TECHNICAL PAPER Linkages between

JOURNAL OF THE SOUTH AFRICAN

INSTITUTION OF CIVIL ENGINEERING
solar activity, climate
Vol 49 No 2, June 2007, Pages 32–44, Paper 659 predictability and water
WILL ALEXANDER is Professor Emeritus of the
resource development*
Department of Civil and Biosystems Engineering
of the University of Pretoria, and Honorary W J R Alexander, F Bailey, D B Bredenkamp, A van der Merwe and N Willemse
Fellow of the South African Institution of Civil
Engineering. He spent the last 35 years of his
career actively involved in the development of
water resource and flood analysis methods as well This study is based on the numerical analysis of the properties of routinely observed
as in natural disaster mitigation and climate change hydrometeorological data which in South Africa alone is collected at a rate of more than
studies. He has written more than 200 papers, presentations and books on
these subjects. He has also presented short courses to some two thousand half a million station days per year, with some records approaching 100 continuous years
practitioners during this period. in length. The analysis of this data demonstrates an unequivocal synchronous linkage
Contact details: between these processes in South Africa and elsewhere, and solar activity. This confirms
Postnet Suite #303, Private Bag X20 009, Garsfontein 0042, South Africa
[email protected] observations and reports by others in many countries during the past 150 years.
It is also shown with a high degree of assurance that there is a synchronous linkage
FREDERICK (FRED) BAILEY is a retired higher between the statistically significant, 21-year periodicity in these processes and the
professional technical officer in the Ministry
of Defence, UK. Being a naval architect required an acceleration and deceleration of the sun as it moves through galactic space. Despite a
advanced understanding of the four-dimensional diligent search, no evidence could be found of trends in the data that could be attributed
behaviour of masses on the move through rough
seas under the influence of gravity and the
to human activities.
elements. This in turn led to his hobby of many It is essential that this information be accommodated in water resource development and
years when he achieved an advanced understanding operation procedures in the years ahead.
of the behaviour of the solar system as it moved through galactic space under
gravitational influences. His findings of the synchronous relationship between
the acceleration and deceleration of the sun, sunspot formation, and together
with the other authors, the connection with hydrometeorological processes BRIEF HISTORICAL BACKGROUND An interpretation of the graph provides
is unique.
Contact details:
an insight into the long term behaviour
[email protected] Rippl diagram 1883 of the time series. It is also a valuable tool
For e-mails greater than 1 MB, please use [email protected] The Rippl diagram is a graph of the accumu- when used in conjunction with statisti-
lated values of a hydrologic quantity, gener- cal serial correlation analyses to identify
DAVID BREDENKAMP is a hydrogeologist who
specialises in groundwater resource assessment. ally as ordinate values, plotted against time synchronous characteristics between mul-
He holds an MSc degree in chemistry and a PhD or date as abscissa. It was first proposed by tiple site and multiple processes, including
degree in geohydrology from the University of the Rippl (1883) as a basis for determining the linkages with solar activity should they be
Free State. He has written 40 reports and
publications on this subject, as well as a handbook volume of storage required to sustain a given present.
on groundwater recharge and aquifer storativity. demand from a river. A modification of the
Contact details: method is to divide all the sequential values Hutchins 1889
[email protected]
by the record mean value, then accumulate In 1889 colleagues of D E Hutchins (1889),
ALWYN VAN DER MERWE has been working and plot the sums one at a time. The graph conservator of forests in Knysna, persuaded
for Eskom for the past 14 years, the last nine years will start and end with zero values. him to publish a compilation of his lectures
with a variety of water management issues. He If straight lines are drawn parallel to the in a book with the title Cycles of drought and
has a degree in chemical engineering as well as a
postgraduate degree in water resources mean value such that they pass through the good seasons in South Africa in which he dem-
engineering, both obtained from the University of maximum positive and negative deviations onstrated a remarkable insight into global
Pretoria. His interest is in the development of from the mean value, the vertical distance climatic processes.
computer programs for the numerical
characterisation of regional rainfall and river flow. His participation in this apart is the range R. This is the required Hutchins was one of a generation of
study was in his private capacity. storage capacity, ignoring all evaporation scientists and civil engineers who served in
Contact details: and other losses, that would have been the British Colonial Office in India and then
[email protected]
required to sustain a yield equal to the mean migrated to South Africa. He was stationed
NICO WILLEMSE graduated with a BS Eng from value during the period of record. in Mysore, India, when, in 1876, 1,5 mil-
the University of Pretoria in 1979 and obtained an Assuming a record length of n years and lion people out of a population of 5 million
MBL from the University of South Africa in 1986. a standard deviation of the time series of S, in the state starved to death during a severe
He spent his first years as engineer with the
Transvaal Roads Department and Van Wyk and then drought. It had been noted previously that
Louw Consulting Engineers. He later joined Fluor there were linkages between sunspot num-
Engineers and Sulzer. After a short spell with Rn /Sn ~ nH bers and famines dating back to 1810. Many
management consultants, he returned to the
consulting engineering industry where he was a partner in consulting scientists at that time, including Hutchins,
engineering practices. He is currently in Ireland where he is involved in the Theoretically, H should have a value of 0,5 were involved in the search for predictable
management and implementation of various water and sewerage schemes. for normally distributed sequences of inde- linkages between droughts and sunspot
He is a member of the South African Institution of Civil Engineering
(SAICE). pendent random values. numbers. Hutchins details their efforts in
Contact details:
[email protected]
* This paper is a compilation of separate studies by the authors using different methods, different data sets, different time
frames and different regions of the globe. For this reason there may be some differences in the analytical methods and
Keywords: solar activity, climate predictability, water resource presentations. This in no way negates the overall conclusion that an unequivocal synchronous linkage exists between the
development hydrometeorological processes in South Africa and elsewhere, and solar activity.

32 Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007
Table 1 Presence of 21-year concurrent periodicity in hydrometeorological data In an objective approach the possible
Periodicity influence exerted by fluctuations in the
Number Record sun’s radiant energy on the incidence
Process Not
of sites years 95 % Present None
determinable of drought is stressed. Some of South
Africa’s most severe and prolonged
Evaporation 20 1 180 0 0 20 0
droughts of the nineteenth and twenti-
Rainfall 93 7 141 18 67 8 0 eth centuries have without doubt coin-
River flow 28 1 877 7 12 5 4 cided with troughs of minimum sun-
spot activity. The sun thus appears to
Flood peak maxima 17 1 235 4 7 2 4
be either directly or indirectly responsi-
ble for abnormal weather. As the exact
his book. His perceptive observations are There are many examples of studies mechanism is not yet clear, research in
discussed again later. by mathematicians, hydrologists and civil this field should evidently receive ear-
engineers at that time (Rodrigues-Iturbe & nest attention (Commission of Enquiry
Kokot 1948 Yevjevich 1968; Mandelbrot & Wallis 1968; into Water Matters 1970:92).
In the late 1940s D F Kokot, a civil engi- Yevjevich 1968; Wallis & Matalas 1971; This report provided the impetus for
neer in the South African Department of Yevjevich 1972; Wallis & O’Connell 1973). Alexander’s studies from the 1970s through
Irrigation (now Water Affairs and Forestry), Klemes (1974) commented that ever since to the present day.
undertook a comprehensive study to deter- Hurst published his famous plots for some
mine whether or not there had been recent geophysical time series, the classical Hurst
climatic changes that could have had an phenomenon continued to haunt statisti- ANALYSES (ALEXANDER 1978–2007)
effect on rainfall and river flow. The results cians and hydrologists, and that attempts The following material is from reports and
of his study were published in an Irrigation to derive theoretical explanations from the published papers prepared by Alexander
Department memoir entitled An investiga- classical theory of stationary stochastic during the past 30 years. It illustrates the
tion into evidence bearing on recent climatic processes have failed. development of analytical methods used to
changes over southern Africa (Kokot 1948). The studies included the effects of serial quantify the multiyear periodicity in the
It included many reports by early travellers correlation. When this could not be identi- hydrometeorological data and the devel-
and missionaries. fied in the data, the assumption had to be opment of climate prediction and water
Kokot concluded that there was no made that no meaningful serial correlation resource development models based on this
evidence of any changes in rainfall and existed. However, once the records become periodicity. The analytical procedures used
river flow that could be ascribed to climate long enough to identify the presence of 21- in the analyses are simple. They can be read-
change or any other causes. year serial correlation and its synchronous ily replicated by anyone with experience in
linkage with sunspot activity, then every- observation theory and time series analyses.
Hurst 1950s thing fell into place. The data used in the analyses are from pub-
In 641 AD – more than a thousand years The accumulated departure method and lished records of the responsible authorities.
ago – a water level gauging structure was serial correlation analyses are fundamental The starting point was the incontest-
built on Rodda Island in the Nile River to time series analyses in situations where able, statistically significant (95 %), 21-year
at Cairo. The record from the Rodda non-stationary processes are involved. It was periodicity in the South African rainfall,
Nilometer is the longest available hydro- only when records became long enough to river flow and other hydrometeorological
logical record in the world. In 1950 the detect the 21-year periodicity in the hydro- data. Table 1 shows the presence of 21-year
civil engineer H E Hurst (1951, 1954) ana- logical data that Hurst’s Ghost was eventu- concurrent periodicity in South African
lysed 1 080 years of data from the Rodda ally laid to rest. hydrometeorological data. The degree of
Nilometer recorded during the period statistical significance is dependent on the
641–1946 which he used to determine the Commission of Enquiry into length of the record as well as the magni-
required storage capacity of the proposed Water Matters, 1970 tude and nature of the variability about the
new Aswan High Dam. The Commission of Enquiry into Water mean. The periodicity is almost certainly
Hurst applied the Rippl method to suc- Matters was appointed by the State President present in all hydrometeorological data
cessive segments of equal length, that is, (Government Gazette 1966). Its terms of ref- series, other than open water surface evapo-
n = 10, 20 etc, and found an unexplained erence included the instruction to inquire ration, but has not yet reached a high level
anomaly in the data. The value of the coeffi- into, report upon and submit recommenda- of statistical significance at some of the sites.
cient H for the Nile River was approximately tions on all aspects of water provision and The commencements of the periods
0,75. He then analysed other long geo- utilisation within South Africa. The follow- are readily identified and predictable.
physical records, where he found the same ing extract from the report is directly rel- They are characterised by sudden reversals
anomaly. These were lake deposits (2 000 evant to this paper: from sequences of years with low rainfall
years, H = 0,69), tree rings (900 years, 0,80), The drought phenomenon remains one (droughts) to sequences of years with wide-
temperature (175 years, 0,70), rainfall (121 of the country’s most vexing problems spread rainfall and floods.
years, 0,70), sunspots (0,70) and wheat and it is in drought prediction that While the reversals are a characteristic
prices (0,69). This anomaly became known long-term forecasting can probably of the start of the periods, the periodicity
as the Hurst phenomenon, or Hurst’s Ghost. be of the greatest value. On the other refers to the whole spectrum of values. For
It is important to note that the same multi- hand, it must be admitted that mete- example, a significant correlation exists
year anomalies that were present in the flow orologists have not yet succeeded in between all the fifth values after the com-
records of the Nile River are also present discovering the fundamental causative mencement of the periods, all the n-th val-
in the proxy data used by today’s climate factors either of drought or of excessive ues, and so on. This relationship is stronger
change scientists. precipitation. The classical attitude, viz than the relationship between successive
For more than two decades these that drought is purely a chance occur- values in the hydrometeorological data
anomalies were studied by hydrologists and rence in the climatic history of the where no statistically significant serial cor-
mathematicians. country, does not appear to be correct. relation exists.

Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007 33
Table 2 Vaal River – annual flow record 1923/24 to 1995/96 (expressed as percentages of the mean, showing the mid- harmonic and spectral analysis methods
period and full period sudden reversals from drought sequences to flood sequences) suppress the important, sudden changes
Year Inflow Year Inflow Year Inflow Year Inflow that are present in hydrometeorological
23/24 39 43/44 353 63/64 58 83/84 79 time series, and may also introduce oscil-
latory behaviour that is not present in the
24/25 246 44/45 87 64/65 149 84/85 30
data.
25/26 42 45/46 66 65/66 27 85/86 36 Standard serial correlation analyses
26/27 66 46/47 58 66/67 175 86/87 46 were sufficient to identify statistically sig-
nificant serial dependence and/or cyclical
27/28 44 47/48 57 67/68 31 87/88 208
behaviour should they be present. This
28/29 83 48/49 33 68/69 35 88/89 165 procedure followed the standard time
29/30 142 49/50 100 69/70 60 89/90 65 series analysis methods that require that
30/31 40 50/51 33 70/71 52 90/91 59 the processes be identified graphically in
the first instance, and only subsequently be
31/32 36 51/52 60 71/72 102 91/92 13
described mathematically (Chatfield 1982).
32/33 24 52/53 100 72/73 23 92/93 26 Additional information on the methodol-
33/34 170 53/54 45 73/74 112 93/94 92 ogy developed for hydrological time series
34/35 131 54/55 181 74/75 295 94/95 17
analyses, is detailed in Alexander (1994,
1995a, 1997).
35/36 87 55/56 80 75/76 247 95/96 464
36/37 225 56/57 277 76/77 123 96/97 N/A Trend analyses
37/38 59 57/58 188 77/78 122 97/98 N/A Annual rainfalls recorded in the Weather
Bureau’s (now South African Weather
38/39 202 58/59 69 78/79 31 98/99 N/A
Service) district rainfall database were stud-
39/40 112 59/60 75 79/80 63 ied (Weather Bureau 1972, updated annu-
40/41 131 60/61 105 80/81 62 ally). Conventional statistical trend analyses
41/42 54 61/62 50 81/82 19 could not be performed in the presence of
the large periodic variations in the data.
42/43 185 62/63 68 82/83 12
However, simple arithmetical and graphical
analyses demonstrated increases in rainfall
It is interesting to note the absence of 21- and –785. The maximum was more than in 75 of the 81 rainfall districts with com-
year periodicity in the evaporation data twice that of the minimum that occurred plete records, totalling 9 % for South Africa
in table 1. Another observation is that the only three cycles earlier. as a whole for the 78-year period 1921 to
magnitudes of the periodic changes rela- The lengths of the corresponding dou- 1999. Forty-two districts had increases of
tive to the long-term mean values, increase ble sunspot cycles were 20, 21, 22 and 20 10 % or more, 12 districts had increases
from evaporation (absent) to rainfall, to river years with a mean of 20,8 years, a minimum of more than 20 %, and four districts had
flow, to flood peak maxima. Together, these of 20 years and a maximum of 22 years. increases of more than 40 %. There was also
characteristics indicate that the periodicity The average number of sunspots in the an increase in the numbers of widespread,
is amplified by the processes involved in the alternate cycles that make up the double heavy rainfall events during the latter
poleward redistribution of solar energy. cycles were +706 and –664, demonstrat- half of the past century (Alexander & Van
Neither the lengths of the periods nor ing a meaningful difference in sunspot Heerden 1991).
the synchronous occurrences are precise in activity in the alternating cycles. As will There were also increases in open water
the mathematical sense, but their presence be seen, the alternating sunspot cycles surface evaporation observed in 14 of the 19
is beyond all doubt. have appreciably different effects on the accepted data sets studied. No trends were
hydrometeorological processes. discernible in any of the other processes
Sunspot database It will later be demonstrated that it is studied. If present, they were overwhelmed
Conventional sunspot cycles were used as not the annual sunspot numbers that are by the natural variability of these processes.
indicators of solar activity. The following important in identifying the relationship,
data are from website information distribut- but the rate of change in the numbers. This Numerical comparison
ed by the World Data Centre for the Sunspot is not apparent in the conventional graphs The next aspect studied is illustrated in
Index (2005). of the sunspot cycles where all numbers table 2, which lists the annual flows in the
There were eight complete cycles during have positive values. The sunspot num- Vaal River at Vaal Dam as percentages of the
the past century. These commenced with bers in the alternating sunspot cycles were mean annual runoff at the site. Vaal Dam is
the sunspot minimum that occurred in June therefore given negative values, and an the major source of water for South Africa’s
1913 and ended with the sunspot minimum arbitrary graph origin of -200 was used for largest metropolitan, industrial and mining
in March 1996. The lengths of the cycles convenience in order to present all values region. This is the most analysed hydrologi-
were 10, 10, 11, 10, 10, 12, 10 and 10 years, as positive numbers. This is a requirement cal record in South Africa. The full period
with a mean of 10,4 years. These values for statistical analyses where logarithms reversals (heavy horizontal lines) refer to
are within a narrow range of between 10 are employed (Alexander 2002b). These are the years when the sudden reversals from
(minimum) and 12 (maximum) years. A cor- graphical datum changes and do not affect low flow sequences to high flow sequences
responding increase in solar activity during the interpretations. occurred. These identified the commence-
the past century is reflected in the increase ment of the 21-year periods. Note that these
in the numbers of sunspots per cycle, com- Methodology are not exactly 21 years apart. The light
mencing with the cycle that started in 1913. The emphasis was on simple arithmetical horizontal lines identify the commencement
Alternating cycles are identified by negative and graphical interpretations rather than of the mid-period reversals.
values. The sunspot numbers per cycle were mathematical interpretations. The reasons The reversals in the flows in the Vaal
+442, -410, +605, -757, +950, -705, +829 were that mathematical analyses such as River from drought sequences to flood

34 Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007
 Sunspots 1913 to 1996 Inflow: Vaal Dam + Grootdraai 1923 to 1996
4 4

3 3

2 2

1 1

0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Annual values/mean (horiz scale = years + 1900, vert scale = x mean) Annualvalues
Annual values/mean (horiz scale
/mean (horiz. scale==years
years++1900,
1900,vert
vert.scale
scale==xxmean)
mean)
4 4
B D F B D F
2 2

0 0
A E G C E
-2 C -2 A
G
-4 -4
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Accumulated departure from the mean Accumulated departure from the mean

0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Observed periodicity – years beginning Oct 1912, 1933, 1954, 1974, 1995 Observed periodicity – years beginning Oct 1912, 1933, 1954, 1974, 1995
1 1
H
H
0 0

-1 5 10 Years 15 20 25 -1 5 10 Years 15 20 25
Correlogram (95 % confidence band) Correlogram (95 % confidence band)

Figure 1 Comparison of the characteristics of annual sunspot numbers with corresponding characteristics of the annual flows in the Vaal River

Table 3 Comparison of sudden changes in the annual flows in the Vaal River with corresponding sudden changes in Graphical comparison
sunspot numbers The next issue is the nature of the solar-
Three-year totals of flows in Vaal River Three-year totals associated with the induced periodicity of the hydrometeoro-
(% of record mean) corresponding sunspot minimum logical processes. Figure 1 shows graphical
Three Three comparisons of the properties of the double
Three subsequent Sunspot Three lowest subsequent sunspot cycle with those of the Vaal River.
Minimum year previous years years minimum years years This follows the method developed by
1932/33 100 388 1933 25 250 Alexander (1978) and successfully used to
predict the climate reversals from drought to
1941/42 297 625 1944 56 277
flood sequences that occurred in 1995 and
1953/54 205 538 1954 50 370 again in 2006 (Alexander 1995b, 2005c).
1965/66 234 241 1964 53 247 A reference datum value of –200 was
used in the sunspot data in order to accom-
1972/73 177 654 1975 73 275
modate the negative values assigned to the
1986/87 112 438 1986 60 400 alternate sequences. This has no effect on the
1994/95 135 464+ 1996 48 277 interpretations as it only affects the position
of the datum line but not the variability about
Average 180 478 Average 52 300
that line. The top panels are the conventional
dimensionless histograms, where all values
sequences evident in table 2 correspond sudden increase in the annual flows in the are expressed as multiples of the record mean
closely with similar reversals in sunspot Vaal River from the three previous years to values. While the cyclicity is apparent in the
numbers. This is evident in table 3. In all the three subsequent years. This is directly sunspot panel, it is not recognisable in the
but one sequence (Vaal River 1965/66, data associated with a six-fold increase in sun- river flow. The river flow histogram shows
not available), the three-year totals after spot numbers. The second important point the high degree of asymmetry about the
the minima of both river flow and sunspot is the consistency in the range of sunspot mean value with many more values less than
numbers, are substantially greater than the numbers before and after the reversal. the mean value than above it. This is typical
three-year totals before the minima. This The totals for the three prior years varied of river flow data in dry climates.
information demonstrates the close associa- between 25 and 60, and the totals of the The most informative graphs are those
tion between major variations in river flow three immediately subsequent years varied in the second panels, which show the accu-
and corresponding variations in sunspot between 250 and 400. It is very clear that mulated departures from the record mean
activity, with a high degree of confidence. these are systematic changes associated values. These are obtained by subtracting
There are several interesting features with the sunspot minima, and are not ran- the mean values (1,0) from each of the val-
in this table. There is an almost three-fold, dom events. ues in the histogram. Some of the values

Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007 35
 the nature of the periodicity in river flow
400%
at a number of representative sites in South
K Africa. The procedure used for each site was
350%
to extract the data in the first 20 years of each
300% sequence starting in October of the follow-
J L N ing years: 1912, 1933, 1954, 1974 and 1995.
250% Then the average values for each year of the
% MAR

sequence divided by the record mean annual

200%
runoff (MAR), were determined and plotted.
M
150% The reference period used for calculating the
MAR was that from 1954 to 1974 as it was
100% the only period that was common to all data
sets. The selection of a reference period does
50%
not affect the results. This procedure was
0% repeated for the other sites. The rainfall and
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 flood peaks exhibited similar characteristics,
Periodic sequence (years) although the rainfall amplitudes were less
and the flood peak amplitudes were greater
Figure 2 Characteristics of the periodic sequences of river flow at representative dam sites. The double sunspot cycle is than those of river flow. The diagrammatic
diagrammatically superimposed double sunspot cycle J-K-L-M-N is included
in the figure for ease of comparison.
will be negative. These are accumulated one tion of the water mass above sea-level. This While there is a large scatter in the plot-
at a time and the sum plotted. residual energy has its origin in solar activity; ted results, the general trend is clear and sev-
An increase in the accumulated depar- followed by the arrival on the earth’s atmos- eral conclusions can be drawn from it. Major
tures of the sunspot numbers during the phere, continents and oceans; followed by the flood events are associated with the first half
period of record is immediately apparent. poleward movement of the energy through of the first sunspot cycle (J-K). This is the
The maximum negative departures occurred complex atmospheric and oceanic processes; sub-period when the rate of increase in sun-
at the start of the 21-year periods, identified followed by the systems that produce the spot numbers is greatest, and is associated
as A, C, E and G. rainfall; and finally by the complex rainfall- with global atmospheric and oceanic turbu-
The comparison with that of the flow runoff processes. The survival of the periodic lence at this time. This in turn generates the
in the Vaal River is very instructive. The signals on its own demonstrates a strong and processes that produce heavy, widespread
reversals at points A, C, E and G are virtually unequivocal relationship between variation in rainfall events that generate river flow.
identical with the corresponding reversals in solar activity and the corresponding variation The characteristics of the second of the
the sunspot data. They occurred during the in climatic responses. two sunspot cycles (L-M-N) are very different
hydrological years beginning October 1933, Note also the increase in sunspot activity from those of the first cycle (J-K-L). Fewer
1954, 1974 and 1995. The rising limbs A-B, during the last century, with some indication heavy rainfall events occur. Droughts become
C-D and E-F are sequences of years where of a decrease E-F-G at the end of the century. increasingly prevalent during this cycle. It is
the inflows were greater than the mean value. postulated that this is the consequence of the
The falling limbs B-C, D-E, and F-G are Time interval of interest differences in solar activity between the two
sequences where the inflows were less than In 1889 D E Hutchins (1889) published his cycles as well as the lesser sunspot numbers.
the mean value. These alternating sequences book Cycles of droughts and good seasons in There is also a clear, diminishing, annu-
were reported in the early hydrological lit- South Africa. It was based on records of three al oscillatory pattern during the beginning
erature where they were referred to as the processes – rainfall, river flow and air tem- of the second sunspot cycle (L-M) that is not
Joseph effect, after Joseph’s biblical prophecy perature – and their linkage with sunspot present in the greater scatter during the first
(Mandelbrot & Wallis 1968). activity. He described the sudden changes cycle. No possible causes can be offered.
The third panels are the correlograms. from drought to flood conditions. The first ten years of the sequence are
This is a standard calculation procedure in Note how his study referred to a combi- associated with the first sunspot cycle J-K-L.
time series analyses. The statistically signifi- nation of eight conditions. These were global These are Hutchins’ good years and the last
cant cyclicity in the sunspot data is clearly and multiyear, alternating, multi-process, ten years L-M-N are his drought years.
apparent. The 95 % confidence limits are sunspot linkage, abrupt changes, and the
± 0,22. The minimum and maximum (H) two hydrological extremes. Characteristics of the periodicity
autocorrelation coefficients occur respectively The key time unit of interest is not The regular grouping of alternate sequences
at 10 (–0,83) and 21 (+0,70) years, which are twelve months, but a sequence of 21 years. of wet and dry years and its linkage with
well in excess of the 95 % confidence limits. Furthermore, each of these sequences has the 21-year (approximately) periodicity is
The statistically significant cyclicity in two components, a good year sequence beyond all doubt. The best description of
the sunspot data is no longer present in the component followed by a drought sequence the periodic behaviour is that provided by
correlogram of the annual flows in the Vaal component. Together they constitute the Hutchins (1889:25): ‘The yellow line ris-
River, where the residual coefficients indicate time unit of interest. It is the numerical ing steeply to a maximum and then falling
random noise. The only, but very important, properties of the time unit of interest that away gradually to a minimum is the sunspot
residual serial correlation, is the statistically have to be determined. Hutchins appreci- curve – a curve which ought to be graven on
significant 21-year periodicity. This is identi- ated this more than 100 years ago. the mind of every man and woman in South
fied at (H) in the bottom panel of the figure. Africa.’ There is no evidence of the sinusoi-
This relationship exists despite the long Predictability dal oscillatory behaviour in the data.
and complex energy path starting at the sun Figure 2 is from Alexander 2005a. Without An important characteristic is that the
and ending in the river flow that enters Vaal going into detail, it shows the properties of most extreme conditions occur at the begin-
Dam. The only residual energy is the poten- the time unit of interest and its association ning of the periods (floods) and at the end of
tial energy, which is a function of the eleva- with the double sunspot cycle. It illustrates the periods (droughts) with sudden revers-

36 Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007
Table 4 Comparison of ranked maximum values with sunspot minima second half-period are the consequence of
South African rainfall Sunspot minima the absence of these events.
Rank Compare the lengths of the sequences
Month mm Year Lag (years)
of wet and dry years with the biblical seven
1 Mar 1925 211 1923 +2 years of plenty followed by seven years of
2 Jan 1974 149 1976 -2 famine. The ancient Egyptians were well
aware of these alternating sequences in the
3 Feb 1939 148 1933 +6 annual flows of the life-giving Nile River.
4 Feb 1988 145 1986 +2 This periodicity is very important for all
those who maintain that global warming will
5 Jan 1923 138 1923 0
result in increased variability in the hydro-
6 Jan 1976 136 1976 0 logical process – specifically floods, droughts
7 Feb 1955 132 1954 +1 and water supplies. If they are to provide
convincing arguments they will have to dem-
8 Jan 1958 130 1954 +4 onstrate (not postulate) how global warm-
Ranked flood peak maxima in the Mkomazi River ing will change the alternating wet and dry
sequences; the associated periodic properties;
Flood maxima Sunspot minima
Rank and the drought and flood severities.
Year m3/s Year Lag (years)
1 Mar 1856 7 000 1856 0 Further confirmation
Further confirmation of the linkage between
2 Mar 1925 6 260 1923 +2
the rate of increase in sunspot numbers
3 May 1959 6 200 1954 +5 and rainfall over South Africa as a whole is
4 ? 1868 6 130 1867 +1 shown in table 4, which shows the relation-
ship between the months during which
5 Mar 1976 2 140 1974 +2
the maximum rainfall occurred and the
corresponding years in which the sunspot
Table 5 Wet and dry sequences minima occurred. The lag is the difference
Length of sequence in years when the sunspot minima are used
Periodic Reversals
Years Wet/dry to predict the rainfall maxima. The lower
sequence number Wet Dry (Alexander)
panel is a repeat of the upper panel, using
Bredenkamp: Mzimgazi + St Lucia + Uitenhage + Wondergat the flood peak maxima observed in the
1919–1924 Wet 08 to 13 5 Mkomazi River, south of Durban.
The 1856 peak was concurrent with
1925–1929 Dry 14 to 18 4 1933
the flood peak in the Mgeni River, where
1930–1939 Wet –03 to 07 9 floodwaters flowed across Durban and into
1941–1953 Dry 09 to 21 12 1954 Durban harbour. These floods occurred in
March 1856. The maximum recorded flood
1955–1962 Wet 02 to 09 7
engraved on the buttress of the Georges V
1965–1971 Dry 12 to 18 6 1973 Bridge built in 1760 across the Loire River in
1972–1978 Wet –01 to 06 6 Orléans, France, occurred in June 1856. The
sunspot minimum occurred in December
1980–1983 Dry 08 to 11 3
1855 (World Data Centre for the Sunspot
1984–1990 Wet 12 to 18 6 1995 Index 2005). This correspondence in time
Tyson: South African rainfall (months) and space (hemispheres apart) is
1905–1915 Dry 14 to 04 10 1912
far too great to be coincidental.
Six of the eight rainfall events and four
1916–1924 Wet 05 to 13 8 of the five flood peak maxima occurred
1925–1932 Dry 14 to 21 7 1933 within two years of the sunspot minima.
1933–1943 Wet 01 to 11 10 This confirms that these extreme events are
sensibly synchronous with the reversals in
1944–1952 Dry 12 to 20 8 1954
sunspot numbers associated with the sun-
1953–1961 Wet –01 to 08 8 spot minima, as shown in table 3.
1962–1970 Dry 09 to 17 8 1973 It is also important to note that these
maxima were recorded 80 years ago
1971–1980 Wet –02 to 08 9
(rainfall) and 149 years ago (Mkomazi
floods) and that there is no evidence
als from droughts to floods that identify the maximum annual river flow values for of an increase in time that could be
the beginning of the periods. Note that this all sites analysed was 675 % of the record associated with global warming. Historical
information on the characteristics of the mean values compared with the average of observations in several other rivers confirm
hydrometeorological data has been known for the following ten years of only 380 % of the that the floods in the mid-1800s remain
more than 100 years. record mean values. This is probably asso- the highest on record.
Analyses showed that the rainfall and ciated with the sign of the sun’s magnetic
river flow during the first half-period (first polarity. Other analyses not reported here Alternating wet and dry sequences
sunspot cycle) are appreciably higher than showed that the high values in the first half- – Tyson and Bredenkamp
the second half-period. For example, for the period are the result of widespread, heavy Table 5 shows the alternating wet and
first ten years of the period, the average of rainfall events, while the low values in the dry sequences in South African hydrom-

Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007 37
Table 6 Sequence numbers of 40 world maximum floods
Sequence Number Sequence Number Rainfall – Mpumalanga Highveld vs model
number of floods number of floods
2,40
1 4 11 2
2,20
2 2 12 - 2,00
3 1 13 1 1,80
4 - 14 2 1,60

MAP
5 3 15 3 1,40
6 6 16 - 1,20
7 2 17 1 1,00

8 1 18 1 0,80

9 2 19 - 0,60
0,40
10 3 20 4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Hydrological year ’02 ’05
21 2
Actual (based on ten sites) 1,0 % (distribution derived
Total 24 Total 16
5,0 % from 36 district rain
25 % zones)
50 %
eteorological data. The following conclu- 75 %
sions can be drawn from the independent 95 %
observations by Tyson and Bredenkamp. 99 %
Each author used different data and differ-
ent analytical methodologies. Bredenkamp Figure 3 Validation of the Alexander climate prediction model based on regional rainfall in the Highveld region
analysed lake and groundwater levels while
Tyson studied areal rainfall over South by the two authors. These are well within Table 6 shows the sequence numbers
Africa. the range of achievable accuracy in most during the time units of interest of the
Tyson (1987) provided evidence support- hydrometeorological estimates. world’s 40 maximum floods. Compare this
ing the presence of alternating sequences of with the information in figure 2 where the
years with high and low rainfall over large Flood alerts sequence numbers are identified.
regions of South Africa. He noted the oscil- In November 2005, during the then There is a wealth of information
latory nature of the data, although he was drought, Alexander (2005c) issued the first worldwide on these alternating sequences
unable to trace its cause. He concluded that of four flood alerts based on the model. and their linkage with the double sunspot
its physical reality was considerable in South They included details of action that local cycle.
Africa and in other countries. He noted that authorities should take to limit the potential
the 11-year solar cycle was mentioned in the loss of life in informal settlements. Mathematical modelling
literature but did not discuss it further. Three months later large regions of The final stage was the development of a
Bredenkamp (2000) studied ground- the African subcontinent were wetter and mathematical simulation model for water
water resources. He used the cumulative greener than at any time in human memory. resource development and management
departure method as his principal tool, for Floods occurred in many rivers from Angola applications that accommodates the char-
which he developed a mathematical rela- in the north through to the coastal rivers of acteristics described in this report. The
tionship. He demonstrated the presence of the southern Cape. Dams filled over most mathematical simulations were based on
wet and dry sequences from 1919 through of the region. The loss of life was minimal the readily quantifiable periodicity of the
to 1992 based on water level observa- thanks to the emergency services in the hydrological data. There was no need to
tions at Lake Mzingazi; discharge from the areas. These observations are further con- invoke linkages with solar activity. Nor
Uitenhage springs corrected for abstractions; firmation of the validity of the prediction was it necessary to include the postulated
water levels at Lake St Lucia; and ground- model. adverse consequences of global warming,
water levels at the Wondergat sinkhole in such as increases in floods and droughts
a large dolomitic formation. These all have World maximum floods and threats to water supplies. The meth-
high storage/input ratios that smooth out the In 1984 the International Association of odology is described in Alexander 1994
short-term fluctuations. Hydrological Sciences (IAHS) published and 1997.
Begin with the right-hand column of a World Catalogue of Maximum Observed
the table. This shows the years in which Floods (Rodier & Roche 1984) as a contri- Development of the climate
the 21-year reversals occurred based on bution to the International Hydrological prediction model
studies by Alexander. The fourth and fifth Programme of UNESCO (the United Very great advantages in the manage-
columns show the alternating wet and dry Nations Educational, Scientific and Cultural ment and practical utilisation of our
sequences. Note how the periods between Organisation). The 40 largest recorded water resources would follow if a meas-
the reversals each consist of a wet sequence floods in the world between 1900 and 1982 ure of reliability could be achieved in
followed by a dry sequence. This confirms from the catalogue were listed on this basis. the long-term forecasting of climato-
the pattern observed in other hydromete- Twenty-four of them occurred during the logical conditions … The Commission
orological data. first of the two sunspot cycles that made up regards it as essential that research
There is also a good correspondence the double sunspot cycle, while 16 occurred and attempts to acquire the necessary
between the dates in the first column of the during the second cycle. This analysis of data to make long-term weather fore-
two panels, as well as between the first and world maxima supports the hypothesis that casting possible be actively supported
last columns in both panels, considering the the linkage with the double sunspot cycle is (Commission of Enquiry into Water
different data sets and methodologies used a global phenomenon. Matters 1970:7).

38 Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007
 Validation: Orange River vs model data Falls dam and power station in 1954, at
4,00
Jinja on the mouth of the Main Nile, flows
3,50 through the power station were maintained
on a ten-day rolling basis so as to match the
3,00
flows which would have occurred, had the
2,50 Ripon Falls still been the control. Thus, in
effect, the natural Nile outflows were main-
MAR

2,00
tained in an unbroken form up until June
1,50 2000. This requirement was enshrined in
the Nile Waters Agreement. In June 2000
1,00
the new Kiira power station was commis-
0,50 sioned in parallel to the original station and
subsequent extractions then exceeded the
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ‘natural’ ones.
Hydrological year ’97 ’02 ’05 The massive Lake Victoria is situated
Actual (inflow into Gariep) 1,0 % (distribution derived on the equator. Kite (1982) established that
5,0 % from 24 sites) 80 % of the lake’s water balance comprised
25 %
rainfall and evaporation with only 20 %
50 %
75 % due to tributary inflows and Nile outflows.
95 % In the period up to the 1960s the lake level
99 % oscillated about a Jinja gauge level of approxi-
mately +11 m. The ‘agreed curve’ flows
Figure 4 Validation of the Alexander climate prediction model based on Orange River flow data corresponding to this were used to size the
original power station and its installed capac-
The prediction model is based on the sta- no year-to-year variability in the probability ity of 150 MW. However, in the early 1960s
tistically significant (and therefore predict- distributions. a dramatic rainfall increase over central and
able) 21-year periodicity in South African The model is based on the observed east Africa raised the lake to unprecedented
hydrometeorological data. Verification is an 21-year periodicity in the data. Although levels. To meet ‘agreement’ requirements from
essential process in any prediction model. not included in the model, the synchronous the 1960s onwards, the sluice gates at the
The model has to pass this test before it linkage with sunspot activity is beyond all dam had to discharge excess water, beyond
can be used with confidence. Alexander’s doubt. Note also the increase in rainfall rela- those which could be used for generation.
climate prediction model was published in tive to the mean values during the past four The rainfall event and the climate anomaly it
Water SA in April 2005 (Alexander 2005a). years 8 to 11 in the figures. Similar increas- produced have been thoroughly reported by
It has now been tested and verified and can es occurred in the river flows in the Western Lamb (1966). Since the 1960s the lake would
be applied in practice for multiyear regional Cape. Note further that the observed annual appear to have been dropping with an overall
rainfall and river flow analyses and pre- rainfall and river flows during the past years mean trend of 29 mm/yr.
dictions. In particular it can be used with were nowhere near the historical maxima In the 1920s it was noted that the level
greater assurance than current methods for and minima. of Lake Victoria seemed to respond to pat-
multiyear simulations required for water Most importantly, refer to figures 2, 3 terns of solar variation in the form of the
resource development and management. and 4 and note that with the exception of sunspot number index. The correlation
year 13 (rainfall and river flow) and year 17 disappeared after the 1920s and when the
(rainfall only) the mean values of the predic- matter was later reviewed by Hurst (1952),
VERIFICATION OF THE CLIMATE tions for the next ten years are all less than he found no statistical correlation for the
PREDICTION MODEL (VAN DER MERWE) the long-term mean annual runoff (MAR). complete record and concluded that the
An Excel program was developed to imple- The predictions for the present hydrological apparent correlation up until the 1920s had
ment the model and tested on a number of year (2006/07) are below average rainfall been a coincidence.
occasions. The following is an example of and river flow. The next climate reversal The issue was re-examined by the
the verification of the model following the from drought to flood conditions based on present writer (Mason 1993, 1998, 2006).
first complete hydrological year after its the analysis of historical data is only expect- When the falling trend of 29 mm/yr is
publication. ed to occur in 2016. This confirms the link- removed, it would appear that the solar
Figure 3 is the annual rainfall for the age with the double sunspot cycle. index correlation has been re-established
Highveld region where the interest is in the The next ten years will be critical for since 1968 (see figures 5a and 5b). Coupled
availability of water for cooling at the coal- water resource development and operation. with the earlier period to 1928, this implies
fired power stations. Figure 4 is the annual This has nothing whatsoever to do with that the correlation covers two-thirds of the
river flow in the Orange River where the global warming. gauged record. The two periods of correla-
interest was in hydropower generation at the tion also correspond to periods of high vol-
Gariep and Van der Kloof dams. These are canic dust (Lamb 1970, 1977, 1983).
both important practical applications on an LAKE VICTORIA AND A RENEWED The reasons for a possible volcanic dust
issue of national importance. SOLAR CORRECTION (MASON) link are not clear, although there has been
The figures shows box and whisker The level of Lake Victoria has been carefully speculation in recent years on the way in
probability plots derived directly from monitored by international survey teams which radiation may affect dust, aerosols
recorded data within the regions of inter- since 1896, as have its outflows into the and water vapour in the atmosphere to affect
est. The outer thin lines show the observed Nile river system. In the 1950s the relation- climate (Lockwood et al 1999). A discus-
maximum and minimum values. Current ship between the lake level and the Nile sion of such linkages is beyond the scope
simulation models used for water resource discharges, controlled by the Ripon falls, of this paper, however, it is hoped that the
analyses assume that all the boxes are in was established in the form of an ‘agreed apparent Lake Victoria example may provide
the same vertical position, that is, there is curve’. After the commissioning of the Owen useful data for those working in that field.

Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007 39
 In particular it would seem useful to explore
whether or not the correlations with the
13,00 sunspot index could be improved still fur-
Falling trend of ther using other variable solar parameters.
29 mm per year If past trends continue, some respite
may occur after 2006, until the next low
Jinja gauge levels (m)

Approx mean level is reached on or around 2016. It remains

12,00 prior to 1962
unclear if the present correlation will con-
tinue. However, the present writer believes
Lake rise from the matter warrants a wider awareness and
1962 to 1964 further investigation. Even if the assump-
11,00 tion of continued correlation is used as just
one possible planning scenario, the poten-
tial benefits for planned lake usage and
for assessing future aid and famine relief
10,00 requirements in areas affected by the Nile
system would seem clear.
1940
1930

2000
1980

1990
1950
1920

1960
1900

1910

1970
Year
IRELAND (WILLEMSE)
Figure 5a Levels of Lake Victoria from 1896 to 2005 Willemse studied the relationship between
annual rainfall and sunspot numbers
for several meteorological stations in the
12,40 Republic of Ireland. These included Dublin,
Malin Head (most northern point of the
Period Period 180,0
1896–1928 1968–2005 island), Valentia (south-west coast) and Birr
(midlands). The data for Dublin date back to

Solar variation (sunspot number)

Natural 150,0 1881 (more than 46 000 data sets), but the
Lake (Jinja gauge) level (m)

lake level other data sets were a lot shorter.

120,0 Figure 6 is the comparison for Dublin.
Sunspot Other analyses not reported here included
11,40 number
90,0 the wet and dry seasonal rainfall periods
separately in an attempt to identify any dif-
Lake level minus ference in the wet/dry cycle.
60,0
the 29 mm per The cusp shape of the cumulative depar-
year falling trend ture plots indicates that both the accumu-
30,0 lated totals of sunspot numbers and annual
rainfall were increasing synchronously with
10,40 0,0 time. It is particularly interesting to note the
1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

close correlation between the sunspot num-

bers and rainfall from about 1940 onwards.
Year
This corresponds with an increase in sun-
spot activity during the latter half of the last
Figure 5b Levels of Lake Victoria from 1896 to 1928 and from 1968 to 2005 compared to solar variation in the form of sunspot century (see figure 1).
number indices, but with the 29 mm per year falling trend in lake level eliminated from the 1968 to 2005 data

Dublin: Annual precipitation 1881–2006 CAUSES OF PERIODICITY IN

1 100,0 30,0
SOLAR ACTIVITY (BAILEY)
The synchronous linkages between sun-
900,0 20,0 spot activity and the hydrometeorologi-
cal processes are demonstrated earlier in
this paper. Bailey identified the processes
700,0 10,0 within the solar system that are responsible
Precipitation in mm

for sunspot activity. His studies require

500,0 0,0 the visualisation of four-dimensional,
time-space acceleration and deceleration
of the sun, and consequent changes in the
300,0 -10,0 distance between the earth and the sun,
Annual precipitation
Cumulative departure as they move in an oblique plane through
100,0 from MAP (scale x 10) -20,0 galactic space.
Sunspot number:
cumulative departure
from mean The earth’s spiral path
-100,0 -30,0 Conventional illustrations show the earth
1881
1885
1889
1893
1897
1901
1905
1909
1913
1917
1921
1925
1929
1933
1937
1941
1945
1949
1953
1957
1961
1965
1969
1973
1977
1981
1985
1989
1993
1997
2001
2005

orbiting around a static sun. This is mislead-

Year ing. First, the sun wobbles through a tube
of space and not along a smooth path at a
Figure 6 Comparison between annual rainfall recorded in Dublin, Republic of Ireland, and annual sunspot numbers constant velocity. Second, the earth orbits

40 Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007
 Visualise the three-dimensional view of this
The ‘tube’ in the middle represents the volume of space that the
sun revolves in and it is about 3,7 x 106 km in diameter. figure with the orbit path spiralling towards
The ecliptic plane is shown as being at 45o to the line of flight. N
the viewer.
The earth to sun distance (the chord length) can vary each year, EARTH Perihelion Starting in 1978, the orbit maintains a
depending upon where the sun is located inside the ‘tube’ SC M nearly constant distance from the SSCM.
k of S
N Trac In 1985 the orbit starts moving closer to
the central point occupied by the SSCM. It
Vernal equinox swings around the SSCM, reaching its clos-
EARTH N Perihelion
est position in 1990. It then spirals away
from the SSCM until 1994. From 1995

e
through to 2000 there is little change in the

lan
Chord

ic p
displacement from the SSCM. From 2001

ip t
through to 2006 it makes another approach

Ecl
Autumn equinox to the SSCM.
Aphelion e) The sun follows a weighted reciprocal
tim
1y
ear
it a l
or b path but its centre of mass is much closer to
r th ’s
(E a the SSCM. It also accelerates and decelerates
synchronously but moves in the opposite
direction in order to maintain the system in
equilibrium.
The sunspot minima occurred in 1986,
Figure 7 Earth’s orbital path through the galaxy 1996 and 2006. The compass points on the
figure are for reference purposes only. Note
that the sunspot minima of 1986 and 1996
1980
both occurred in the SW quadrant of the
N 1979 Perihelion
1981 figure, and that of 2006 in the NW quad-
Jan
rant when viewed from a position ahead
Orbit of Jupiter of the approaching solar system. This is in
1982 an anticlockwise direction relative to the
Mar 1978
2004 forward clockwise movement of the spiral
2005
paths about the SSCM followed by the orbit-
ing components of the solar system. The
angular distance followed by the orbit from
1983 2006 2003
1986 to 1996 was 360° when it returned to
1992 1991 Nov the same quadrant. It was only 270° from
1996 to 2006 when it did not complete a
1993 full 360° rotation around the SSCM. The
1984 1990 2002
angles are approximate but are amenable to
calculation.
1994
Table 7 shows the positions of the
1985 1989 planetary system’s centre of mass (PCM) at
May the time of the sunspot minima during the
2001
1995
period 1902–2006.
1986 The information in this table provides
1988 the first positive linkage between solar activ-
ity and the hydrometeorological time series.
1996 1987 2000 Alexander’s analyses showed a statistically
Sept significant linkage with the double sunspot
cycle. He found no statistically significant
1997 July linkage with the single, 11-year cycle. His
1998 1999
analyses showed that these alternating
cycles are associated with different hydrom-
Figure 8 The orbit of the centre of mass of the four major planets about the SSCM from 1978 to 2006. The movement is eteorological characteristics.
towards the viewer The periodic behaviour of the solar sys-
tem has a duration of 21 years (actually 20,8
the solar system’s centre of mass (SSCM) length) varies, depending on where the sun years during the past century), not 11 years.
and not the sun’s centre of mass. The earth is located in the tube. While the paths of the (Note the sums of the pairs in the last col-
therefore follows a spiral path as it moves sun and the earth are closely linked as they umn of table 6.) This explains why scientists
through space. This is illustrated in figure 7. move through space, the changing relative have been unable to find a linkage with the
(It is important to note that the scales in the positions result in corresponding changes in 11-year cycle, from which they erroneously
figures are highly compressed so that they the distance between them. concluded that there is no linkage with solar
can fit.) activity. While the relative positions of the
The tube in the middle represents the Influence of the planets planets are closely grouped in space at 21-
volume of space that the sun revolves in and Figure 8 shows the path of the combined year intervals, they are not precise in either
is about 3,7 106 km in diameter. The ecliptic centre of mass of the four major planets, time or space. This is the reason for longer
plane is at a 45° angle to the line of move- Jupiter, Saturn, Uranus and Neptune, rela- period cyclicity including 178 years and
ment. The earth to sun distance (the chord tive to the SSCM for the period 1978–2006. longer cycles.

Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007 41
Table 7 Angular shift in the position of the planetary centre of mass at the time of sunspot minima for the period Alexander’s studies of the hydromete-
1902–2006 orological data indicate the presence of an
Rotation angle Cycle length instability phenomenon associated with the
Cycle number Cycle period PCM quadrant
(approx) (years) predictable, sudden climatic reversals from
14 1902–1913 SW 360 11 drought to widespread rainfall conditions.
These are closely synchronous with the
15 1913–1923 SW 270 10 sunspot minima. It is not yet clear whether
16 1923–1934 NW 360 11 this instability is associated with solar
17 1934–1944 NW 270 10 processes or with the responding climatic
processes.
18 1944–1954 NE 360 10
19 1954–1965 NE 270 11
VARIATIONS IN SOLAR IRRADIANCE
20 1965–1975 SE 360 10
21 1975–1986 SE 270 11 The sun’s wobble
22 1986–1996 SW 360 10 The distance of the sun from the SSCM is
the weighted reciprocal of the distance of
23 1996–2006 SW 270 10 the combined centre of mass of the orbiting
planets. Consequently, both the sun’s dis-
tance from the SSCM and its galactic veloc-
N ity are continually changing. This creates a
1950
1952
wobble in its path through space. This can
200 be calculated given the knowledge of the
1948
1954 masses and orbits of the four major planets.
1946 1944
Astronomers make use of the wobbles in
1958
100 the trajectories of distant stars to determine
1956 whether or not they are accompanied by
(Viewed in the direction of the sun’s travel)
18 19 0 1
orbiting objects. Figure 9 shows the sun’s
Denotes track of SSCM wobble as it moved through galactic space
1950 1960 x 106 km N
e
during the period 1944 to 1958. During
an

Curve of Sunspot numbers

pl

most of this time its orbit was below that of

t ic
lip

the SSCM in this view. While the SSCM lies

Ec

250 km s –1 1960 Track of SSCM 1954 1944

1950
within the body of the sun most of the time,
1952
1948 there are occasions when the sun wobbles
1954
1958
1956 1955 1946 1944 outside the SSCM. This figure provides an
indication of the extent of its wobble as the
sun moves through space.
Figure 9 The location of the sun’s centre of mass relative to that of the SSCM from 1944 to 1958
Earth to sun chord distance
Sunspot production move backwards in their orbits around the As a result of the sun’s wobble, the chord
The plane of the path of the orbiting plan- SSCM. length between the earth and the sun and
ets and the sun must be at 45° to the line The net result is that the galactic veloci- the amount of energy received by the earth
of motion of the solar system. This is in ties equal that of the SSCM when the bod- will change accordingly. The next exercise
order to balance the gravitational forces of ies directly trail or lie directly ahead of the is therefore to determine the corresponding
a three-dimensionally balanced group of SSCM. The galactic velocities increase as changes in the distance between the earth
objects travelling at constant forward speed they move forward around the SSCM, and and the sun and thereby the changes in the
relative to that of the SSCM. Each body in they decrease as they move backwards about rate of solar energy reaching the earth. This
the solar system will follow a three-dimen- the SSCM. The galactic velocity of each is amenable to precise calculation.
sional spiral track around the SSCM thus body in the solar system, including the sun, The calculation of the chord length
maintaining the group’s constant forward therefore alternately accelerates and deceler- between the earth and the sun at any par-
speed. This path will also be influenced by ates within the galactic plane as it orbits the ticular time has two components. The first is
the changing positions of the major planets SSCM. This is the crux of the issue. Once the position of the sun relative to the SSCM
relative to one another and the sun’s recip- it is appreciated that the reference system is at that time. The second is the elliptical path
rocal movement. the galactic plane and not the plane of the of the earth about the SSCM.
All bodies of the solar system therefore solar system, then everything else falls into The sun’s displacement from the SSCM
have a combination of two velocities. The place. changes relatively slowly but the ecliptic
dominant velocity component is the con- Sunspot production is a direct function direction of the earth about the sun changes
stant galactic velocity that is followed by of the sun’s galactic acceleration and deceler- with the seasons. Figure 10 shows the dis-
the SSCM. The orbital velocities of the indi- ation, with sunspot minima occurring when placement of the position of the sun from
vidual bodies around the SSCM are super- the sun is directly ahead or trailing the the SSCM during 1993 and its effect on
imposed on the galactic velocity. As they SSCM. There can be no doubt that it is the variations in solar energy received on earth
orbit the SSCM their net forward velocity influence of the changing relative positions during that year.
will be the galactic velocity plus the orbital of the major planets that is the direct cause The Intergovernmental Panel on Climate
velocity (corrected for the 45° slope of the of sunspot activity. The actual mechanism Change (IPCC) (2001) dismisses the view
solar orbits) as they move forward in their for sunspot production as a result of galactic that solar activity has a meaningful influ-
orbits around the SSCM, and the galactic velocity changes has yet to be determined, ence on global climate. The basis for this
velocity minus the orbital velocity as they although several theories exist. view is that variations in the receipt of solar

42 Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007
 mathematical analyses. The other authors
N Perihelion
have also undertaken extensive studies on
this subject.
Direction of We believe that our combined stud-
Received wattage earth in
1 417 January ies have made a major contribution to the
advancement of international science in
this field. Our findings are reproducible by
anybody with sufficient knowledge in these
fields.

147,3 x 106 km Uncertainties

A study of the literature shows that there
are still large uncertainties in three related
Received wattage Received wattage issues. These are the following:
1 365 1 385 ■ The physical causes of the regular sunspot
Spring Autumn
150,2 x 106 km activity
equinox 149,1 x 106 km equinox
Direction of ■ Reasons for the different climatic respons-
Direction of
earth in earth in es to the alternating sunspot cycles
April September ■ The mechanisms that link changes in
sunspot activity with corresponding
changes in climate
151 x 106 km We believe that we have thrown light on
answers to the first two questions and that
he outstanding uncertainties relating to
causal linkages do not invalidate the results
of the fundamental studies described in
this paper.

Received wattage Direction of Recommendations

1350 earth in It is extremely important that all those
June
Aphelion involved with water resource studies should
appreciate that there are fundamental flaws
Figure 10 Received wattages per square metre for the year 1993, based on the calculated 0,63 106 km solar in current global climate models used for
displacement from the SSCM climate change applications. These models
fail to accommodate the statistically signifi-
activity are too small to account for varia- from orbiting satellites and other sources cant, multiyear periodicity in the rainfall
tions in the climatic responses. These vari- that are incomplete in both space and time. and river flow data observed and reported
ations were determined from satellite and Furthermore, the calculated wattage changes by South African scientists and engineers
other observations. What the IPCC scientists are appreciably greater than those derived for more than the past 100 years. They also
failed to appreciate is that changes in the from the sources quoted in the literature. failed to predict the recent climate revers-
level of solar radiation received on earth are As can be seen in figure 8, these changes als based on Alexander’s model (Alexander
amenable to precise calculation. The varia- are not regular in time. They were relatively 1995b, 2005a). The global climate model
tions are well in excess of the IPCC value of unchanged from 1979 to 1985, and again outputs can therefore not be used for adap-
+0,3 Wm–2 quoted earlier. from 1995 to 2000. They changed rapidly tation studies.
Although the chord length is of para- from 1986 through to 1994, when they All those who have a genuine interest
mount importance on a month-by-month closely orbited the SSCM. in the prosperity of our country and the
or year-by-year basis, whereby the actual welfare of its peoples are urged to assist in
wattage received by the earth can be deter- the development of joint strategies to opti-
mined, it does not compare with the accu- CONCLUSIONS mise future water resource development and
mulative effect of high or low wattages in a management.
given direction on the ecliptic plane, which Transparency and reproducibility Those who have expressed concerns
determines the type of weather that certain This paper is a brief summary of the results regarding the environmental consequences
parts of the globe will experience for a given of our studies. Alexander’s studies are associated with water resource development
21-year period. described in his comprehensive, 474-page should appreciate that the provision of water
The cold periods that were experienced technical report entitled Climate change to meet rising demands is essential for the
by the northern hemisphere in the 1400s, and its consequences – an African perspective prosperity of any nation. South Africa does
1600s and 1800s would have been coun- (Alexander 2006). It includes 51 tables, 33 not have the luxury of abundant water sup-
terbalanced by opposing hot spells in the figures and 218 references. The data used in plies. The provision of water and conserva-
southern hemisphere. This is because of the report were obtained from readily avail- tion of the water environment are non-com-
the reciprocal effect of the changes in chord able records published by the responsible mensurate objectives in that one cannot be
length. national authorities. The report is available met without sacrificing the other. Reasoned
Probably the most important short- on CD. compromises will have to be made. This
coming in current climate change science Bailey’s studies are detailed in his can only happen if all parties have a sound
is the failure to appreciate that variations 88-page book The sunspot mystery – who knowledge of the multiyear properties of
in received solar energy are amenable to is listening? (Bailey 2006). It includes 23 rainfall and river flow. It is hoped that the
precise calculations, instead of attempting detailed figures. His calculations are based information in this paper will go a long way
to derive these changes from observations on well-established theoretical physics and towards meeting this requirement.

Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007 43
REFERENCES Plymouth, Devon: Frederick Bailey (ISBN 978-0- Lamb, H H 1983 Update of the chronology of assess-
Alexander, W J R1978. Long range prediction of river flow – a 9551202-1-3). ments of the volcanic dust veil index. Climate Monitor,
preliminary assessment. Technical Report TR80. Pretoria: Bredenkamp, D B 2000. Groundwater monitoring: a critical 12:79–90.
Department of Water Affairs. evaluation of groundwater monitoring in water resources Lockwood, M, Stamper, R, Wild, M N, Balogh, A and Jones,
Alexander, W J R and Van Heerden, J 1991. Determination of evaluation and management. Water Research Commission G 1999. Our changing sun. Astronomy & Geophysics,
the risk of widespread interruption of communications due to Report No 838/1/00. 40(4).
floods. Report of the Department of Transport Research, Chatfield, C 1982. The analysis of time series. London/New Mandelbrot, B B and Wallis, J R 1968. Noah, Joseph,
Project No RDAC 90/16. York: Chapman and Hall. and operational hydrology. Water Resources
Alexander, W J R 1994. Alexander, Anomalies in the Commission of Enquiry into Water Matters 1970. RP Research,4(5):909–918.
stochastic properties of river flow and their effect on 34/1970. Pretoria: Government Printer. Mason, P J 1993. Hydrology or climate change, a case
reservoir yield. In Jan-Tai Kuo and Ko-Fei Liu (eds), Government Gazette 1966. No 1480 of 1 July. Pretoria: study. Dams and Reservoirs (Journal of the British Dam
Proceedings of the Republic of China – South Africa Government Printer. Society), 3(2).
Bilateral Symposium on Water Resources, Taipei. Hurst, H E 1951. Long-term storage capacity of reservoirs. Mason, P J 1998. Sunshine over Lake Victoria. The Chemical
Alexander, W J R 1995a. Detection of climate change. Transactions of the American Society of Civil Engineers, Engineer, 6 August 1998.
In Proceedings of the IGBP Conference on Global Paper 2447. Mason, P J 2006. Lake Victoria: a predictably fluctuating
Environmental Change – Implications for Southern Hurst, H E 1952. The Nile. London: Constable. resource. Hydropower & Dams, 13(3).
Africa. Pretoria. Hurst, H E 1954. Measurement and utilisation of the water Rippl, W 1883. The capacity of storage reservoirs for water
Alexander, W J R 1995b. Floods, droughts and climate resources of the Nile Basin. Proceedings of the Institution of supply. Proceedings of the Institution of Civil Engineers,
change. South African Journal of Science, 91:403-408. Civil Engineers, volume 3, part III, pp 1–26, April 1954: 71:270–278.
Alexander, W J R 1997. Predictability of widespread, discussions pp 26–30, correspondence pp 580–594. Rodier, J A and Roche, M 1984.World catalogue of maximum
severe droughts, and their effect on water resource Hutchins, D E 1889. Cycles of drought and good seasons in observed floods. International Association of Hydrological
management. In Proceedings of the 5th International South Africa. Wynberg Times Steam Printing Office. Sciences Publication No 143.
Conference on Southern Hemisphere Meteorology and Intergovernmental Panel on Climate Change 2001. Rodrigues-Iturbe, I and Yevjevich, Y 1968. The investigation
Oceanography. Pretoria. (Invited guest presentation.) Summary for policymakers. of relationships between hydrologic time series and sunspot
Alexander, W J R 2002a. Climate change – the missing Intergovernmental Panel on Climate Change 2007. numbers. Colorado University Hydrology Paper No 26.
links. Science in Africa, September 2002. Summary for policymakers. Fort Collins.
Alexander, W J R 2002b. Statistical analysis of extreme Kite, G W 1982. Analysis of Lake Victoria levels. Tyson, P D 1987. Climatic change and variability in Southern
floods. Journal of the South African Institution of Civil Hydrological Sciences Journal, 27(2):99–110. Africa. Cape Town: Oxford University Press.
Engineering, 44(1):20–25. Klemes, V 1974. The Hurst Phenomenon: a puzzle? Water Wallis, J R and Matalas, NC 1971. Correlogram analysis
Alexander, W J R 2004. Climate change – there is no need Resources Research, 4. revisited. Water Resources Research, 7(6).
for concern. Science in Africa, April. Available at http:// Kokot, D F 1948. An investigation into the evidence bearing Wallis, J R and O’Connell, P E 1973. Firm reservoir yield.
www.scienceinafrica.co.za/2004/april/climate.htm. on recent climatic changes over southern Africa. Irrigation How reliable are historic hydrologic records? IBM Research
Alexander, W J R 2005a. Development of a multi-year Department Memoir. Pretoria: Government Printer. RC 4298.
climate prediction model. Water SA, 31(2). Available at Kruger, A C and Shongwe, S 2004. Temperature trends Weather Bureau (now South African Weather Service)
http://www.wrc.org.za/downloads/watersa/205/Apr- in South Africa: 1960–2003. International Journal of 1972. Climate of South Africa: Part 10, District rainfall and
05/1788.pdf. Climatology, 24:1929–1945. the annual march of rainfall over Southern Africa. WB 35,
Alexander, W J R 2005b. Linkages between solar activity Lamb, H H 1966, Climate in the 1960s. The Geographical Pretoria: Weather Bureau.
and climatic responses. Energy & Environment, 16(2). Journal, 132(2):183–212. World Data Centre for the Sunspot Index 2005. Yearly sun-
Alexander, W J R 2005c. Flood alert. Issued in November Lamb, H H 1970, Volcanic dust in the atmosphere; with a spot numbers from 1750 to 2004. Available at http://sidc.
2005. chronology and assessment of its meteorological signifi- oma.be.
Alexander, W J R 2006. Climate change and its consequences cance. Transactions of the Royal Society London, A, 266 Yevjevich, Y 1968. Misconceptions in hydrology and their
– an African perspective. Technical report (available (1178):425–533. consequences. Water Resources Research, 4(2).
on CD). Lamb, H H 1977. Volcanic activity and climate. Climate Yevjevich, Y 1972. Stochastic processes in hydrology. Water
Bailey, F 2006. The sunspot mystery – who is listening? Monitor, 6:54–67. Resources Publications, Fort Collins, USA.

44 Journal of the South African Institution of Civil Engineering • Volume 49 Number 2 June 2007