EARLY SPACE WEATHER FORECASTING

Space Weather John Kennewell

Phil Wilkinson

An early version of these notes was presented
at the 2011 IUGG conference in Melbourne

INTRODUCTION

The prologue to the Space Weather Forecasting History session of the 2011 IUGG conference in Melbourne Australia stated:

Long before the term "space weather" was invented, the need to forecast the effect of extra-terrestrial agents on terrestrial systems was recognized. The first attempts at forecasting were probably testing scientific hypotheses. When the global nature of the processes was recognized, exchanges were arranged to distribute data needed to make forecasts as well as the forecasts themselves. The International URSIGRAM and World Day Service (IUWDS) carried out that function for many years and had a change of name to the International Space Environment Services (ISES) to better reflect this role. For many, the functional history of forecasting starts in 1956-57 - the International Geophysical year, when global forecasting agencies around the world coordinated their forecasts in an effort to help researchers make additional observations during and before disturbances of the Earth's environment. Some agencies came into being at this time; others started to recognize their role. The success of this era was built on earlier efforts to forecast the effects of the ionosphere on HF radio and changes in the Earth's magnetic field. The 1979 Boulder Solar-Prediction Workshop was a watershed and the follow-on workshops brought many in the field together to exchange ideas. This session will explore the development of space weather forecasting from its inception, when it was little more than a scientific curiosity, to the present day, when it is a recognized service.

The term "space weather" used to describe the state and dynamics of the solar-terrestrial environment is of very recent origin. We believe it was first used in a research paper in 1988 by Tsugunobi Nagai of the Japanese Meteorological Research Institute. The paper was published by the American Geophysical Union (AGU) in their journal Geophysical Research Letters (GRL) and was titled ' "Space Weather Forecast": Prediction of Relativistic Electron Intensity at Synchronous Orbit '. The term was later adopted in the USA to replace the phrase 'solar-terrestrial physics', in order to make the subject more appealing to the public and politicians in the hope of raising awareness and attracting more funding for the discipline. The term quickly spread into widespread use around the globe by the mid-1990's.

However, as mentioned in the quote above, forecasts of 'space weather' phenomena were being made up to a couple of centuries before this time. We discuss the early history of this field in the sections that follow.

THE BIRTH OF SOLAR TERRESTRIAL PHYSICS

Discovery of the sunspot cycle, with a period of around 11 years, is usually attributed to the German astronomer Heinrich Schwabe. Observing the Sun regularly from 1825, he came to the conclusion that sunspots showed a periodic increase and decrease after about 17 years of observations. He announced his findings in 1843. However, this study was not generally known, particularly in the English speaking world, until the publication in 1851 of a massive work called 'Kosmos' by Baron Alexander von Humboldt. Humbolt drew the world's attention to Schwabe's results. These were later confirmed in an extended study of sunpsots by Rudolf Wolf (which is why the sunspot number is sometimes referred to as the Wolf number).

Edward Sabine Edward Sabine (1788 - 1883) was a military officer and scientist who spent much time analysing geomagnetic data from observatories established in Toronto (Canada) and Hobart (Tasmania).

Upon the publication of Schwabes's results in Kosmos, Sabine realised that there was a strong correlation between his magnetic declination disturbance studies and this newly discovered sunspot cycle. In April 1852 he published his deductions in volume 142 of the journal Philosophical Transactions. He noted that:

"I have had the satisfaction of finding that the observations of these years confirm every deduction which I had ventured to make from the analysis of the former period; whilst new and important features have presented themselves in the comparison of the frequency and amount of the disturbances in the different years, either from a causal connection (meaning thereby there being the possibly joint effects of a common cause), or by a singular coincidence, corresponds precisely both in period and epoch, with the variation in the frequency and magnitude of the solar spots, recently announced by M. Schwabe as the result of his systematic and long-continued observations."

In fact, at least three other researchers had arrived at the same conclusion independently and around the same time. These were Rudolf Wolf and Jean-Alfred Gautier, both in Switzerland, and Johannvan Lamont in Germany. However, probably because of his publication in the prestigious British journal, Sabine is generally given the credit for the birth of Solar-Terrestrial studies. Sabine also went on to discover a link between the Moon and magnetic variations, although his explanation was not correct. It was not a lunar magnetic field that caused the changes, but lunar gravitational tides in the Earth's ionosphere that were responsible. Sabine also noted, in 1856, that auroral displays appeared with the same general period as magnetic disturbances and sunspots.

THE FIRST SPACE WEATHER FORECAST

Pehr Wargentin William Cade (2013) has identified the first 'space weather' prediction as being made by the Swedish astronomer and mathematician Pehr Wargentin. This prediction was reported after the fact by Wargentin in a 1750 paper. His prediction of auroral activity was made possible by the linkage between geomagnetic disturbance and auroral occurrence by earlier Swedish scientists reported in a 1747 paper.

Image credit - William Cade III

There appears to be a gap of over 100 years between this verbal prediction and the first published space weather forecast. This later forecast, addressing solar effects on current technology, was made by William Ellis of the Royal Observatory at Greenwich (RGO) in the year 1879.

William Ellis Ellis joined the RGO in 1841 when he was 13 years old as a computor employed on lunar reductions. He spent a year as an astronomical observer at Durham University in 1852 but then returned to RGO working as a transit circle observer on geodetic measurements. In 1871 he was promoted and transferred at his request to the magnetic and meteorological department, where he worked on the relationship between terrestrial magnetism and sunspots. Ellis wrote a seminal paper in 1880 on the correlation between the two and it was widely accepted as proof of the relation suggested by Sabine in 1852. His last paper "Sunspots and Terrestrial Magnetism" was written at age 88, in 1916.

In his 1880 paper Ellis presented tables and graphs showing how the magnetic parameters of declination and field intensity compared with the Wolf sunspot number for three complete sunspot cycles.

Graph from Ellis paper The graph here shows the last of the three cycles Ellis studied, what we now call sunspot cycle 11 (1867 - 1879). It was a large cycle, reaching a peak smoothed sunspot number of 139. During this cycle there had been a substantial increase in the size and sensitivity of electric telegraph networks throughout the world.

William Ellis was well aware that geomagnetic disturbances could cause problems with the electric telegraph. Sometimes induced currents would cause automatic and random operation of the apparatus and cause distortion of transmitted messages. Occasionally the induced voltages could be large enough to cause electric shocks to unsuspecting operators touching the metal parts of their apparatus.

In 1879 when sunspot numbers were just starting to rise toward the peak of cycle 12, Ellis penned a short note to the Journal of the Society of Telegraphic Engineers and Electricians (volume 8, page 214) in which he informed the telegraphic community that sunspots are correlated with auroral and magnetic activity and that the next sunspot cycle would reach a maximum in 1882 (actual maximum was in 1884). He noted that in recent years there had been little magnetic activity and that telegraphic technology had taken a turn toward even more sensitive apparatus and thus "I would therefore ask whether any of the new apparatus possesses such peculiarity in their principle or construction as would render it more liable than were the older forms to be temporarily deranged or interfered with by earth currents?"

THE FIRST SPACE WEATHER RADIO FORECAST

The Union Radio Scientifique International (URSI) was founded in 1919, under the auspices of the International Research Council, for the study of radio science. It was soon recognised that changes in the space environment (ie space weather as we now call it) affected radio signals. In order to bring such changes to the prompt notice of radio observers around the world, URSI suggested to the French government that a daily service of radio-cosmic bulletins should be broadcast. These bulletins came to be known as URSIgrams.

Eiffel Tower The first such bulletin was broadcast from the Eiffel tower in Paris on 1 Dec 1928. The form of the broadcast was a series of coded groups that were sent in the international Morse code. The groups contained data of solar and terrestrial changes that could affect radio propagation. These broadcasts were received and retransmitted around the world. A long wave station at Lafayette near Bordeaux and a short wave station at Poutoise near Paris were two French stations that contributed to getting these forecasts around the world.

In 1929 the American section of URSI recognised the value of these messages. The broadcast on US soil was made in 1930 from the US Navy station NAA at Arlington, Virginia, near Washington DC. These were not just retransmissions of the French broadcasts but employed data from US institutions. Solar data was provided from Mount Wilson Observatory in California, magnetic observations from the Coast and Geodetic Survey and auroral observations in Alaska from the Carnegie Institute of Terrestrial Magnetism.

WWV radio station In 1937 these broadcasts were switched to the standard time and frequency station WWV at Fort Belvoir, also in Virginia. In 1968 WWV was moved to Fort Collins, near Boulder in Colorado. These broadcasts can still be heard worldwide today on 5, 10, 15 and 20 MHz at 18 minutes past each hour. Today the data is provided by the Space Weather Prediction Center (SWPC) of the US National Weather Service. SWPC is located in Boulder, not far from the WWV transmitters.

THE FIRST QUANTITATIVE FLARE FORECASTING ALGORITHM

One of the desirable qualities of any forecasting method is its ability to make quantitative predictions. We believe that the Australian physicist Ronald Gordon Giovanelli was the first person to derive a method to make quantitative flare forecasts. He published his method in the Astrophysical Journal in 1939 (volume 89, pages 555-567) with the title "The Relations Between Eruptions and Sunspots".

Ronald Giovanelli Working at the Commonwealth Solar Observatory on Mount Stromlo in the years before World War II, Giovanelli used data on 1399 solar flares (which he termed 'eruptions') that occurred during the years 1935 to 1937 inclusive. He investigated the 'statistical relationships between sunspots and the solar eruptions that are associated with them'. His analysis dealt with 'the probability of an eruption in relation to the size, type and development of the associated spot group'. The result of his analysis was a simple formula for the probability of a flare in terms of the type of spot group, its area and rate of increase of area. He also noted that the probability did not depend on the maximum magnetic field in the group nor on the age of the group.

Giovanelli's method is expressed by the formula:

where P is the probability of a flare occurring near a given sunspot group during one 24 hour day, k is a value assigned according to the type of spot group (see table below), A is the total area of the spot group expressed as millionths of the solar hemisphere and f(i) is a function related to the increase (i) in area of the spot group (in millionths of the solar hemisphere over the previous 24 hours). An analytic expression for f(i) is given below:

Spot group typek value
Alpha 0.81 x 10-3
Beta 0.96 x 10-3
Beta Gamma 1.20 x 10-3
Gamma 2.05 x 10-3

f(i) = 0.90 for i<75 (incl neg values of i)
f(i) = (0.70 + 0.0025 i) for i>75

Flare Loops

WALDMEIER'S RULES FOR SUNSPOT PREDICTION

Max Waldmeier Max Waldmeier (1912 - 2000) was director of the Swiss Federal Observatory in Zurich from 1945 to 1979, and was involved in many aspects of solar research, including the reduction of sunspot observations from 30 observatories around the world to form the Zurich sunspot number. This activity was started by Rudolph Wolf in 1855, which is why the sunspot number is sometimes referred to as the Wolf number. Waldmeier retired from the position of director in 1979 and no one at Zurich wanted to take over the arduous task of producing the sunspot number. Eventually the Royal Observatory of Belgium offered to take on the job and from 1980 onward the Zurich sunspot number became known as the International sunspot number. Waldmeier, along with others, investigated various relationships between characteristics of the smoothed sunspot number variation in an attempt to be able to predict the sunspot cycle. His doctoral thesis, completed in 1935, was written concerning the 'laws' that govern sunspot activity. He summarised these in five rules:

Waldmeier also devised formulae to express these rules. He initially derived separate formulae for even and odd cycles, justifying this as the sunspot magnetic polarity would alternate with adjacent cycles.

It should be noted that these five rules are not all independent of one another. In particular, rule 5 implies rule 1. Rule 2 implies both rules 3 and 4.

For prediction purposes, rule 5 is probably the most useful. It expresses rule 1 a little more quantitatively. It is also the most consistent of the rules from cycle to cycle. From this rule we can estimate the maximum smoothed sunspot number and the time of maximum as soon as we can obtain a reasonable value for the average rate of rise. This can normally be computed moderately accurately at around 18 months into the cycle, giving a forecast lead time of anywhere from about 1.5 to 3 years.

EARLY IONOSPHERIC FORECASTING

Following Marconi's success in transatlantic communication in 1901, it was speculated that a reflecting layer in the upper atmosphere was the responsible agency (although studies of geomagnetic variation in the previous century gave a prior indication of the existence of such a layer). JE Taylor in 1903 and JA Fleming in 1906 suggested that this layer was produced by the action of ultrviolet light from the Sun on the Earth's upper atmosphere. This implied solar control of radio propagation. Practical investigations of the ionosphere also showed that geomagnetic events had an effect on radio propagation. Appleton was probably the first to notice this in 1927. The Internation Polar Year of 1932-33 provided a lot of statistical information on auroral and magnetic changes on the ionosphere. Some radio engineers evidently made use of the 27-day recurrence period, due to solar rotation, to provide local predictions of radio propagation quality. However, it was not until World War II that routine forecasts of the ionosphere related to radio propagation were made.

Edward Appleton The English physicist Edward Victor Appleton is one of the earliest researchers who not only proved the existence of the ionosphere and measured some of its properties, but who was also very active in developing theoretical tools to allow quantitative study of ionospheric refraction of radio waves.

With the outbreak of WWII came a desperate need for reliable long distance communication between fixed and mobile stations around the world. Various governments around the world set up groups to provide predictions of the appropriate frequencies to use for these communications. Radio scientists were linked to solar astronomers to help in producing these predictions. In Australia the Commonwealth Solar Observatory at Mount Stromlo near Canberra was the agency to provide radio propagation predictions to the armed services.

The normal circuit frequency predictions were linked to sunspot number and disturbance forecasts were made using knowledge of the 27 day recurrence period in both geomagnetic activity and in solar flare activity (due to asymmetric distribution of sunspot groups around the rotating Sun). John Dellinger, in 1937, had discovered that sudden ionospheric disturbances (SID), which created shortwave fadeouts (SWF), were due to solar flares. The causal agency would later be found to be a soft X-ray component accompanying the visible light increase in a flare. This produces increased ionisation in the D-layer, the lowest region of the ionosphere.

REFERENCES


ASAAustralian Space Academy