Correlation 29(1) 2013 53

A Signal from Saturn in Daily Temperature Data

by
Bruce C. Scofield
Abstract

Historic annual prognostications published by Renaissance scientists Thomas Digges,

Tycho Brahe and Johannes Kepler, and “farmers” almanacs published in Europe and New
England, contained weather forecasts based on geocentric planetary configurations. The
syzygy between the Sun and Saturn, in this tradition thought to produce cold weather, is
here tested using daily surface temperature records. A daily time-series from Central
Massachusetts from 1971-2000 shows lower temperatures within one day before the
geocentric Sun-Saturn opposition with more extreme departures when the opposition
occurred near the equinoxes, when Saturn was at perigee, and when the Moon was an
addition to the syzygy. The “Saturn effect” was found in other Northeastern US datasets
including one of 120 years from Amherst, MA. Two other long daily temperature datasets,
Central England and Prague, and also other datasets from temperate to subarctic latitudes
in the Northern and Southern hemispheres were examined, and in most cases similar results
were found. Significant decreases in temperature within one day of the exact opposition
occurred more frequently when the opposition was near the equinoxes or when the Sun and
Saturn shared the same declination. These observations lead to a hypothesis: a
gravitational resonance between the Sun and Saturn slightly increases atmospheric tides.
When the opposition is near the equinoxes the gravitational forces are orthogonal to the
surface at the equator and parallel to the surface at the poles; the consequence is
atmospheric compression at the poles with movement of cold air southward.

Introduction
Extraterrestrial influences on climate are generally associated with variations in solar
radiation reaching the Earth due to orbital parameters, i.e. Milankovitch cycles, and the
Schwabe, Hale, and other solar cycles. A number of previous studies also indicate a
possible lunar effect on precipitation, thunderstorm frequency, diurnal pressure changes,
hurricanes, among other meteorological phenomena. i No specific mechanism is currently
accepted in regard to correlations between the Moon and surface weather, although most
assume a lunar-driven tidal mechanism on the atmosphere or magnetic field may be
involved. Correlations between astronomical bodies other than the Sun or Moon and surface
weather, however, have not been studied. Historic annual prognostications written by
Renaissance scientists Thomas Digges, Tycho Brahe and Johannes Kepler, and “farmers”
almanacs published in Europe and New England, typically contained weather forecasts
based on geocentric planetary alignments (aspects) associated with meteorological
phenomena including heat, cold, rain, winds, storms, etc. From this tradition, one
established family of alignments for predicting cold weather are those between the Sun and
Saturn. This study tests primarily the opposition of the Sun and Saturn, and in certain cases
the conjunction. The opposition (Saturn-Earth-Sun) and conjunction (Earth-Sun-Saturn)
form straight lines in celestial longitude, a syzygy, in both geocentric and heliocentric
coordinate systems.

A signal from Saturn: Bruce Scofield

54 Correlation 29(1) 2013

Data Sources and methods

To test for evidence of a correlation between Sun-Saturn syzygies and cold temperatures a
study was designed utilizing daily surface temperature records. Initially, a daily surface
temperature time-series from Central Massachusetts for the period 1971-2000 was
converted to daily anomalies from the period mean was also obtained. ii Secondly, station
data from ten other locations in Northeastern United States for the period 1971-2000, and a
longer dataset, 1885-2005, from one of these sources (Amherst, MA) were prepared. iii
Thirdly, two long daily temperature datasets of over 200 years, Central England and Prague,
and 27 other daily weather datasets from temperate to subarctic latitudes in the Northern
and Southern hemispheres were prepared in the same way. iv Ephemeris data for Sun-Saturn
syzygy was calculated using software and published tables. v Given that the orbital cycle of
Saturn is approximately 30 years, daily temperature records of 30 years, or multiples
thereof, are appropriate for this study as they eliminate any seasonal distortions that records
of other lengths may accrue.

The method of data analysis used in the study compared samples of surface temperature
data (and in some cases temperature data at a geopotential height of 500 mb) centered on
Sun-Saturn syzygies with controls taken at regular and random intervals in the time-series.
For each individual study, daily temperature anomalies were extracted from prepared time-
series data in 21-day blocks centered on the exact date of each Sun-Saturn syzygy
designated as day 0. Although the precise time of a Sun-Saturn syzygy could be located
anywhere within the 24 hour period of local time on day 0, the study used only 24-hour
daily average temperatures, not hourly data, and day 0 was treated as the midpoint of the
21-day samples. All 21-day samples were then stacked and a mean for each of the 21 days
of the total set was calculated. Temperature anomalies on days close to day 0, the exact
syzygy, and the amplitude of any variation were noted. The data was then sampled
according to seasonality and also to the syzygy’s location relative to the equinoxes, in both
celestial longitude and equatorial declination, for any amplification or diminishing of
apparent effect. Confidence intervals used for estimating significance relative to the mean
were calculated for each day of a test, and also for the entire time series. Control tests were
designed using the same methodology applied to regularly spaced once-yearly dates,
randomly generated dates, dates equidistant from the test dates, and tests of alignments
involving planets other than Saturn. Some tests focused on periods when the Sun and/or
Saturn were at significant astronomical locations, such Saturn at perigee or Sun and Saturn
near the equinoxes.

Results
The Central Massachusetts 1971-2000 test data displays a mean daily temperature low on
the day before the Sun-Saturn geocentric opposition. The 500 mb height for the same period
in Central Massachusetts was also tested and shows an similar pattern with the low a day
earlier. The sums of temperatures for days -2, -1, and 0 in this dataset, 0 being the syzygy,
are lower than those of days -3, -4, and -5 in 23 of the 30 years. A composite of ten
Northeastern United States station records for the period 1971-2000, and a 121-year record
from Amherst, MA, show a 3 degree drop in temperature from 5 days before the syzygy to a
low on day -1, the day before. vi

A signal from Saturn: Bruce Scofield

 Correlation 29(1) 2013 55

Figure 1a

Figure 1b

Figure 1a: Saturn-Sun geocentric oppositions 1971-2000 Petersham, Central

Massachusetts; surface temperature (solid line) and 500 mb height (dashed line) both
expressed as anomalies from the dataset mean. Figure 1b: Saturn-Sun geocentric
A signal from Saturn: Bruce Scofield
56 Correlation 29(1) 2013

oppositions 1971-2000, a composite of ten Northeast stations (solid line) and temperatures
at Amherst, Massachusetts 1885-2005 (dashed line) expressed as anomalies from the
dataset mean. The correlation coefficient r for the two sets of variables in Figure 1b =
0.876. The roughly 3 degree drop in temperature from day -5 to day -1 is not quite one
standard deviation from the mean, which for both datasets is about 4 degrees. In the
Amherst 1850-2005 dataset mean temperatures declined between day -7 and the day of the
syzygy 75% of the time.

Control tests were performed on data for the period 1971-2000. The test methodology was
applied to randomly generated dates derived from online software (www.random.com), one
date for each year of the time-series which produced a sample size consistent with the
study. Random dates were also generated within the 30-year period regardless of year, and
also dates equidistant from the Sun-Saturn opposition for each year were calculated. These
controls display the general noise found in the data but did not display a consistent low
temperature correlation with day 0 of the test methodology. Controls were then divided into
two sets, the first and second half of the period, and also by alternation of years, the later
eliminating the temperature gradient which rises during the period. Both methods produced
very weak or negative correlation when compared to each other suggesting the signal was
inconsistent through the period examined. The same method applied to the actual test itself
showed greater similarity between the halves quantified by a high correlation coefficient,
suggesting that the signal was constant through the period. Another control tested Sun-
Jupiter oppositions for the same period where an increase in temperature during the 21-day
period occurred. The same data tested for precipitation (not shown) showed a maximum
centered around day 0 which is consistent with almanac lore.

Figure 2a

A signal from Saturn: Bruce Scofield

 Correlation 29(1) 2013 57

Figure 2b

Figure 2c

A signal from Saturn: Bruce Scofield

58 Correlation 29(1) 2013

Figure 2d

Figure 2e

A signal from Saturn: Bruce Scofield

 Correlation 29(1) 2013 59

Figure 2f

• Figure 2a: Comparison of two randomly generated datasets for the 30-year period
tested. The correlation coefficient r = -0.037.
• Figure 2b compares the Petersham 1971-2000 temperature record of Sun-Saturn
oppositions in two equal parts derived by yearly alteration of data points. Part 1,
the solid line, is the mean of dates beginning with 1971 then skipping every other
year. Part 2, the dashed line, is the mean of the remainders. The correlation
coefficient r = 0.578, evidence of some correlation.
• Figure 2c and 2d are equal sections of the two randomly generated controls
(shown in Figure 2a) divided by the same method used in Figure 2b. The
correlation coefficent r for Figure 2c = -.402 and for Figure 2d = -0.663,
indications of some negative correlation. Figure 2e displays the results of
performing the test on Sun-Jupiter oppositions. Figure 2f displays Sun-Saturn
opposition temperature data from the Amherst 1885-2005 data and data from a
control dataset built on samples taken annually on the first of every succeeding
month through the 121-year period.

A weak signal associated with the Saturn-Sun syzygy may have beeen detected in these
temperature records. If gravity is the force modulating the atmosphere during a Sun-Saturn
syzygy, it follows that the additional mass of the Moon should register in the data. The
seven Sun-Saturn geocentric oppositions occurring on the same day as a new or full moon
for the 1971-2000 period for Amherst were taken as a sample and temperature data for
these events were stacked and a mean calculated. In the Amherst 121-year record a sample
of the 9 Sun-Saturn oppositions that occurred within 12 hours of a full or new moon was
likewise taken. These two small samples are graphed in figure 3a, the correlation coefficient
r = 0.595. The amplitude of the temperature drop from day -10 to day 0 is well over one

A signal from Saturn: Bruce Scofield

60 Correlation 29(1) 2013

Figure 3a

Figurre 3b

Figure 3a: Solid line: New and Full Moon events (7) occurring < 1 day from Sun-Saturn
A signal from Saturn: Bruce Scofield
 Correlation 29(1) 2013 61

oppositions; 1971-2000 Amherst. Broken line: New and Full Moon events (9) occurring <
0.5 day from Sun-Saturn oppositions; 1885-2005 Amherst.
Figure 3b: New and Full Moon events occurring < 2 days from Sun-Saturn oppositions (7)
and Sun-Saturn conjunctions (4); 1971-2000 Amherst.

standard deviation which is roughly 4.5 degrees from the mean, calculated for either
individual days or the entire dataset. Temperatures were below the mean for six days before
and after the exact opposition. In Figure 3b the eleven Sun-Saturn syzygyies, seven
oppositions and four conjunctions, occurring within 2 days of new moon or full moon for
1971-2000 are graphed.

Two longer datas sets from Europe, the 237-year Central England Temperature Series and a
230-year blended Prague temperature time series from European Climate Assesment, were
tested using the same methodology. Results were also sorted by months and by the
declination of the syzygy. Unlike the Northeastern United States weather data, the
opposition (Figure 4a) does not show an observable correlation between the opposition and
low or declining temperatures – except for those that occurred in April and October in
Prague and those within 3 degrees of the equinoxes in England and Prague (Figures 4b and
4c). This finding implies low declination may be an influencing factor as these points are
the intersection of the celestial ecliptic and equator. In the sample used for Figure 4c,
oppositions near the vernal equinox show a more precise correlation than those near the
autumnal equinox. Johannes Kepler, who lived in Prague for many years, regarded the Sun-
Saturn conjunction as a good example of the efficacy of planetary weather forecasting and
so, in deference to Kepler, this syzygy was investigated. Conjunctions in the full Central
England and Prague datasets showed no significant temperature decline, but the Sun-Saturn
conjunctions that occurred within 30 days of an equinox show a correlation.

Figure 4a

A signal from Saturn: Bruce Scofield

62 Correlation 29(1) 2013

Figure 4b

Figure 4c

A signal from Saturn: Bruce Scofield

 Correlation 29(1) 2013 63

Figure 4d.

Figure 4a: Temperature data for CET and Prague for 230+ years of Sun-Saturn oppositions.
Figure 4b: Prague Sun-Saturn oppositions that occurred in April (n=19) and October
(n=18). The 3 degree C decline of temperature during April approaches the standard
deviation of 3.7 degrees of the entire datset. Figure 4c: Sun-Saturn opppositions, CET and
Prague, (n=6) limited to 3 degrees from equinoxes. Figure 4d: CET (n=77) and Prague
(n=73) conjunctions within 30 days of equinoxes.

Temperature studies centered on Sun-Saturn oppositions and conjunctions were done for
selected locations at higher northern and southern latitudes around the Earth. In North
America, datsets for Fairbanks, AK, Yellowknife, Canada, Saskatoon, Canada, Ottawa,
Canada, Mount Washington Observatory, New Hampshire, Halifax, Canada, Igaluit, Baffin
Island, Canada, and Tasiilaq, Greenland. In Europe and Asia datasets for Oslo, Norway,
Moscow, Russia, Novosibirsk, Russia, and Jakutsk, Russia were analyzed (graphs are not
shown). In general, the opposition appears to correlate more consistently with lows near day
0 in the eastern North American datasets and Greenland. It appears as if cold temperatures
move southwest across North America with the lowest temperatures reaching the
Northeastern United States on or about day -1. Correlation with the opposition in Europe
and Asia is weaker or non-existent. The conjunction seems, in Europe, at least, to
corroborate Kepler’s observation that it marks warmer temperatures.

In regard to the southern hemisphere, limited land mass at latitudes near or greater than 40
degrees presents two problems in the analysis of weather data required for this study. The
first is that the regions of this category, the southern tip of South America and New
Zealand’s South Island, have relatively limited daily weather datasets and many of these are
not complete. The second is that nearly all the stations are near the coast and the climate is
therefore maritime and not continental. In spite of these constraints, data was obtained from

A signal from Saturn: Bruce Scofield

64 Correlation 29(1) 2013

the Online Climate Data Directory of the NOAA Satellite and Information Service for six
southern hemisphere stations: Christchurch, NZ, Invercargill, NZ, Nelson, NZ, Neuquen
Aero., AR, Rio Gallegos Aero., AR, Puerto Montt, CH. Missing data was then replaced
with a mean calculated from the full length of the dataset. A composite of the three South
American stations shows a nearly 1.5 degree temperature drop following day +1. This is not
found in the composite of the New Zealand stations. A graph of all six southern hemisphere
stations combined appears to register a cooling trend lasting for several days following the
alignment, this being most obvious in the South American data alone. (Data not shown).

4. Discussion

The opposition between the Sun and Saturn, which occurs when the Earth is closest to
Saturn, appears to correlate in some datasets with a small drop in temperature within a day
of the precise alignment, but this deviation is amplified when the event occurs near the
equinoxes, in proximity to a lunation (new or full moon), or when Saturn is at perigee.
Coincidence with a lunation and Saturn at perigee suggest that gravity may be force behind
this observation. Gravity may also be a factor when the Sun is at the equinoxes, a time of
the year when two geophysical conditions are operative. First, a seasonal reversal begins
that is driven by the rapidly changing day to night ratio and redistribution of incoming solar
radiation. At the vernal equinox warm equatorial air is moving north, the reverse occurs at
the autumnal equinox. Air mass movement and circulation changes at the equinoxes cause
weather conditions to become more volatile. Secondly, for a body orbiting on the plane of
the ecliptic, only positioning at the equinoxes places it at right angles to the axial pole. The
tidal effect of an ecliptic body on the Earth’s atmosphere, when at the equinoxes, produces
a bulge at the equator and a flattening at the poles, a process by which colder air from the
higher latitudes could be brought to the surface and pushed southward. Around the time of
the equinoxes, when the balance between tropical and polar air is more fragile, this effect
may be enhanced. Lunar tidal effects on the Earth’s atmosphere are known and correlations
have been reported between lunar declination and tropospheric circulation patterns. vii

Differences in regional temperature response to a gravitational effect from Saturn may be

due to geography – the proximity and size of land mass between the source of the dataset
analyzed and the poles. It is possible that a weak gravitational effect from Saturn, somehow
resonating with the tidal force of the Sun, is capable of compressing air masses near the
poles which then push southward to the stations used in this study. The calculated
gravitational force of the planet Saturn on the Earth’s atmosphere, based on Saturn’s gravity
and that of the mass of the Earth’s atmosphere, is weaker compared to that of the Moon and
Sun, by four and six orders of magnitude respectively, even when the planet is at perigee. If
there is such a tidal effect from this alignment to the Earth’s atmosphere, then some
amplification of the signal must be occurring. Tides in the Sun’s atmosphere driven by
planetary gravitational fields with resonance amplification have been proposed and it is
possible that the tests performed in this study may be evidence of a similar terrestrial
effect.viii If there is a gravitational factor behind the cooling trend at the time of Sun-Saturn
oppositions, then it follows that a similar trend may also occur at the time of the full Moon.
Several tests in regions where Sun-Saturn oppositions showed a strong correlation with a
cold trend did not show a similar response to full Moons, however. This may be due to the
rapid motion of the Moon. Using a range of 3 degrees celestial longitude before and after
the opposition, the lunar alignment with the Sun is relatively brief (hours) in comparison
A signal from Saturn: Bruce Scofield
 Correlation 29(1) 2013 65

with the several days duration of Sun-Saturn opposition.

The findings of this study may be of use in modern meteorology. Knowledge of Sun-Saturn
alignments, in combination with modern meteorological data, could improve forecasting,
both short and long range. Further studies of other traditional astrometeorological
techniques may broaden this approach. A more detailed study of geopotential height
changes over the poles during Sun-Saturn oppositions may confirm that tides in the
atmosphere are the cause of the temperature variations found. Correlations between the
Sun-Saturn syzygy and southward moving cold air masses over the 21-day study interval are
often observable in daily climatic surface maps generated from the NCEP/NCAR
Reanalysis data base but analyzing this data requires computer programming, which has not
yet been done. ix In general, longer range datasets are needed for stronger evidence for or
against a possible influence of Saturn on Earth’s weather. Longer datasets would also allow
for the investigation of recurrent planetary alignments in the decadal range which, if Kepler
and others were right, may turn prove be modulating factors in weather and climate studies.

5. References

iBradley, D. A., M. A. Woodbury, G. Brier. (1962) Lunar Synodical Period and Widespread
Precipitation. Science. 137, 748-749. E. E. Adderley and E. G. Bowen. (1962) Lunar
Component in Precipitation Data. Science. 137, 749-750. Lethbridge, Mae DeVoe. (1970).
Relationship between Thunderstorm Frequency and Lunar Phase and Declination. J.
Geophys. Res. Oceans 75, 5149-5154. T. H. Carpenter, R. L. Holle, J. J. Fernadez-Partagas.
(1972) Observed relationships between lunar tidal cycles and formation of hurricanes and
tropical storms. Monthly Weather Review, 100:451-460. Hanson, Kirby, George A. Maul,
William McLeish (1987) Precipitation and the Lunar Synodic Cycle: Phase Progression
across the United States. J. Climate Appl. Meteor., 26:1358–1362. Balling, Robert C.,
Randall S. Cerveny. (1995) Influence of Lunar Phase on Daily Global Temperatures,
Science, 267, 1481-1483. Ebby K. Anyamba, Joel Susskind. (2000) Evidence of lunar phase
influence on global surface air temperature. Geophysical Research Letters 27:18, 2969.
Shouraseni Sen Roy. (2006) Impact of lunar cycle on the precipitation in India. Geophysical
Research Letters 33:1, L01707.

ii Boose E, Gould E. 1999. Shaler Meteorological Station. Harvard Forest Data Archive.

iiiBrown, P.J., Bradley, R. S. , and Keimig, F. T. 2010. Changes in Extreme Climate Indices
for the Northeastern United States, 1870-2005. J. Climate, 23, pp. 6555-6572. Daily
temperature and precipitation data from the United States Historical Climatology Network
(USHCN) (Easterling et al. 1999), the National Weather Service Cooperative Observers
Network (COOP), and CDMP eighteenth- and nineteenth-century Forts datasets were
utilized. Adjustments were made to daily temperatures to account for changes in mean,
variance, and skewness resulting from inhomogeneities, The ten Northeast stations are
Burlington, VT, Farmington, ME, Hanover, NH, Cooperstown, NY, Amherst, MA, West
Point, NY, Groton, CT, New Brunswick, NJ, Franklin, PA, State College, PA.
A signal from Saturn: Bruce Scofield
66 Correlation 29(1) 2013

ivParker, D.E., T.P. Legg, and C.K. Folland. 1992. A new daily Central England
Temperature Series, 1772-1991. Int. J. Clim., Vol 12, p317-342. Manley (1953,1974)
compiled most of the monthly series, covering 1659 to 1973. These data were updated to
1991 by Parker et al (1992), when they calculated the daily series. Both series are now kept
up to date by the Climate Data Monitoring section of the Hadley Centre, Met Office. Since
1974 the data have been adjusted to allow for urban warming. European Climate
Assessment & Dataset (eca&d), November 2008: Klein Tank, A.M.G. and Coauthors, 2002.
Daily dataset of 20th-century surface air temperature and precipitation series for the
European Climate Assessment. Int. J. of Climatol., 22, 1441-1453. Data and metadata
available at http://eca.knmi.nl This is the blended series of location CZECH REPUBLIC,
PRAHA (location-ID: 24).

v
Solarfire 5.1 software from Astrolabe Software, Brewster, MA. Michelsen, The American
Ephemeris for the 20 th Century, (1980, 1997).

vi STDEV = 3.87, Burlington, VT, Cooperstown, NY, Franklin, PA, Groton, CT, Hanover,

NH, NB, NJ, Farmington, ME, West Pt, NY, Amherst, MA, State Col, PA.

vii Krahenbuhl, Daniel S., Matthew B. Pace, Randall S. Cerveny, Robert C. Balling Jr. 2011.

Monthly lunar declination extremes' influence on tropospheric circulation patterns. Journal

of Geophysical Research: Atmospheres.Volume 116, Issue D23, 16.

8Seymour, P.A.H, M. Willmott, and A. Turner. 1992. Sunspots, Planetary Alignments and
Solar Magnetism: A Progress Review, Vistas in Astronomy 35:39-71.

ix http://www.esrl.noaa.gov/psd/data/gridded/data.nmc.reanalysis.html

A signal from Saturn: Bruce Scofield