[Table of Contents]These articles are scanned and OCRed from old editions of the
ARRL's QST magazine. Here is a list of the
QST articles I have already posted. All copyrights (if any) are hereby acknowledged.
(and its predecessor NACA), and private and public operators have been
monitoring solar events in the optical realm for many decades while
attempting to correlate terrestrial phenomena with it. Auroral light
displays in the extreme polar regions have long been known to be caused
by solar flare and coronal mass ejections (CME). With the advent of
radio, the electrical nature of the upper atmosphere became evident
when static (AM) and long range propagation affected long range communications.
Extreme CME activity eventually was associated with behavior of the
electrical power grid; indeed, massive blackouts and brownouts are to
blame for many. Last but not least came concern for sun-sourced electrons
regarding satellites. More than one 'bird' has been smoked by the sun's
November, on Thanksgiving Day, actually, comet
, which could produce the most spectacular astronomical display
in our lifetime ("the comet of the century"), might actually impact
the sun. Depending on where it strikes (if it strikes), some of the
most violent solar electrons ever could be hurdled toward the earth.
The effects cannot be predicted; we'll all find out real-time what our
fate is. You might want to make sure all your electronics devices have
a full charge, just in case.
See all available
vintage QST articles
432-Mc. Solar Patrol
A Study of Solar Noise in Relation to Radio Propagation
M, Wilson, W4HHK/A4HHK
the spring of 1966, the "big dish" at W4HHK was comparatively idle.
Oscar IV was silent, moonbounce activity on 432 Mc. was nil, and there
were no troposcatter schedules in prospect. Then an item in Sky and
Telescope caught my eye. It told how NASA planned to maintain a solar
flare patrol as part of the space program. Why not a solar patrol on
This seemed like a worthwhile effort. Though the flux
density of radio noise from the sun is measured daily by observatories,
their records are not immediately available to the amateur. Our project
would be a means of keeping track of solar activity on a day-to-day
basis. In addition it would be a way to evaluate the antenna and receiving.
system, and periodically check it.
Until now only an occasional
"look" at the sun had been made. Observation had been inconvenient because
each session meant trips up and down the tower to release and stow the
dish. By the middle of April, 1966, several events had made daily observations
feasible. A stowing device was built that permitted operation from the
ground. A commercial step attenuator was obtained for calibrating solar
noise recordings accurately. Last but not least, a 432-Mc. preamplifier
built with low-noise kMc. transistors made a noticeable improvement
in 432-Mc. receiver performance. Solar Signals
At this point it is desirable to examine the source of the noise
to be observed and measured. The sun is some 92 million miles from earth,
or 8.3 minutes of travel at the speed of light and radio emissions.
It is not a true point source, but for most amateur purposes can be
considered as one. According to Bray and Kirchner,1
be represented as a ring of about one degree angular diameter on the
outside, and about one-half degree on the inside. The sun is a source
of noise on radio frequencies, caused by plasma oscillations and gyro-oscillations
in the solar atmosphere, as well as noise originating in random collision
of electrons. Noise from the quiet sun is of the latter type.2
A solar noise recording made when the sun was relatively quiet is shown
Fig. 1-Noise of the sun on a relatively-quiet day, Dec. 1, 1966.
Noise was running between 3 and 3.5 db. American sunspot count
was 32. A 1-db. calibration dip is seen near the center of the
recording. A burst of ignition noise is seen at the left.
Fig. 2-Block diagram of the 432-Mc. solar-noise measuring setup
Fig. 3-Solar noise level of 5 db. recorded March 3, 1967. Read
chart from right to left. A chart speed of 3 inches per hour
was used at the start, as the dish was centered on the sun for
the recording in the upper right corner of the chart, then changed
to 3 inches per minute for the measurement. Next come three
levels of attenuation, inserted for calibration, and another
15 seconds of solar noise. The dish is then turned away from
the sun, and the quiet-sky noise is recorded. A 50-ohm termination
is then substituted, giving about 0.5 db. more noise than the
Fig. 4-Movement of the sun across the dish is shown in this
3-hour recording made with slow chart speed. The antenna was
aimed at a point where the sun would be later, and left there
to record the rise in noise level as the sun passes. Calibrations
of 4 and 5 db. inserted at the noise peak show the strength
of the solar noise. Numerous spikes are ignition and other noise.
Fig. 5-Correlation between sunspot number and 432-Mc. solar
noise is shown clearly by this graph. Dates when aurora was
observed are indicated by the circled A.
Fig. 6 - 90-second burst of solar noise recorded at 1613
GMT Feb. 23, 1967. Peaks Were some 3.5 db. above the normal
solar noise level.
Paul Wilson, W4HHK, at work on the 18-foot dish used to make
solar noise measurements on 432 Mc. The driven element and its
circular plane reflector are just above the author's head.
Table 1-432 Mc. Solar Noise June, 1966-February, 1967
From Sky & Telescope Magazine.
Noise from a quiet sun appears to have random polarization, but bursts
of high intensity at 432 Mc. are elliptically to linearly polarized.3
Sunspot activity of the current 11-year cycle (Cycle 20) is expected
to peak some time in 1968. The quiet sun looks like an approximate 500,000-degree
Kelvin source, rising to about 1,000,000 degrees K. when disturbed."
The amplitude of solar emission may remain relatively constant for long
periods, and then will be greatly enhanced during a "noise storm." Such
storms are often associated with solar flares and certain geophysical
disturbances, and may last for hours or days.5
flares are sometimes, but not always, followed by aurora some 20 to
40 hours later. The rotation period of the sun is about 27 days, and
there is a tendency for aurora to recur at this rate. Some sunspots
may survive several rotations before disappearing. In a 1951 QST article,6
Moore pointed out that correlation between sunspot number and aurora
is not as great as one might expect. But what about solar noise and
aurora, or sporadic E, F2
or transequatorial v.h.f. propagation?
It was hoped that regular solar noise observation might serve as an
indicator of propagation conditions, and possibly give advance warning
of events such as major auroras. Equipment Evaluation
Besides helping you to keep up to date on what the
sun is doing, regular solar noise measurements are a means of evaluating
system performance, from antenna to receiver output. For given antenna
size, feed line loss and receiver temperature (or noise figure) a certain
minimum amount of solar noise should always be obtainable. WIFZJ described
how a 17-foot parabola and a receiver-feed line combination with a 751-degrees
Kelvin temperature would observe a 3-db. increase in noise when aimed
at the quiet sun.7
Another reference gives the quiet-sun
noise level at 432 Mc. as 21 db. below the receiver noise level when
using a perfect receiver and a dipole antenna. An increase of 10 db.
could be expected from an active sun.8
At the outset
several questions were raised. Was the writer's dish performing up to
specifications? What short-term variations would be observed? Could
solar noise be used for reasonably accurate comparisons of equipment?
Text books on hand didn't have all the answers. Thus one goal of the
project was to resolve some of these questions. The
The equipment used for solar
noise measurement at W4HHK is shown in block-diagram form in Fig. 2.
It doesn't have the simplicity of visual devices such as the Aurorascope9
for checking sunspot activity, but it does work, rain or shine. The
antenna is an 18-foot parabola, with a focal length of 90 inches, and
a diameter-to-focal-length ratio of 2.45. The slide-rule specifications
for 432 Mc. are 26 db. gain over isotropic, and a beam-width of 9 degrees
at the half-power points. It was made by the D. S. Kennedy Company,
and obtained through the Army MARS program. The center is 35 feet above
ground, at a 380-foot elevation. Geographic location is latitude 35
02 48 North and longitude 89 40 04 West. The dish is fully steerable
in azimuth and elevation by an SCR-584 pedestal.
FPS-3 radar platform supports the pedestal and dish. Selsyn indicators
at the control point show the azimuth and elevation in one-degree intervals.
Steering is done by applying d.c. power to the 1/2-h.p. pedestal motors.
The feed for the dish is a horizontal folded dipole, with a
plane reflector 16 inches in diameter. An air-dielectric balun mounted
at the feed matches 50-ohm coax to the balanced 300-ohm dipole. A 38-foot
length of foam RG-8 connects the balun to a 60-foot run of 7/8-inch
Heliax, from the base of the pedestal to the shack. Overall line loss
is about 2.25 db., according to published figures.
antenna is also used for transmitting, a coaxial relay, not shown, is
employed for transmit-receive switching. It is worth mentioning that
after several months of operating a kilowatt transmitter alongside,
no measurable change in receiver noise figure has been observed. The
feed line is disconnected manually when the equipment is not in use,
as a precaution against damage from lightning and heavy static.
Each r.f. stage is in a separate aluminum box, with power supplied
by individual 9-volt batteries. The first r.f. amplifier uses a KMC
n.p.n. experimental transistor, with a factory-measured noise figure
of 2.8 db. (for the transistor). The second uses a TIXMO5 p.n.p. transistor.
Gain per stage is about 9 db. Each has the common-emitter configuration,
and is unneutralized. Both are stable, even in the absence of input
load. Both have double-tuned input circuits, to minimize response to
out-of-band signals. In-band signals are not a problem, as the nearest
432-Mc. station is 70 miles distant! The second r.f. stage feeds a 1N21F
mixer, which works into a 50-Mc. converter at 49.5 Mc., converting again
to 7 Mc., and followed by a 75A3 receiver and chart recorder. The converter
and 75A3 are always kept on standby when not in use, to minimize trouble.
The r.f. amplifiers and 432-Mc. converter were built by the writer.
A constant-voltage line transformer was found to be necessary
to maintain reasonably good gain stability. Without it a recording of
a constant-amplitude signal would vary considerably, especially during
lengthy recordings. Apparently the main cause of this was heater-voltage
change, as the converter plate voltages are all regulated. The converter
seemed more sensitive to line-voltage change than did the 75A3. This
and other problems were ironed out before regular observations began
on June 1, 1966.
A 20-db. fixed pad is used at the converter
output and a 10-db. fixed pad at the receiver input, with the step attenuator
in the 50-ohm line between them. This was done to insure operation of
the attenuator at its design impedance. The 75A3 mode switch was modified
to permit reception with the b.f.o. and a.v.c. off. The receiver 500-ohm
output is connected to a bridge rectifier, which drives the Esterline-Angus
1-ma. strip-chart recorder. Rectifier output is not linear, crowding
at the low-signal end of the scale. Esterline-Angus chart paper, type
132020, that closely matches the rectifier response, was obtained surplus.
A chart speed of 3 inches per minute has been found a good compromise
between easily-read recordings and paper conservation. For long-duration
recordings a speed of 3 inches per hour has been used.
Radio observatories measure
the flux density of solar radio emissions by using the unit of 10-22
watts per square meter per cycle per second. The method used by the
writer measures the ratio of solar noise to quiet-sky (background) noise
in decibels. The results of a typical measurement, made as described
below, are shown in Fig. 3. The feed line is connected to the first
r.f. amplifier and power is applied to both r.f. stages. Converter crystal
mixer current is adjusted to read 0.5 ma. The 75A3 S-meter is observed
to verify that two S-units of converter noise is read with the receiver
in the a.m. mode. The 75A3 a.v.c. and b.f.o. are turned off, and the
a.m. and c.w. limiters are disabled. Audio gain is advanced to 75 percent
of full volume, and the" r.f." gain is reduced until the chart recorder
reads about one-third scale, as seen at the right edge of the recording.
The dish is then steered toward the sun, and the azimuth and
elevation controls are adjusted until maximum noise is indicated on
the recorder. Audio gain is readjusted for a meter reading of about
0.9 ma. Chart speed is shifted from the 3 inches per hour, used while
steering, to 3 inches per minute, and solar noise is recorded for about
30 seconds. This is the first 5-db. peak at the right. At this point
several steps of attenuation are inserted to calibrate the chart, usually
beginning with the 3-db. step. Normally the 3-, 4-, and 5-db. steps
are each recorded for 15 to 30 seconds, depending on the undesired responses
that may be present, such as from ignition or radar. In this instance
steps of 4, 5 and 6 db. were used. The audio is monitored on a speaker,
and the chart trace is observed, to make certain that interference is
not spoiling calibration and reception in general.
the last level of calibration the attenuator is switched out of the
circuit, and full solar noise is again recorded on the chart. If nothing
has changed (dish heading by wind gust, solar noise level, etc.) the
reading will be about the same as at the start. Though the sun is not
tracked during the measurement, the noise amplitude should remain constant,
because of the antenna beamwidth and the shortness of the measurement
The dish is then steered away from the sun to a point
in the northwest sky, the quietest heading for this location. Quiet-sky
noise level is recorded for 30 seconds to a minute, again making certain
that unwanted noise does not obscure the quiet-sky noise level. As each
measurement is being made the GMT date-time, calibration steps, and
other information are noted quickly on the moving chart. Three to six
measurements are made each day in this manner, the entire operation
taking 30 minutes or so. More measurements are made if conditions warrant.
Recordings are examined at the end of the session, or later, and solar
noise readings determined. Periodically the noise of a 50-ohm termination
at the receiver input is recorded, in addition to quiet-sky noise, to
confirm receiver performance. Termination noise is usually about 0.5
db. higher than quiet-sky noise, as seen at the left edge of Fig. 3.
All readings are entered in the solar noise log, and the section
of chart with the highest reading is mounted and filed. The highest
reading is of particular interest because it shows maximum solar activity
during the observing period, and because it is usually the most accurate.
Any errors caused by steering, noise, and equipment failure tend to
A solar noise recording made while the sun
moved across the dish is shown in Fig. 4. This was obtained by pointing
the dish at a point in the sky where the sun would be at a later time.
The Greenwich hour angle and declination of the sun were determined
from the Nautical Almanac.10
This information was converted
to azimuth and elevation by tables found in a Hydrographic Office publication.
Receiver gain was adjusted so that background noise deflected
the recording pen about one-third scale, and the chart speed set for
3 inches per hour. Recording was started about an hour in advance of
the time for which the dish was positioned. When the sun came into view
the solar noise rose from the background level, peaked, and then slowly
fell. Insertion of suitable levels of attenuation, in this instance
4 and 5 db., provide a check on the performance and establish the accuracy
of the recording. Only the main lobe is seen at present levels of solar
activity. Minor lobes of the antenna are too far down to show.
The sun's position can be used to "boresight" an amateur antenna.
The dish is peaked carefully on the sun at a time and date listed in
the Nautical Almanac, and the azimuth and elevation indicators are adjusted
to read the azimuth and elevation given in the tables for that moment.
This requires knowing the latitude and longitude of the antenna site
precisely, and making the test at the correct time.
A Quiet Summer Ends Dramatically
Noise varied monotonously
between 3 and 4 db. through the summer of1966. QST reported an aurora
on July 7-8, and Sky and Telescope mentioned a big flare on July 11,
but solar noise recordings on or about these dates gave no hint of anything
out of the ordinary. Obviously measurements made during a brief period
daily do not tell the whole story, and short-term events can be missed
entirely. By late August the daily routine of shooting the sun had become
just that: a routine. Interest was lagging.
Then on August 29
a reading of 4.7 db. was obtained. Sunspot records indicate that it
was probably higher the previous day, when a Class 3 flare and associated
sudden ionospheric disturbance (SID) occurred, but the 28th had been
missed by this observer. Here was the big change sought all summer.
Would there be an aurora? Indeed there was! QST reported auroral v.h.f.
communication on Aug. 29, 30 and Sept. 1 through 4. The evening display
of Sept. 3 was the best aurora in years, according to Sky and Telescope,12
and amateur results on 6 and 2 bear this out. An inconvenient work schedule
and too much reliance on "N7" transmissions by WWV caused the writer
to miss the big event!
This called for careful monitoring of
the sun for the remainder of the month, with results shown in Fig. 5.
Solar noise is plotted by the solid line, and American sunspot numbers13
by the broken line. Aurora dates are identified by the symbol, A. Note
the high level of solar noise and sunspot number on Aug. 29, followed
by a rapid drop to the low point Sept. 3, the period of auroral activity.
Solar noise peaks were observed again on the 16th and 20th, and corresponding
sunspot maxima occurred on the 17th, 19th and 21st. Both started dropping,
and reached a low point on the 27th and 28th. No radio aurora was reported
around the noise peaks of the 16th and 20th, but visual sightings were
reported by Sky and Telescope. QST reported 50-Mc. DX Sept. 21.
The graph indicates that 432-Mc. solar noise and sunspot numbers
are related, and vary in the same general way. One can see readily that
there is also a correlation with aurora. It would also appear that during
September the sunspot maxima and minima tended to trail solar noise
highs and lows by a day or so. This is not always the case. Records
for eight other months show noise peaks and sunspot maxima coinciding,
and occasionally the sunspot count has peaked a day in advance of the
noise. Fall and Winter
solar noise was measured every day but three. Auroral propagation on
50 and 144 Mc. was reported by W1HDQ Oct. 3, but noise and sunspot figures
for the period didn't indicate it. Presumably this was a 27-day recurrence
of the disturbance of early September. There was a noise peak on the
20th that coincided with maximum sunspot number (82) for the month.
The Ottawa Algonquin Radio Observatory report for October shows solar
flux at 2800 Mc. also peaked this day, but there were no radio aurora
reports. Sky and Telescope confirms that there was a small aurora on
the 20th. Transequatorial propagation on 50 Mc. was observed by PY5GK
on 14 days in October, including the 21st, but not the 20th.
Shortened daylight hours reduced our opportunities for daily observations
in November and December. Solar noise high for November was only 4.1
db., but activity picked up in December. High noise reading for the
last month of the year was 4.75 db. at 1900 GMT, Dec. 13. There were
no reports of aurora. (Statistically, aurora is rare in this hemisphere
in December - Editor). Sagamore Hill Observatory reported an outstanding
solar radio emission on 2695 Mc. at this time, and American sunspot
numbers were at their month high on the 13th. A week later noise readings
reached a low point, and then began an upward climb toward the end of
the month. A good indicator of general solar activity, for the writer,
has been the number of days per month that solar noise reached 4 db.
or higher. December had 13 such days, compared with only 5 in November,
for almost the same number of observing days.
were made the first few days of 1967, but when they were resumed Jan.
5 a whopping 5.1 db. was recorded, and it was up to 5.6 db. the following
day. The alert was sounded, and participants in the V.h.f. Sweepstakes
the weekend of Jan. 7-8 don't need to be told what happened. Working
schedules the next few days prevented observations, but on Jan. 10 a
high of 4.75 db. was recorded. Three days later, Jan. 13-14, another
auroral session was reported. Like that of Jan. 7-8, it was violent
and widespread, yet January is normally a relatively quiet month for
aurora. Solar activity continued high, and every day when observations
could be made found noise exceeding 4 db.
The first day of February
a high noise reading of 4.6 db. was obtained, and on the 4th an all-time
high of 6 db. was logged! Three days later, on the 7th, another aurora
was enjoyed by the v.h.f. gang. Activity continued high, with readings
of 4.7 db. on the 7th and 4.5 db. on the 8th. Readings then settled
down to about 4 db. There was no indication of aurora on Feb. 16, about
the time that a recurring disturbance might have been expected. It would
appear that some of the "repeaters" (auroras recurring on the 27-day
solar rotation cycle) are not associated with solar noise readings,
or the related noise peaks are of short duration and are easily missed.
Readings continued at 4 db. until Feb. 23, when a moderate rise
to 4.5 db. was observed. Something new to the writer was recorded on
this date: a noise burst some 3 db. above the average level, lasting
about 90 seconds. See Fig. 6. On the 25th solar noise was peaking at
4.6 db. A minor aurora was reported on the 27th. The
Highlights for the period June 1, 1966,
to March 1,1967, are shown in Table I. The monthly mean value for American
sunspot numbers is included to show the relationship of sunspot count
to solar noise. No doubt results might have been different if noise
readings had been made every day, or for longer periods of time. Nevertheless,
the figures do show that solar radio noise and solar activity have been
increasing since June, 1966, and are still climbing. During the entire
period there were no changes made in equipment or measuring procedure.
Any known causes of error have been noted. Various r.f. amplifiers were
tested and compared from time to time, but only the units described
were used for the recorded data. Periodic checks of tubes and regular
measurement of noise figure was done in an effort to obtain consistent
results. The unchanging performance of the transistor r.f. stages contributed
much to the system reliability.
Our 432-Mc. solar patrol has
provided some answers to questions raised at the beginning, but more
information is needed. The receiving system has performed about as expected,
except that the azimuth beamwidth of the antenna is almost double what
it should be, probably because an 18-foot dish is "small" in terms of
wavelength at 432 Mc., and thus the dipole does not simulate a point
source, as it must for optimum performance with a parabolic reflector.
Vertical beamwidth is correct, about 9 degrees.
A minimum of
about 3 db. of solar noise was always obtained, even during periods
of little or no solar activity. When used with care, solar noise readings
under quiet conditions can be used to evaluate systems, and make comparisons
between antennas, r.f. amplifiers and so on. By comparing readings with
other amateurs, a v.h.f. operator can determine if his system is "in
the ball park," which can be very helpful in setting up for moonbounce
work, for example. But beware! Solar noise measurements are subject
to error and variations, just as noise figure measurements are.
Regular observations do give an indication of solar activity, but
the solar noise level is not an exact indicator of sunspot number. For
example, the AA VSO daily sunspot count has varied from 1 to 89 for
noise readings of 3.5 to 3.6 db. Sunspot number is usually in the range
of 1 to 50 for a 3.5-db. noise level. Auroras are associated with high
solar noise levels, the .( magic number" for the writer apparently being
a minimum level or about 4.7 db. This level and higher has been observed
on ten separate occasions, and seven individual auroras, apparently
related, have been reported. Aurora appears more likely when high noise
levels are observed on two or more successive days. Delay between noise
peak and aurora has been one to three days, though some occurred on
the same date. More observations should give a clearer picture, and
improve forecasting. Thus far the relationship of F2
transequatorial 50-Mc. propagation to solar noise is not well defined,
though this could improve when the F2
layer rises into the
50-Mc. region more consistently.14
Sporadic-E seems hardly
related, if at all.
The equipment and techniques described should
not be considered the ultimate or best way. This is simply an account
of a 432-Mc. experiment, in the hope that it will shed some light on
the subject as it relates to amateur v.h.f. enthusiasts, and perhaps
stimulate others to do similar work. Possibly observations by a number
of stations could be coordinated to provide an auroral warning service
for the v.h.f. community, during the peak years of Cycle 20.
The writer owes special thanks to Third Army MARS for the dish and
related equipment; to WIHDQ for propagation reports; to Sky and Telescope
for auroral reports; to the American Association of Variable Star Observers
for sunspot data; to K2TKN, W3GKP, W3O11, WA0IQN and others for their
assistance in making this project possible.
Late Report On the
afternoon of May 23, W4HHK recorded solar noise on 432 Mc. in excess
of 15 db. Visual observation showed exceptionally large spots approaching
the center of the solar disk. On May 25 came the most widespread auroral
display in recent years. V.h.f. communication via the aurora was reported
as far south as Sarasota, Florida, and auroral contacts were made in
Southern California for the first time. The disturbance continued through
May 30, with peaks on the 25th, 28th and 30th. For more details. see
The World Above 50-Mc., July QST.
1. Bray and Kirchner,
"Antenna Patterns from the Sun," July, 1960, QST, p. 13.
NBS Monograph 80, "Ionospheric Radio Propagation," April 1, 1965, p.
38. Price $2.75, from Supt. of Documents,
Washington, D. C. 20402.
3. M. H. Cohen, Cornell University
"Measurement of Solar Radiation at 430 Mc ." Quarterly Status Report,
Feb. 1-July 31, 1965.
4. "The World Above 50 Mc.," October,
1965, QST, p. 112
5. Reference 2, p. 43.
"Aurora and Magnetic Storms," June, 1951, QST, p. 16.
World Above 50 Mc., October, 1960, QST, p. 112.
8. "IT &
T Reference Data for Radio Engineers, 4th Edition, p 764.
Tomcik, "The Aurorascope;" July, 1964, QST, p 43.
Almanac for 1967," Supt. of Documents, Washington, D. C. 20402. Price
11. "U. S. Navy Hydrographic Office Tables of Computed
Altitude and Azimuth," Latitudes 30 to 39 degrees, inclusive.
Pub. No. 214, Vol. IV. Order volume applicable to your latitude, from
Supt. of Documents, Price $3.00.
12. "Auroral Activity Increases,"
Observer's Page, Sky and Telescope, December, 1966, p. 380.
American sunspot numbers derived by the American Association of Variable
Star Observers (AAVSO), solar division,
reported monthly in Sky and Telescope.
14 The correlation
between solar activity and extreme peaks of F2-layer m.u.f.
was first observed by G6DH in the 1930's. For
his slant on it, relating to 50-Mc. DX, see "Any DX Today?" January,
1948. QST, p. 27. The correlation with transequatorial
.50-Mc. DX is indicated in a summary of 1950 work by South American
50-Mc. operators, in "The World Above 50 Mc.,"
May, 1950, QST, p. 49.
Posted September 24, 2013