flight lines at each
altitude. The following methods were tried for removing the
geomagnetic noise prior to building up a geological noise
model:
1) no basestation removal at all, 2) subtract the Sable Island
basestation TF measurements directly, 3) apply a low-pass smoother
to the Sable Island TF measurements,
then subtract them, 4) subtract the Greenwood basestation TF
measurements directly, 5) subtract the smoothed Greenwood
basestation TF measurements.
The technique which gave the best results was subtracting a
lightly smoothed (5 points = 1.25 second boxcar smoother) version
of the Sable Island total-field. The
gradient-and-basestation-corrected TF is denoted TF’’.
DRDC Atlantic TM 2006-004 16
-
5.3 Combining several lines The simplest method for estimating
the geological noise is to re-sample the
gradient-and-basestation-corrected-TF along each profile to common
(latitude, longitude) points using Akima Splines, and compute the
average. However, other techniques can be used. If one wishes to
use the data collected at other altitudes (say data from the 500’
lines to calculate the geological model for 1000’), then there is a
common geophysical technique known as upward/downward continuation
that allows one to estimate the total-field at one altitude based
on measurements at another. If a 2-dimensional survey has been
conducted, then the upward/downward continuation algorithms work
quite well unless one attempts to estimate the total-field quite
close to the source, based on measurements taken quite far away.
However, in our case, we have only a series of flight lines at
various altitudes. If we again use the assumption that the magnetic
sources are 2-dimensional and oriented perpendicular to the flight
line, then we can not only estimate the vertical gradient along the
flight line, but the 2nd and 3rd-order vertical gradients as well.
A Taylor expansion can then be used to estimate the total-field
along a flight line at a higher or lower altitude as follows:
TFi(z) = TFi + (zi · Gzi) + (zi2 · Gzzi)/2 + (zi3 · Gzzzi)/6 + …
(5) where z is the new height relative to the old height, i denotes
the ith point along the profile,
Gz = - Hilbert Transform (Galong) Gxz = d(Gz)/dx Gzz = - Hilbert
Transform (Gxz) Gxzz = d(Gzz)/dx Gzzz = - Hilbert Transform (Gxzz)
and (d/dx) denotes the spatial derivative of a quantity along the
flight track. When applying this algorithm, there is often some
smoothing applied to the higher-order vertical gradient estimates
because they tend to be dominated by high-frequency noise. Lines
from different altitudes can then be used in the average, thereby
possibly improving the geological noise model. Once again, if a
high-quality 2-dimensional total-field survey was conducted, then
Gz, Gzz, and Gzzz could be calculated without the assumption of
linear, two-dimensional sources perpendicular to the flight path.
Finally, instead of merely calculating the average of multiple
estimates of the gradient-and-basestation-corrected-TF, or multiple
upward/downward continued lines, it is possible to take only the
coherent part of the signals as the estimate for the geology
(denoted TFgeo in subsequent text). The technique which gave the
best results was simply to average the three lines from each
altitude, although the other techniques gave only marginally
different results.
DRDC Atlantic TM 2006-004 17
-
5.4 Re-sampling and reverse-gradient correction back to the
original sampling positions Once an estimate of the geological
noise along each desired path has been obtained, it can be
re-sampled with the same number of data points as in the original
flight line. Then it is reverse-corrected with the same gradient
vector (Gxi, Gyi, Gzi) and deviation from the desired path vector
(ΔXi, ΔYi, ΔZi) yielding an estimate of the geological noise at the
exact location where the original TF measurement was made. This
geological noise estimate is denoted “TFest”. Subtracting this
geological noise component from the original TF measurement gives a
first-order estimate of the geomagnetic signal along each flight
path. Note that up until this point, everything has been calculated
right down to DC, but eventually we wish to filter the residual
data in order to accentuate the geomagnetic activity in the
frequency band of interest. In order to avoid large discontinuities
which may cause filter ringing in later processing, the time series
TFest is adjusted in the following manner:
1) during the turns between the lines, TFest is set equal to the
actual flight data during these turns,
2) the DC level of TFest along each line is shifted slightly so
there is no discontinuity at the first or last point of the
line.
These corrections have almost no effect on the geological
estimate within the frequency band of interest, but do allow
subsequent processing to be carried out over the entire flight
line. The difference between the original TF and the
gradient-corrected geological noise estimate (TFest) is denoted
“Resid”.
5.5 High-Pass filter the original TF and geological noise
estimate For most of the flights performed during this experiment,
the geomagnetic field was active near or above 0.02 Hz. A 2nd-order
digital Butterworth high-pass filter with a 3-dB point at 0.02 Hz
was applied to the original airborne TF measurements and the
geological noise estimate (TFest). These quantities will be
referred to as TFHP and TFestHP in subsequent text. Other filters
were investigated, including a similar filter with a 3 dB point at
0.01 Hz high-pass, and multiple applications of the 0.02 Hz
high-pass filter. The results were very similar so only a single
application of the 0.02 Hz filter was used in subsequent analyses.
The difference between the TFHP and TFestHP is denoted
“ResidHP”.
5.6 Re-compensate the filtered residual Even though the standard
compensation and adaptive compensation algorithms remove the vast
majority of the aircraft interference, it is possible that there is
some residual noise remaining. To remove this noise, a 25-term
model was fit to ResidHP on each line. The model consisted of the
standard 18 Leliak terms generated from the vector magnetometers,
plus 7 terms from the
DRDC Atlantic TM 2006-004 18
-
pitch, roll, DGPS altitude, left and right accelerometers,
aileron, and (Gacross· ΔY). Each of the 25 terms was filtered with
the same 0.02 Hz high-pass filter used to generate ResidHP. The
result is denoted ResidHPC The effect of this re-compensation was
to reduce the noise where the aircraft had done slight banks to
stay on track. In these cases, the magnetometers located on the
wingtips moved up, or down, through the vertical gradient in the
Earth’s magnetic field. Also, the aircraft moved back and forth
through the lateral gradient in the Earth’s field in response to
these slight banks and that is why a term for (Gacross· ΔY) was
included. Note that although a similar term was used for the
original gradient correction (Section 5.1), it was mainly the
very-low frequency part of the term that was important in trying to
build a better geology model. In this case it is only the
variations above 0.02 Hz that are important and any small DC errors
in Gacross are not as important. The residual during periods where
the aircraft was not banking was barely affected. It should be
noted that most of the Leliak terms, aileron, pitch, and
accelerometer terms had very little effect on the re-compensation.
Once the re-compensation coefficients have been calculated, it is
possible to apply them to the unfiltered versions of the 25-terms
and perform a “wide-band re-compensation” on the original airborne
TF measurements. The result is denoted TFC. Passing TFC through the
0.02 Hz Butterworth high-pass filter yields the quantity TFCHP.
Care must be taken not to introduce any DC offsets during this
process, and the re-compensation coefficients must only be applied
to the original flight line on which they were calculated. This is
because there is very little variation in some of the 25 terms and
there are many colinearities between them. The entire process of
building up a geological model and removing the geology from the
airborne TFC measurements can then repeated. It was determined that
iterating more than once did not significantly improve ResidHPC. It
is worth noting that in standard aircraft compensation, the signals
are bandpass filtered near 0.1 Hz instead of 0.02 Hz in order to
separate the aircraft motion noise from the geological signals. In
this case we have attempted to remove as much geological noise as
possible and then perform a re-compensation.
5.7 Plotting the results The residual after all the processing,
ResidHPC, was compared to the similarly-filtered Sable Island
basestation total field (TFSHP) in the following ways:
1) overplot the two time series and their difference (DIF1). 2)
perform a frequency-domain coherence analysis of ResidHPC and
TFSHP
and overplot the coherence residual (DIF2). 3) plot the
coherence vs frequency and the PSD of the basestation data vs
frequency to determine if the coherence is better for large
geomagnetic signals.
4) plot DIF1 or DIF 2 vs latitude (or longitude depending on the
flight line direction) to determine if the places where the
differences are the greatest occur at the same position.
5) determine the portions of each line where the geomagnetic
activity was “large” or “small” and compare the DIF1 (or DIF2) in
these regions.
DRDC Atlantic TM 2006-004 19
-
6) Perform a frequency-domain coherence analysis of ResidHPC vs
only the North component of the geomagnetic field measured at Sable
Island (NorthSHP). This did not lead to any better results than
against the TFSHP and so won’t be presented here.
DRDC Atlantic TM 2006-004 20
-
6. Results
6.1 Sable Island North/South Lines Figure 14a shows the raw
total-field aircraft data (TF) vs the gradient- and
basestation-corrected total-field (TF”) for the twelve North/South
lines flown near Sable Island. In general the blue lines overlap
better than the black lines, indicating that corrections have at
least been applied in the correct sense.
Figure 14a. Comparison of the raw total-field (TF) and the
gradient- and basestation-corrected total-field (TF”) along the
three North/South lines near Sable Island at each altitude (Black
vs,
Blue).
Figures 14b-e show the horizontal gradient corrections applied
to each line. They are predominantly low-frequency and correspond
to the aircraft manoeuvres as the pilots attempted to maintain the
desired track. These corrections are typically less that 1 nT in
amplitude.
DRDC Atlantic TM 2006-004 21
-
Figure 14e. Horizontal gradient correction
applied to each North/South line at 5000’ near Sable Island,
based on IGRF gradients.
Blue=ΔNorth x GNorth; Red=ΔEast x GEast.
Figure 14d. Horizontal gradient correction applied to each
North/South line at 2000’ near
Sable Island, based on IGRF gradients. Blue=ΔNorth x GNorth;
Red=ΔEast x GEast.
Figure 14c. Horizontal gradient correction applied to each
North/South line at 1000’ near
Sable Island, based on IGRF gradients. Blue=ΔNorth x GNorth;
Red=ΔEast x GEast.
Figure 14b. Horizontal gradient correction applied to each
North/South line at 500’ near
Sable Island, based on IGRF gradients. Blue=ΔNorth x GNorth;
Red=ΔEast x GEast.
DRDC Atlantic TM 2006-004 22
-
Figure 14f shows the individual geology estimates and the
averages for each altitude. The three geology estimates for each
altitude overlap almost perfectly. There is a small amount of
very-low-frequency difference between the three estimates which
suggests that there may in fact be some phase lag between the
very-low-frequency geomagnetic activity measured at the Sable
Island basestation and that measured along the flight lines.
However, this will have no effect on the final high-pass filtered
geology estimate.
Figure 14f. Comparison of geology estimates along the three
North/South lines near Sable Island at each altitude (Black,
Blue,
Green) vs the average (Red). TFgeo is set to the average.
Figures 14g-j compare the original high-pass filter total-field
measurements with the geology estimate removed (ResidHP) to the
re-compensated quantity (ResidHPC). By comparing these figures to
Figures 14b-e, it can be seen that the only significant changes
occur during the aircraft manoeuvres where the pilots were bringing
the aircraft back onto the desired track. Three things occurred at
these time – the aircraft rolled a few degrees, the magnetometers
in the wingpods moved up (or down) through the vertical gradient,
and the aircraft moved laterally through the horizontal gradient in
the ambient field. All of these effects are modelled by the 25-term
model used for re-compensation. It is equally important to note
that when the pilots were not manoeuvring the aircraft to get back
onto the desired track, the re-compensation process does not
increase the magnetic noise.
DRDC Atlantic TM 2006-004 23
-
Figure 14j. Effect of extra compensation along the 5000’
North/South lines near Sable Island: ResidHP
(Black) vs ResidHPC (Blue).
Figure 14i. Effect of extra compensation along the 2000’
North/South lines near Sable Island: ResidHP
(Black) vs ResidHPC (Blue).
Figure 14h. Effect of extra compensation along the 1000’
North/South lines near Sable Island: ResidHP
(Black) vs ResidHPC (Blue).
Figure 14g. Effect of extra compensation along the 500’
North/South lines near Sable Island: ResidHP
(Black) vs ResidHPC (Blue).
DRDC Atlantic TM 2006-004 24
-
Figures 14k-n compare the filtered re-compensated airborne
measurements to the geology estimates (TFCHP vs TFestHP), the
residual after subtracting the geology estimate to the Sable Island
basestation (ResidHPC vs TFSHP), and the two residuals of the
latter using either simple subtraction or frequency-domain
coherence processing (DIF1 vs DIF2). The coherence and the PSD of
the geomagnetic field as measured at the Sable Island basestation
are also shown. The four figures correspond to the four different
flight altitudes. Each figure contains individual plots of the
three lines flown at each altitude. The first thing to note is that
the geology estimate tracks the aircraft data extremely well (see
top traces in each figure). This indicates that the geology
estimates, and the process used to correct for the basestation and
gradients is reasonable. It may not be optimum, but it is
reasonable. The second thing to notice is that when the geology
estimate is subtracted from the aircraft measurements, the
low-frequency residual looks very much like the Sable Basestation
total-field (see middle traces in each upper figure). There are a
few locations where the two differ, but these are at places where
the geological signal is quite large and variable between the three
lines; e.g. near 43.98º, 44.05º, 44.07º and 44.10º in the 500’ and
1000’ data (Figures 14k-l). These are the locations where the
simple IGRF gradient corrections are the least valid so it is not
surprising that the geological model is poorest in these areas.
However, the differences tend to be at a higher frequency than the
geomagnetic noise. The third thing to notice is that there is very
little difference between the residuals formed by simple
subtraction of TFSHP from ResidHPC (DIF1) and the frequency-domain
coherence processing of the two time series (DIF2). This suggests
that in fact there is not a simple change in amplitude or phase of
the geomagnetic field (as measured at the Sable Island basestation)
and the flight lines nearby at 500-5000 feet of altitude. In
general the geomagnetic activity was quite low when these lines
were flown, but in the few places where the geomagnetic signals
were significant in the Sable Island basestation data, they were
highly correlated with the airborne data (e.g. south ends of L1 and
L3 @ 500’, L2 and L3 @ 2000’, and L1 and L3 @ 5000’). The coherence
is usually < 0.8, but where the geomagnetic signal is larger (L2
@ 2000’ and L3 and 5000’), the coherence is higher. This strongly
suggests that the poorer coherence seen on the other lines is
because there is excess noise from some other source such as
imperfect geological noise cancellation or ocean dynamics. If the
lack of coherence was because of something changing the local
amplitude or phase of the geomagnetic signal, then the coherence
would not be larger for larger geomagnetic signals. However,
because we do not have large geomagnetic signals during the low
altitude flights where the effect should be greatest, we cannot
definitively conclude this.
DRDC Atlantic TM 2006-004 25
-
Figure 14k. Upper Trace: Comparison of signals measured along
the three 500’
North/South lines near Sable Island: TFCHP vs TFestHP, (Black vs
Red at the top of each upper trace); ResidHPC vs TFSHP (Blue vs
Green in the middle of each upper trace); DIF1 vs DIF2 (Black vs
Orange at bottom of each
upper trace).
Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of
ResidHPC (blue), TFSHP (black), and DIF2 (red).
DRDC Atlantic TM 2006-004 26
-
Figure 14l. Upper Trace: Comparison of signals measured along
the three 1000’ North/South lines near Sable Island: TFCHP vs
TFestHP, (Black vs Red at the top of each upper trace); ResidHPC vs
TFSHP (Blue vs Green in the middle of each upper trace); DIF1 vs
DIF2 (Black vs Orange at bottom of each upper
trace).
Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of
ResidHPC (blue), TFSHP (black), and DIF2 (red).
DRDC Atlantic TM 2006-004 27
-
Figure 14m. Upper Trace: Comparison of signals measured along
the three 2000’
North/South lines near Sable Island: TFCHP vs TFestHP, (Black vs
Red at the top of each upper trace); ResidHPC vs TFSHP (Blue vs
Green in the middle of each upper trace); DIF1 vs DIF2 (Black vs
Orange at bottom of each
upper trace).
Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of
ResidHPC (blue), TFSHP (black), and DIF2 (red).
DRDC Atlantic TM 2006-004 28
-
Figure 14n. Upper Trace: Comparison of signals measured along
the three 5000’
North/South lines near Sable Island: TFCHP vs TFestHP, (Black vs
Red at the top of each upper trace); ResidHPC vs TFSHP (Blue vs
Green in the middle of each upper trace); DIF1 vs DIF2 (Black vs
Orange at bottom of each
upper trace).
Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of
ResidHPC (blue), TFSHP (black), and DIF2 (red).
DRDC Atlantic TM 2006-004 29
-
Figure 14o is a plot of DIF1 vs latitude. It is clear that the
largest residuals are at the lowest altitudes, and are clustered
towards the northern end of the lines where the geological signals
are the largest and most complex (compare Figures 14k and 14o).
Figure 14o. DIF1 vs Latitude for North/South lines near Sable
Island.
DRDC Atlantic TM 2006-004 30
-
6.2 Sable Island East/West Lines Figure 15a shows the raw
total-field aircraft data (TF) vs the gradient- and
basestation-corrected total-field (TF”) for the twelve East/West
lines flown near Sable Island. In general the blue lines overlap
better than the black lines. The amplitude of the horizontal
gradient corrections and improvement obtained by re-compensation
were not very different than for the North/South lines shown in
Figures 14b-e and 14-g-j, so they have not been included.
Figure 15a. Comparison of the raw total-field (TF) and the
gradient- and basestation-corrected total-field (TF”) along the
three East/West lines near Sable Island at each altitude (Black
vs,
Blue).
Figure 15b shows the individual geology estimates (TFgeo) and
the averages for each altitude. The three geology estimates for
each altitude are very similar but there is a small amount of
very-low-frequency difference between the three estimates. Again
this suggests that there may in fact be some phase lag between the
very-low-frequency geomagnetic activity measured at the Sable
Island basestation and that measured along the flight lines even
though they are separated by less than 25 km. However, this will
have no effect on the final high-pass filtered geology
estimate.
DRDC Atlantic TM 2006-004 31
-
Figure 15b. Comparison of geology estimates along the three
East/West lines near Sable Island at each altitude (Black, Blue,
Green)
vs the average (Red). TFgeo is set to the average.
Figures 15c-f compare the filtered re-compensated airborne
measurements to the geology estimates (TFCHP vs TFestHP), the
residual after subtracting the geology estimate to the Sable Island
basestation (ResidHPC vs TFSHP), and the two residuals of the
latter using either simple subtraction of frequency-domain
coherence processing (DIF1 vs DIF2). The coherence and the PSD of
the geomagnetic field as measured at the Sable Island basestation
are also shown. The four figures correspond to the four different
flight altitudes. Each figure contains individual plots of the
three lines flown at each altitude. Just as was seen in the
North/South results, the geology estimate tracks the aircraft data
extremely well (see top traces in each figure). This indicates that
the geology estimates, and the process used to correct for the
basestation and gradients are reasonable. The quantity ResidHPC
looks very much like the Sable Basestation total field (see middle
traces in each figure). There are a few locations where the two
differ, but these are at places where the geological signal is
quite large and variable between the three lines; e.g. near
-60.02º, -59.95º, and -59.81º in the 500’ and 1000’ data (Figures
15c-d). Again this is consistent with the
DRDC Atlantic TM 2006-004 32
-
results from the North/South flight line data. These differences
also tend to be at a higher frequency than the geomagnetic noise.
Once again there is very little difference between the residuals
formed by simple subtraction of ResidHPC and TFSHP (DIF1) and the
frequency-domain coherence processing of the two time series
(DIF2). This suggests that in fact there is no appreciable change
in amplitude or phase of the geomagnetic field variations near
0.02-0.05 Hz as measured at the Sable Island basestation and the
nearby East/West flight lines at 500-5000 feet of altitude.
(Remember there does appear to be some phase lag at much lower
frequencies because the various estimates for TFgeo do have some
very-low-frequency variation in them.) The geomagnetic activity was
quite variable when the East/West lines were flown, but in the
places where the geomagnetic signals were large in the Sable Island
basestation data, they were highly correlated with the airborne
data (e.g. West end of L1 & L2 @ 500’; L2 and East end of L3 @
1000’; L1 @ 2000’). The coherence, even at low altitude, is >0.8
when there is a substantial geomagnetic signal and smaller when
there is less geomagnetic activity. If we compare the same
geographic area on lines flown at the same altitude, but at times
when the geomagnetic field was quiet vs active (East ends of L1
& L2 @500’ in Figure 15c or the West ends of L1 & L2 @
1000’ Figure 15d), we see that the nature of DIF1 does not change.
Taken together, these two results suggest that the remaining
magnetic noise seen in DIF1 at the lower altitudes for the
East/West lines near Sable Island is independent of the geomagnetic
field amplitude. This implies that there are no conductivity
effects at low altitude. Remember, though, that the separation
between the basestation and these flight lines was less than 25 km,
and the water depth is less than 100 m near Sable Island.
DRDC Atlantic TM 2006-004 33
-
DRDC Atlantic TM 2006-004 34
Figure 15c. Upper Trace: Comparison of
Ea
Middle Trace: Cohe ween ResidHPC
Lower Trace: Pow al Density of R
signals measured along the three 500’ st/West lines near Sable
Island: TFCHP vs
TFestHP, (Black vs Red at the top of each upper trace); ResidHPC
vs TFSHP (Blue vs Green in the middle of each upper trace); DIF1 vs
DIF2
(Black vs Orange at bottom of each upper trace).
ren e betc
and TFSHP.
e Spectrr esidHPC (blue), TFSHP (black), and DIF2
(red).
-
Figure 15d. Upper Trace: Comparison of signals measured along
the three 1000’
East/West lines near Sable Island: TFCHP vs TFestHP, (Black vs
Red at the top of each upper trace); ResidHPC vs TFSHP (Blue vs
Green in the middle of each upper trace); DIF1 vs DIF2
(Black vs Orange at bottom of each upper trace).
Middle Trace: Coherence between ResidHPC
and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2
(red).
DRDC Atlantic TM 2006-004 35
-
Figure 15e. Upper Trace: Comparison of signals measured along
the three 2000’
East/West lines near Sable Island: TFCHP vs TFestHP, (Black vs
Red at the top of each upper trace); ResidHPC vs TFSHP (Blue vs
Green in the middle of each upper trace); DIF1 vs DIF2
(Black vs Orange at bottom of each upper trace).
Middle Trace: Coherence between ResidHPC
and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black), and DIF2
(red).
DRDC Atlantic TM 2006-004 36
-
Figure 15f. Upper Trace: Comparison of signals measured along
the three 5000’ East/West lines near Sable Island: TFCHP vs
TFestHP, (Black
vs Red at the top of each upper trace); ResidHPC vs TFSHP (Blue
vs Green in the middle of each upper trace); DIF1 vs DIF2 (Black vs
Orange at bottom of each upper
trace).
Middle Trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of
ResidHPC (blue), TFSHP (black), and DIF2 (red).
DRDC Atlantic TM 2006-004 37
-
Figure 15g is a plot of DIF1 vs longitude. It is clear that the
largest residuals are at the lowest altitudes, and are clustered
towards the middle of the lines where the geological signals are
the largest and most complex (compare Figures 15c and 15g).
Figure 15g. DIF1 vs Longitude for East/West lines near Sable
Island.
DRDC Atlantic TM 2006-004 38
-
6.3 Gully North/South Lines Figure 16a shows the raw total-field
aircraft data (TF) vs the gradient- and basestation-corrected
total-field (TF”) for the twelve North/South lines flown near the
Gully. Again the horizontal gradient correction and re-compensation
improvement plots have been omitted for brevity.
.
Figure 16a. Comparison of the raw total-field (TF) and the
gradient- and basestation-corrected total-field (TF”) along the
three North/South lines near the Gully at each altitude (Black
vs,
Blue). Figure 16b shows the individual geology estimates and the
averages for each altitude. The three geology estimates for each
altitude overlap almost perfectly for the three lower altitudes,
but there is a significant very-low-frequency difference between
the three estimates at 10,000 ft altitude. However, this will have
no effect on the final high-pass filtered geology estimate.
DRDC Atlantic TM 2006-004 39
-
Figure 16b. Comparison of geology estimates along the three
North/South lines near the Gully at each altitude (Black, Blue,
Green)
vs the average (Red). TFgeo is set to the average.
Figures 16c-f compare the filtered re-compensated airborne
measurements to the geology estimates (TFCHP vs TFestHP), the
residual after subtracting the geology estimate to the Sable Island
basestation (ResidHPC vs TFSHP), and the two residuals of the
latter using either simple subtraction of frequency-domain
coherence processing (DIF1 vs DIF2). The coherence and the PSD of
the geomagnetic field as measured at the Sable Island basestation
are also shown. The four figures correspond to the four different
flight altitudes. Each figure contains individual plots of the
three lines flown at each altitude. Again the geology estimate
tracks the aircraft data extremely well (see top traces in each
figure). The geological noise in the Gully area is generally less
than in the Sable Island area. Unfortunately the geomagnetic field
was fairly quiet during the three 1000 ft altitude lines so it is
difficult to draw firm conclusions on the how well ResidHPC
correlates with the Sable Island Basestation signal (TFSHP) at low
altitudes. The coherence is
-
Figure 16c. Upper Trace: Comparison of signals measured along
the three 1000’
North/South lines near the Gully: TFCHP vs TFestHP, (Black vs
Red at the top of each upper trace); ResidHPC vs TFSHP (Blue vs
Green in the middle of each upper trace); DIF1 vs DIF2
(Black vs Orange at bottom of each upper trace).
Middle trace: Coherence between ResidHPC
and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black) and DIF2
(red).
DRDC Atlantic TM 2006-004 41
-
Figure 16d. Upper Trace: Comparison of signals measured along
the three 2000’
North/South lines near the Gully: TFCHP vs TFestHP, (Black vs
Red at the top of each upper trace); ResidHPC vs TFSHP (Blue vs
Green in the middle of each upper trace); DIF1 vs DIF2
(Black vs Orange at bottom of each upper trace).
Middle trace: Coherence between ResidHPC
and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black) and DIF2
(red).
DRDC Atlantic TM 2006-004 42
-
Figure 16e. Upper Trace: Comparison of signals measured along
the three 5000’
North/South lines near the Gully: TFCHP vs TFestHP, (Black vs
Red at the top of each upper trace); ResidHPC vs TFSHP (Blue vs
Green in the middle of each upper trace); DIF1 vs DIF2
(Black vs Orange at bottom of each upper trace).
Middle trace: Coherence between ResidHPC
and TFSHP.
Lower Trace: Power Spectral Density of ResidHPC (blue), TFSHP
(black) and DIF2
(red).
DRDC Atlantic TM 2006-004 43
-
Figure 16f. Upper Trace: Comparison of signals measured along
the three 10,000’ North/South
lines near the Gully: TFCHP vs TFestHP, (Black vs Red at the top
of each upper trace);
ResidHPC vs TFSHP (Blue vs Green in the middle of each upper
trace); DIF1 vs DIF2 (Black vs Orange at bottom of each upper
trace).
Middle trace: Coherence between ResidHPC and TFSHP.
Lower Trace: Power Spectral Density of
ResidHPC (blue), TFSHP (black) and DIF2 (red).
DRDC Atlantic TM 2006-004 44
-
Only the northern half of L3 at 2000 ft altitude has significant
geomagnetic activity and the phase/amplitude match of ResidHPC and
TFSHP appears better than in the segment at 1000 ft. Again, with so
little data it is difficult to draw any firm conclusions. There is
a great deal of geomagnetic activity in all the lines at 5000 and
10,000 ft altitude and the phase and amplitude of ResidHPC and
TFSHP match very well. In all cases there is very little difference
between the residuals formed by simple subtraction of ResidHPC and
TFSHP (DIF1) and the frequency-domain coherence processing of the
two time series (DIF2). This suggests that in fact there is no
consistent change in amplitude or phase of the geomagnetic field
between Sable Island and the Gully between 1000-10,000 feet of
altitude. There may still be local variations at the lower
altitudes as described above. Figure 16g is a plot of DIF1 vs
latitude. It is clear that the largest residuals are at the lowest
altitudes, and are near where the geological signals are the
largest (compare Figures 16c and 16g).
Figure 16g. DIF1 vs Latitude for North/South lines near the
Gully.
DRDC Atlantic TM 2006-004 45
-
6.4 Gully East/West Lines Figure 17a shows the raw total-field
aircraft data (TF) vs the gradient- and basestation-corrected
total-field (TF”) for the twelve East/West lines flown near the
Gully. The two traces for the 10,000’ lines fall almost on top of
each other because there was very little geomagnetic activity at
this time. The horizontal gradient correction and re-compensation
improvement plots have been omitted for brevity.
Figure 17a. Comparison of the raw total-field (TF) and the
gradient- and basestation-corrected total-field (TF”) along the
three East/West lines near the Gully at each altitude (Black vs,
Blue).
Figure 17b shows the individual geology estimates and the
averages for each altitude. The three geology estimates for each
altitude overlap very well at 1000 and 5000 ft altitude, but there
is some very-low-frequency difference between the three estimates
at 2000 and 10,000 ft. Again this will have no effect on the final
high-pass filtered geology estimate.
DRDC Atlantic TM 2006-004 46
-
Figure 17b. Comparison of geology estimates along the three
East/West lines near the Gully at each altitude (Black, Blue,
Green) vs
the average (Red). TFgeo is set to the average.
Figures 17c-f compare the filtered re-compensated airborne
measureme