Journal of
Applied Physics ARTICLE scitation.org/journal/jap
Do cities have a unique magnetic pulse?
Cite as: J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264
Submitted: 15 February 2022 · Accepted: 23 April 2022 ·
View Online Export Citation CrossMark
Published Online: 31 May 2022
V. Dumont,1,a) T. A. Bowen,2,3 R. Roglans,2,3 G. Dobler,4,5,6,7 M. S. Sharma,7 A. Karpf,8 S. D. Bale,2,3
A. Wickenbrock,9,10 E. Zhivun,2 T. Kornack,11 J. S. Wurtele,2 and D. Budker2,9,10
AFFILIATIONS
1Computational Research Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA
2Department of Physics, University of California, Berkeley, Berkeley, California 94720-7300, USA
3Space Sciences Laboratory, University of California, Berkeley, Berkeley, California 94720-7300, USA
4Biden School of Public Policy and Administration, University of Delaware, Newark, Delaware 19716, USA
5Department of Physics and Astronomy, University of Delaware, Newark, Delaware 19716, USA
6Data Science Institute, University of Delaware, Newark, Delaware 19713, USA
7Center for Urban Science and Progress, New York University, Brooklyn, New York 11201, USA
8Civil and Urban Engineering, Tandon School of Engineering, New York University, Brooklyn, New York 11201, USA
9Institut für Physik, Johannes Gutenberg-Universität Mainz, Staudingerweg 7, 55128 Mainz, Germany
10Helmholtz Institut Mainz, Staudingerweg 18, 55128 Mainz, Germany
11Twinleaf LLC, 300 Deer Creek Drive, Plainsboro, New Jersey 08536, USA
a)Author to whom correspondence should be addressed: vincentdumont11@gmail.com
ABSTRACT
We present a comparative analysis of urban magnetic fields between two American cities: Berkeley (California) and Brooklyn Borough of
New York City (New York). Our analysis uses dataptffiaffiffikffiffiffien over a four-week period during which magnetic field data were continuously
recorded using a fluxgate magnetometer with 70 pT/ Hz noise. We identified significant differences in the magnetic signatures. In particu-
lar, we noticed that Berkeley reaches a near-zero magnetic field activity at night, whereas magnetic activity in Brooklyn continues during
nighttimpe.ffiffiffiffiWffiffi e also present auxiliary measurements acquired using magnetoresistive vector magnetometers (VMRs), with the noise of
300 pT/ Hz, and demonstrate how cross correlation, and frequency-domain analysis, combined with data filtering can be used to extract
urban-magnetometry signals and study local anthropogenic activities. Finally, we discuss the potential of using magnetometer networks to
characterize the global magnetic field of cities and give directions for future development.
© 2022 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://
creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/5.0088264
I. INTRODUCTION the analysis of the work/sleep patterns of urban dwellers (with mea-
Cities are among the most complex systems that are of utmost surements carried out in a way to ensure privacy of individuals
8).
importance for humanity. The multifaceted and dynamic properties With regard to the magnetic field in urban environments,
of cities are determined by intricate combinations of natural, anthro- studies have generally been limited to particular applications, such
9
pogenic, and socio-economic factors. In recent years, a novel as health and safety or geophysical prospection of archeological
1 sites.10,11approach to the study of the cities was introduced, in which a city is Motivated by the success of the multispectral approach,
12
studied, similar to an astronomical object in the multi-messenger we built a prototype network for urban magnetometers and con-
astronomy approach, with an array of observational instruments, such ducted measurements in the San Francisco Bay Area, analyzing the
as, for example, multispectral cameras.2–4 The analysis of such data dominant sources of magnetic signals and learning to extract subtle
has led to important insights into the working of cities,5–7 of impor- information in the presence of much larger backgrounds.
tance in such diverse areas as improving energy efficiency, reducing Here, we report the next step in the urban-magnetometry
pollution, and increasing our understanding of social organization via program, in which we compare the magnetic signatures of two
J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264 131, 204902-1
© Author(s) 2022
 17 October 2023 18:58:05
Journal of
Applied Physics ARTICLE scitation.org/journal/jap
cities, Berkeley (CA) and Brooklyn Borough of New York City Data from each magnetic field direction (that is, X, Y, and Z) are
(NY). Apart from the anticipated result that “New York never stored hourly in separate binary files.
sleeps,” our measurements indicate that each city has distinct mag- Field measurements were performed using magnetoresistive
netic signatures that can, perhaps, be exploited for the analysis of vector magnetometers (VMRs) mpaffiffinffiffiffiuffi factured by Twinleaf LLC
anomalies in city operation and long-term trends of the develop- with the noise at 1 Hz of 300 pT/ Hz.16 In this work, we analyze
ment of cities. the total scalar field and not individual vector measurements from
each axes.17 In terms of acquisition, the Twinleaf sensors do not
require any data-acquisition device and can be powered directly
from a laptop USB port, making them ideal for field measurements
II. EXPERIMENTAL DETAILS (see Sec. III C).
A. Sensor type and data acquisition The geomagnetic field was acquired from the United States
Two types of magnetometers were used to measure the mag- Geological Survey (USGS) using the open-source library Geomag18
netic field in Brooklyn. The base stations were built using eFM-3A Algorithms. The USGS station (FRN) nearest to Berkeley is
three-axis fluxgate magnetometers manufactured by BioMed Jena located 200 miles away, in Fresno, California. For the Brooklyn
data, the nearest USGS station (FRD) is in Corbin, Virginia, about
GmbH.13 These Biomed sensors apreffiffiffitffiffiiffied to a specific location and
have a noise level of about 70 pT/ Hz.14
300 miles away from New York City.
We note that sensors that
have lower noise may not be of much advantage for urban magne-
tometry because the environmental noise by far exceeds the sensor B. Activity period
noise; see, for example, Fig. 1. A power supply from the same man- Data from the Biomed sensors were obtained over four weeks
ufacturer is used to connect the magnetometers to a computer. from each city during the calendar year 2016 for Berkeley and 2018
Digitized magnetic field measurement data are transferred using a for Brooklyn. More specifically, the data used from Berkeley were
universal serial bus (USB) connection. The data streamed from the taken from Monday, March 14, 2016, to Monday, April 11, 2016.
Biomed sensors are sampled at 3960 Hz and recorded on the com- The data from Brooklyn were acquired from Monday, May 7, 2018,
puter using the publicly available URBANM 15AGNETOMETER software. to Monday, June 4, 2018; this period included the US Federal
FIG. 1. Full four-week time series of urban magnetic field data for Berkeley (top) and Brooklyn (bottom). Data downsampled to 1 Hz are shown in blue, while the data in
yellow represent the downsampled hour-rate time series. The weekends are highlighted by the light green regions and holiday (Memorial Day) in light red. We note the dif-
ferences in vertical scales between both cities; in particular, the excursions of magnetic field are significantly larger in Brooklyn. The geomagnetic field taken from the
closest USGS station is shown in orange. The relative variation of the magnetic field around its mean value is shown on the right hand side with the distribution from the
geomagnetic field represented in orange.
J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264 131, 204902-2
© Author(s) 2022
 17 October 2023 18:58:05
Journal of
Applied Physics ARTICLE scitation.org/journal/jap
Memorial Day holiday, observed the last Monday of the month III. COMPARATIVE DATA ANALYSIS
(May 28, 2018), which is highlighted in red in Fig. 1. Holidays are A. Time-domain observations
usually characterized by a quieter magnetic environment due to a
dropping off human activities; this can be particularly noticeable The total scalar field for the entire four-week period for both
when the holiday falls on a weekday and the nearby environment cities is shown in Fig. 1. While daily variations of the magnetic
surrounding the sensor has an overall magnetic field of higher field in Berkeley appear similar regardless of the day of the week
amplitude during working hours. (weekday fluctuations are similar to weekend fluctuations), the
Finally, an examination of the geomagnetic field measured periodic behavior observed in the Brooklyn weekday data appears
by the respective USGS station closest to each city shows a to stop on weekends and holidays. We note, however, that since the
decrease of 0.2 μT from the first measurement period in March sensor in Brooklyn is placed within a business building, the drop in
2016 to the second period in May 2018. This downward trend of activities within the building during weekends and holidays (e.g.,
the global geomagnetic field has been subject to several stopped elevators, lights off ) represents a direct cause for the drop
studies.19,20 However, in the context of our work, the seasonal/ in magnetic field activities observed by the magnetometer. One
annual variations of the natural magnetic background and its dif- should also note that the measured field is relatively far from the
ferences between the two locations are negligibly small and repre- geomagnetic mean, indicating that the field on the sensor has a
sent a small fraction of the urban variation studied in this work, large contribution from a local source.
i.e., 3% and below. The change in variance during weekdays and weekends is
shown on the top plots of Fig. 2. We note that the dispersion of the
magnetic field in Berkeley is two orders of magnitude less at night
than during the day, dropping from 10
2 to 10
4 μT2, while
nightly variations in downtown Brooklyn remain high with a vari-
C. Sensor locations ance lying above 0.1 μT2 all the time. Decreased amplitude fluctua-
The Berkeley measurements (originally presented in tions in Berkeley occur roughly between 1 and 4:30 a.m. when
Ref. 12) were conducted using geographically separated magne- BART is not in service. We also note that nighttime activities differ
tometers in the city of Berkeley. The four-week data used in this slightly from weekday to weekend; this is a direct consequence of
work were generated by one of the Biomed sensors located in a reduced public transport activities on weekends. In Brooklyn,
residential area 90 m away from the Bay Area Rapid Transit however, the changes between daytime and nighttime variations are
(BART) rail system. The city of Berkeley has about 120 000 resi- less pronounced and the weekend variation has only minor day/night
dents, living primarily in houses with a few low-rise buildings in variability. While a decrease in the anthropogenic activity during
the downtown. A BART line, which crosses the city, is the domi- weekdays is usually observed at around 4–7 p.m., when business
nant source of magnetic field above the natural background activities are reduced, the decrease in the magnetic field only starts to
during daytime.21 be seen at around 11 p.m., thereby suggesting that the magnetic field
In Brooklyn, the Biomed sensor was placed on the 12th floor measured by the Biomed sensor is not solely driven by the occupancy
of the downtown-located Transit Building (370 Jay Street) in one of of the building.
the corner offices of NYU’s Center for Urban Science and Progress. The bottom plots of Fig. 2 show the distributions of magnetic
In sharp contrast to Berkeley that has a population density of about field data for the full dataset as well as for daytime and nighttime
4600 people per square kilometer (2020 census), Brooklyn is over 3 periods. For each city, day and nighttime distributions were fitted
times denser with a population density approaching 15 000 people independently using a skewed Gaussian profile,
per square kilometer, and its downtown constitutes a major trans-
portation axis connecting it to downtown Manhattan. Located A  2 
 2
40 m underneath the sensor’s position, beneath the Transit f ðx; A, μ, σ, γÞ ¼ pffiffiffiffiffi exp 
(x 
 μ) = 2σσ 2π
Building, is the Jay Street-MetroTech subway station, which is   
served by three subway lines at all times and by several additional  þ γ(xp
ffiffiμ)1 erf , (1)
lines during commute hours. σ 2
Due to limited resources, we were only able to stream data
seamlessly from one base station in Brooklyn, located in the Jay where A, μ, σ, and γ correspond, respectively, to the amplitude,
Street building. As the observations are limited to a single-point mean, standard deviation, and skewness of the profile and erf[ ] is
measurement from the building, the question arises as to whether the error function. The best-fit results for each distribution are pre-
or not these magnetic field fluctuations are characteristic of the sented in Table I. Two observations can be made that distinguish
magnetic environment of Brooklyn or, alternatively, if the fluctua- the Berkeley magnetic field variations from Brooklyn. First, we note
tions in the magnetic field are the result of a geographically local- that while day and night time distributions recorded by the sensor
ized set of magnetic sources in the Jay Street building. To address in Brooklyn are centered around a consistent mean magnetic field
this concern, we performed a series of auxiliary experiments using of about 92.8 μT, the mean of both distributions in Berkeley is dif-
multiple portable magnetic sensors (see Sec. III C) and demonstrate ferent with high significance, from a mean of 49.361(6) μT during
that while local processes measured by the Jay Street sensor are pre- the day to 49.925(9) μT at night. The second observation that can
dominant, characteristic observations of the Brooklyn urban mag- be made is the change in skewness of the distribution in Berkeley
netic field can also be extracted from the data. where the nighttime distribution profile sees an increase in
J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264 131, 204902-3
© Author(s) 2022
 17 October 2023 18:58:05
Journal of
Applied Physics ARTICLE scitation.org/journal/jap
FIG. 2. Variance and distribution from Berkeley (left) and Brooklyn (right) data. The daily average variance was calculated over all the days for each consecutive 20 min
time series. The vertical dashed lines on the top figures highlight the transition between day and night times. Nighttime has been set from 1 to 4:30 a.m. for Berkeley and
11 p.m. to 7 a.m. for Brooklyn. The bottom plots show the daytime (red) and nighttime (green) distributions as well as for the full day (blue). A skewed Gaussian was fitted
to both daytime and nighttime histograms independently (see Table I for best-fit results). The blue skewed Gaussian profiles represent the sum of both daytime and night-
time profiles.
skewness compared with daytime variations. In Brooklyn, on the Low-frequency signals for both daytime and nighttime periods
other hand, the distribution remains roughly Gaussian all the time are shown in Fig. 4. We notice that a 20min signal at 8:3 10
4 Hz
with a low absolute skewness of around 1. is observed during daytime in Berkeley, which is known to be asso-
ciated with the BART activities.12 In Brooklyn, a similar signal is
B. Frequency content also observed, but at nighttime. In order to identify this 20 min
periodic signal in the time-series data, we made 100 min averages
In Fig. 3, we show the power spectral density (PSD) at of the daytime Berkeley and nighttime Brooklyn data. Applying a
high-frequency for both Berkeley and Brooklyn data. The drop in high-pass filter with a cutoff frequency at 0.001 Hz allows us to
magnetic field activities in Berkeley at nighttime, identified in the improve the extraction and visibility of the 20 min periodic signal
variance plot (see top panels from Fig. 2), can be explained by (see bottom panels in Fig. 4). While the signal is already visible in
the decrease in low-frequency signals (up to 10 Hz) in the PSD. the average 100 min data for Berkeley, the noise in the unfiltered
We also note a significant difference in amplitude of the power line data from Brooklyn provides greater challenges to identifying the
and other high-frequency signals between both cities (see bottom 20 min signal.
panels from Fig. 3). In Fig. 5, we show a scalogram that demonstrates the richness
of urban magnetic field data. The quiet nighttime perturbations in
Berkeley, previously shown in Ref. 12, are recovered. Using the full-
TABLE I. Best skewed Gaussian fit of daytime and nighttime magnetic field distribu- rate data, one can see how high-frequency ranges are richer in
tions for Berkeley and Brooklyn data. anthropogenic activities. In particular, irregular signals below the
Berkeley Brooklyn power frequency can often be seen and are more prominently in
the Berkeley data.
Params Daytime Nighttime Daytime Nighttime
A 67 980(407) 11 488(654) 158 547(504) 79 559(321) C. Auxiliary field measurements
μ (μT) 49.361(6) 49.925(9) 92.802(14) 92.833(18)
12
σ (μT) 0.281(5) 0.212(18) 0.930(12) 0.617(12) The measurements previously made in Berkeley revealed
γ 1.39(8) −4.61(148) −1.15(5) −0.93(7) coherent magnetic field fluctuations in a geographically distributed
magnetometer array. The significant correlations between stations
J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264 131, 204902-4
© Author(s) 2022
 17 October 2023 18:58:05
Journal of
Applied Physics ARTICLE scitation.org/journal/jap
FIG. 3. High-frequency power spectral density (PSD) from Berkeley (left) and Brooklyn (right) in both logarithmic (top) and linear (bottom) scales. The PSDs were pro-
duced using the full-rate data, sampled at 3960 Hz. The 60 Hz power line and its harmonics can be seen clearly in the linear-scale PSD (bottom panel) and are highlighted
by the thin dashed vertical lines.
FIG. 4. Low-frequency PSDs with 20 min signal extraction. The top panels show the logarithmic-scale PSDs produced using decimated data at 1 Hz. The black arrows
show a 20 min periodic signal (8:3 10
4 Hz) that can be found in the daytime variation from the Berkeley data and nighttime variations in Brooklyn. The bottom panels
show the 20 min periodic signal extracted from an ensemble average of 100 min data regions after applying a high-pass filter with a cutoff frequency at 0.001 Hz.
J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264 131, 204902-5
© Author(s) 2022
 17 October 2023 18:58:05
Journal of
Applied Physics ARTICLE scitation.org/journal/jap
FIG. 5. Wavelet scalogram for a full day (left) and 5 min of data (right) for the first day (Monday) of both Berkeley (top) and Brooklyn (bottom) datasets. The full-day scalo-
gram was achieved using the downsampled 1 Hz data and plotted from the lowest available frequency, i.e., inverse of sampling rate to 500 mHz. The 5 min scalograms
were, on the other hand, produced using the full-rate data, thereby showing frequency content up to the Nyquist limit, i.e., half the sampling rate.
allowed identification of these fluctuations with a “global” magnetic the magnetic field recorded from inside the twelfth floor of the Jay
field, which characterizes the magnetic signature of the city of Street building and identified a correlated signal with a lag time of
Berkeley (or, rather, the broader East Bay). Therefore, in order to 5 min between both stations. Similarly, when recording the mag-
fully determine the signature of Brooklyn, or of New York City netic field from the inside of two buildings located across the street
(NYC) at large, a comparative analysis of in situ data taken from from each other (see the last column in Fig. 7), a noticeable anti-
two different environments must be made. correlated behavior can be observed.
In Fig. 6, we show the behavior of the magnetic field in five
distinct locations throughout Brooklyn. Each measurement was IV. DISCUSSION
acquired using magnetoresistive vector magnetometers manufac-
tured by Twinleaf LLCpwffiffiiffitffihffiffi data sampled at 200 Hz and the noise A. Optimal sensor network distributions
level as low as 300 pT/ Hz at 1 Hz. As one can notice, downtown This work represents an initial demonstration of the potential
Brooklyn is an urban environment with a high diversity of mag- complexity in small-scale magnetic field variability in dense urban
netic field sources (e.g., elevators moving in buildings, cars on environments. Indeed, in the context of the “magnetic field of a
surface streets, subways crossing the Manhattan Bridge), thereby city,” our observations show that the power in high spatial fre-
making the identification of a more global magnetic signature more quency modes is larger in a dense city like Brooklyn; this may be a
challenging. feature of larger cities with a wider variety of magnetic sources.
All field measurements were acquired using two stations to Small-spatial-scale effects must, therefore, be considered when
allow cross correlation between both instruments. Figure 7 demon- designing optimal sensor network systems so they can map out the
strates how challenging the cross correlation between individual spatial variability of magnetic field on multiple scales.
stations that are geographically separated can be. While two stations
placed close to each other (i.e., within a few meters, see the first
column) hold highly cross-correlated data, the information quickly B. Impact of small-scale effects on global field
becomes uncorrelated the further away one station is from another. In this work, we attempt to characterize the global magnetic
However, using low-pass filters, it becomes possible to corre- field of cities, that is, the portion of the magnetic field that has
late signals from different environments. For instance, in the spatial and temporal variation, but is observed to have spatiotem-
second column of Fig. 7, we cross correlate the magnetic field poral correlations over the extent of the city system. A global mag-
recorded from the sidewalk in front of the Jay Street building with netic system can be defined as the set of extended and point
J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264 131, 204902-6
© Author(s) 2022
 17 October 2023 18:58:05
Journal of
Applied Physics ARTICLE scitation.org/journal/jap
FIG. 6. Five samples of magnetic field time series at five different locations in Brooklyn. From left to right: (1) Elevator measurements were taken on the twelfth floor of
Transit Building, (2) subway measurements were acquired from the Jay Street Metro Tech station, (3) Brooklyn bridge measurements were taken underneath the bridge,
(4) street measurements were obtained on the sidewalk in front of the Transit Building in downtown Brooklyn, and (5) the Manhattan Bridge measurements were taken on
top of the bridge from the middle of the walkway.
sources that contribute non-negligibly to its magnetic field. In the extraction of the underlying global city-system field more challeng-
case of a city system, this often contains multiple subsystems, such ing to perform.
as individual buildings and trains.
A subsystem generally contains a geographically localized
(within a volume) set of magnetic sources, which are both extended C. Inferring global properties from local
and point-like in nature. For instance, building-specific fluctuations measurements
represent subsystems within larger systems where the time depen- Local measurements in a dense urban environment may have
dence includes effects from both the structure itself and the behav- periodicities similar to the daily/weekly trends observed in all
ioral signals from the population that is using the structure. In our urban systems, and one could probably argue that the subsystems
work, the base station located within the Jay Street building is are coupled to these large-scale systems. However, periodicities in
subject to fields due to sources with a spatial extent comparable to the global field (e.g., the extended urban environment) are harder
that of the building, as well as any point sources within its subsys- to measure.
tem. Presumably, the field measured with a sensor in the building We further point out that point-source perturbations can be
can also have contributions from other subsystems, such as trains, used to understand buildings in a global field but from an in situ
for example, from the subway station underneath the Jay Street experimental standpoint. However, it is hard to constrain any
building. dynamics using a point-source assumption and a small number of
Spatial and temporal variations in the magnetic field are often sensors. Our interpretation of “subsystem” urban magnetic fields
observed in the data and can be due to a variety of reasons, includ- (i.e., local fluctuations) is that they basically consist of multiple
ing the motion of magnetized objects or time variations in the dipoles (or multipoles) moving in potentially complex ways. A
current generating the magnetic field. While not all point sources, single sensor (even with vector measurements) is unable to
that is, sources confined to a certain small volume, have their mag- uniquely determine a dipole moment/orientation. A minimum of
netic field measurable beyond the vicinity of a few nearby sensors, two sensors is, therefore, needed; the same is true for fields gener-
these sources come with some characteristic radial dependence that ated by line current. An added challenge in understanding the vari-
perturb the larger spatial scale magnetic field, thereby making the ation of a localized source using a few measurement sites comes
J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264 131, 204902-7
© Author(s) 2022
 17 October 2023 18:58:05
Journal of
Applied Physics ARTICLE scitation.org/journal/jap
FIG. 7. Cross-correlated data between two sensors. (1) Sensors placed 7 m apart from each other on the sidewalk; a low-pass filter with a cutoff frequency at 0.1 Hz was
applied to the data; (2) First sensor (blue) placed on the sidewalk and the second sensor (red) in the CUSP office on the 12th floor, low-pass filter applied with a cutoff fre-
quency at 10 Hz; (3) sensors placed at opposite ends of one of the platforms in the Jay Street-MetroTech subway station, low-pass filter applied with a cutoff frequency at
10 Hz; (4) asynchronous measurements from 10 a.m. to 8 p.m., one recorded data (blue) were streamed from the old CUSP office at One MetroTech Center on Monday,
October 9, 2017, while the second set of measurements (red) was made from across the street in the Transit Building on Monday, October 10, 2017, low-pass filter
applied with a cutoff frequency at 0.001 Hz masking periodicity with timescale less than 16.7 min.
from the time dependence of the signals. Further investigations will (nonmagnetic) data, for instance, those from multispectral
need to address to what extent localized fields can, in practice, be cameras.4,5
isolated and identified using a magnetometer network. A specific advantage of magnetometry for urban studies is
The magnetic environment within a subsystem may be highly that it can provide information on the functioning of infrastructure
chaotic. A determination of the large-scale field properties from within its boundaries, but at a distance (e.g., a moving elevator or
local measurements requires an analysis of the statistical variability operating machinery within a building that can be detected from
in the measurements. An exception to a purely statistical approach the outside); therefore, uses might include post-disaster assessment
might be when there is a single dominant source in the local sub- (e.g., vulnerability of partially destroyed buildings), infrastructure
system, which can be subtracted from the local field. monitoring (e.g., assessment of sensorless bridges with short bursts
of observations), monitoring the stability of the power grid (with
instabilities being precursors of outages), monitoring climate and
V. CONCLUSION weather events (e.g., detection of correlated lightning strike signa-
In this pilot study, magnetic signatures obtained in different tures), etc.
urban environments (Berkeley and Brooklyn) were compared. We Some interesting multidisciplinary questions one could
find that there are major differences in magnetic signatures in these address include: How does an anomalous event, such as epidemic
two test cases, for example, the difference in the contrast of mag- or pandemic, affect the urban magnetic signature? Are there signifi-
netic signatures between daytime and nighttime. cant monthly and/or seasonal variations of magnetic signatures?
There are many ways to analyze the rich urban magnetic data. What are the origins of these variations? Are they the same for dif-
As we have shown in this work, some of them allow reducing the ferent cities?, etc. It is the authors’ belief that answering these ques-
complex data stream to a few key parameters that can be used to tions of “comparative urban magnetometry” will teach us a lot
monitor the dynamics of the city. about cities, and this knowledge will eventually translate into tangi-
The results of this work point toward a number of possible ble economic and social benefits.
future directions. For example, this is an extension to sensor There are also technical improvements that can benefit future
networks as introduced in Ref. 12 and correlation with other urban-magnetometry studies. For example, if a measurement is
J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264 131, 204902-8
© Author(s) 2022
 17 October 2023 18:58:05
Journal of
Applied Physics ARTICLE scitation.org/journal/jap
done near a local source, vibrations of the sensor can lead to spuri- 8C. E. Kontokosta, “Urban informatics in the science and practice of planning,”
ous signals. These, however, can be identified by correlating the J. Plan. Educ. Res. 41, 382–395 (2021).
9
magnetic readout with accelerometer data. In fact, the Twinleaf M. Lindgren, M. Gustavsson, Y. Hamnerius, and S. Galt, “Mapping of magnetic
magnetometers that we used are already equipped with such auxil- fields in city environment,” in Electricity and Magnetism in Biology and
Medicine, edited by F. Bersani (Springer US, Boston, MA, 1999), pp. 821–824.
iary sensors. 10J. Fassbinder, H. Becker, and M. van Ess, “Magnetometry at Uruk (Iraq): The
city of king Gilgamesh,” in EGS–AGU–EUG Joint Assembly (European
AUTHOR DECLARATIONS Geophysical Society, 2003), p. 9152.
11H. Becker, Caesium-Magnetometry for Landscape-Archaeology (CRC Press,
Conflict of Interest 2008), pp. 129–165.
12
The authors have no conflicts to disclose. T. A. Bowen, E. Zhivun, A. Wickenbrock, V. Dumont, S. D. Bale, C. Pankow,
G. Dobler, J. S. Wurtele, and D. Budker, “A network of magnetometers for
multi-scale urban science and informatics,” Geosci. Instrum. Methods Data Syst.
DATA AVAILABILITY 8, 129–138 (2019).
13
The entirety of the data and codes used in this work have More details on the characteristics of the Biomed’s eFM-3A sensor can be
been made publicly available. Detailed instructions on how to found at http://citymag.gitlab.io/nuri/biomed.html.14
access the data and reproduce the results presented here can be The noise of the magnetometer is specified at 10 Hz. The 1=f knee for these
devices12 is at around 1 Hz.
found in the online documentation accessible at http://citymag. 15L. Zhivun and V. Dumont, URBANMAGNETOMETER version 1.0.0, 2018.
gitlab.io/nuri/paper. The analysis was performed on 86 GB worth 16See https://twinleaf.com/vector/VMR/ for VMR Magnetoresistive Vector
of magnetic field data using the open-source software NURI,22 spe- Magnetometer.
cifically designed to analyze time-series data produced by our 17There may be a small systematic effect in these measurements: The VMR
urban-magnetometry network. sensors are based on HMC1001 series sensors and have variation over time and
temperature such that the calibration of each axes can be off by as much as 10%,
REFERENCES which may cause the recorded “scalar” measurements to not be purely scalar.18See https://github.com/usgs/geomag-algorithms for Geomag Algorithms.
1G. Dobler, M. Ghandehari, S. E. Koonin, R. Nazari, A. Patrinos, M. S. Sharma, 19P. Alken, E. Thébault, C. D. Beggan, H. Amit, J. Aubert, J. Baerenzung,
A. Tafvizi, H. T. Vo, and J. S. Wurtele, “Dynamics of the urban lightscape,” Inf. T. N. Bondar, W. J. Brown, S. Califf, A. Chambodut, A. Chulliat, G. A. Cox,
Syst. 54, 115–126 (2015). C. C. Finlay, A. Fournier, N. Gillet, A. Grayver, M. D. Hammer,
2V. Lebourgeois, A. Bégué, S. Labbé, B. Mallavan, L. Prévot, and B. Roux, “Can M. Holschneider, L. Huder, G. Hulot, T. Jager, C. Kloss, M. Korte, W. Kuang,
commercial digital cameras be used as multispectral sensors? A crop monitoring A. Kuvshinov, B. Langlais, J.-M. Léger, V. Lesur, P. W. Livermore, F. J. Lowes,
test,” Sensors 8, 7300–7322 (2008). S. Macmillan, W. Magnes, M. Mandea, S. Marsal, J. Matzka, M. C. Metman,
3H. Park and K. B. Crozier, “Multispectral imaging with vertical silicon nano- T. Minami, A. Morschhauser, J. E. Mound, M. Nair, S. Nakano, N. Olsen,
wires,” Sci. Rep. 3, 1703 (2013). F. J. Pavón-Carrasco, V. G. Petrov, G. Ropp, M. Rother, T. J. Sabaka, S. Sanchez,
4L. Duempelmann, B. Gallinet, and L. Novotny, “Multispectral imaging with D. Saturnino, N. R. Schnepf, X. Shen, C. Stolle, A. Tangborn,
tunable plasmonic filters,” ACS Photonics 4, 236–241 (2017). L. Tøffner-Clausen, H. Toh, J. M. Torta, J. Varner, F. Vervelidou, P. Vigneron,
5G. Dobler, M. Ghandehari, S. E. Koonin, and M. S. Sharma, “A hyperspectral I. Wardinski, J. Wicht, A. Woods, Y. Yang, Z. Zeren, and B. Zhou, “International
survey of New York City lighting technology,” Sensors 16, 2047 (2016). geomagnetic reference field: The thirteenth generation,” Earth Planets Space 73,
6G. Dobler, F. B. Bianco, M. S. Sharma, A. Karpf, J. Baur, M. Ghandehari, 279 (2021).
J. Wurtele, and S. E. Koonin, “The urban observatory: A multi-modal imaging 20C. Constable and M. Korte, “Centennial- to millennial-scale geomagnetic field
platform for the study of dynamics in complex urban systems,” Remote Sens. 13, variations,” in Treatise on Geophysics, 2nd ed., edited by G. Schubert (Elsevier,
1426 (2021). Oxford, 2015), pp. 309–341.
7F. Qamar and G. Dobler, “Pixel-wise classification of high-resolution ground- 21A. C. Fraser-Smith and D. B. Coates, “Large-amplitude ULF electromagnetic
based urban hyperspectral images with convolutional neural networks,” Remote fields from BART,” Radio Sci. 13, 661–668 (1978).
Sens. 12, 2540 (2020). 22V. Dumont, NURI version 1.0.0, 2022.
J. Appl. Phys. 131, 204902 (2022); doi: 10.1063/5.0088264 131, 204902-9
© Author(s) 2022
 17 October 2023 18:58:05