Struggling to cope without computers

Originally published 5 January 1987

My word pro­cess­ing com­put­er is on the fritz, so I am writ­ing this col­umn with a pen­cil on a pad of paper. It took me a long time to find the pen­cil. I had to buy the pad of paper.

Some lit­tle giz­mo in my com­put­er went pop and sud­den­ly I’m back in the Stone Age. Or it feels like the Stone Age. It has been so long since I’ve used a pen­cil that it might as well be a quill pen, or a piece of char­coal on the wall of a cave. It is hard to believe there was ever a time when get­ting the right words onto paper was such a labo­ri­ous craft.

I want to write about a book called Weath­er Pre­dic­tion by Numer­i­cal Process, by Lewis F. Richard­son. It is a sto­ry that will illus­trate our grow­ing part­ner­ship with machines. The book was first pub­lished by Cam­bridge Uni­ver­si­ty Press in 1922. My copy is a Dover reprint, pub­lished in 1965, and for most of what fol­lows I am indebt­ed to mete­o­rol­o­gist Syd­ney Chap­man, who wrote the intro­duc­tion to the Dover edition.

The sto­ry begins at a vil­lage called Eskdale­muir in the Scot­tish low­lands. To this place of “bleak and humid soli­tude,” Lewis Richard­son went in 1913 to super­in­tend a mete­o­ro­log­i­cal sta­tion for the British Mete­o­ro­log­i­cal Office. In his off-hours, he devel­oped ideas about a math­e­mat­i­cal way to pre­dict the weather.

Calculating the weather

Richard­son knew that changes in the atmos­phere hap­pen in accor­dance with well-known laws of physics. From obser­va­tions of the state of the weath­er at a cer­tain moment over a broad area of the globe, it should be pos­si­ble, in prin­ci­ple, at least, to cal­cu­late math­e­mat­i­cal­ly the state of the weath­er at some future time. Richard­son set down the equa­tions of pres­sure, tem­per­a­ture, and wind veloc­i­ty as they apply to the atmos­phere, and began to work out a way to test his ideas.

The Great War intrud­ed. Richard­son was a Quak­er, and his reli­gious con­vic­tions pre­vent­ed him from tak­ing the role of a com­bat­ant in the war. He was torn, he wrote, “between an intense curios­i­ty to see war at close quar­ters and an intense objec­tion to killing peo­ple.” He joined an ambu­lance unit, and in 1916 – 1918 was in France trans­port­ing wound­ed sol­diers. Dur­ing that ter­ri­ble time, at odd moments snatched “on a heap of hay in a wet rest bil­let,” he put his ideas of weath­er pre­dic­tion to the test.

For the pur­pos­es of cal­cu­la­tion, he divid­ed a map of part of cen­tral Europe into a grid of 25 squares, each 20 kilo­me­ters on a side. From pub­lished obser­va­tions of pres­sure, tem­per­a­ture, humid­i­ty, and wind veloc­i­ty at sta­tions with­in the cells of the grid, he com­put­ed for a place near the cen­ter of the grid the state of the weath­er three hours before and three hours after the time of the obser­va­tions. These the­o­ret­i­cal results he com­pared to the weath­er that had actu­al­ly been record­ed. The agree­ment was not good.

But Richard­son knew what was wrong. First, bet­ter obser­va­tions were need­ed over a wider part of the Earth­’s sur­face. He imag­ined a grid of 3200 squares cov­er­ing most of the globe, each with a mete­o­ro­log­i­cal sta­tion near its cen­ter. Many of the sta­tions, of course, would have to be ships. The cost of estab­lish­ing and main­tain­ing such a net­work he esti­mat­ed at $1 bil­lion a year. It was a huge­ly imprac­ti­cal sum.

The busi­ness of com­pu­ta­tion was even more imprac­ti­cal. For math­e­mat­i­cal weath­er pre­dic­tion to work, the cal­cu­la­tions would have to be ready and dis­trib­uted well in advance of the time to which they referred. To do this with pen­cil and paper (the only option at that time), Richard­son fig­ured, would take an army of 64,000 math­e­mat­i­cal clerks, all cal­cu­lat­ing in unison.

Here was a man who was way ahead of his time.

Lost in the coal

Nev­er­the­less, the Roy­al Soci­ety, the most pres­ti­gious sci­ence orga­ni­za­tion in Britain, con­tributed $500 toward the pub­li­ca­tion of Richard­son’s math­e­mat­i­cal the­o­ry. Dur­ing the bat­tle of Cham­pagne in April 1917, he sent a copy of the man­u­script to the rear, where it was lost, to be dis­cov­ered months lat­er under a heap of coal. With lat­er revi­sions, this was the man­u­script pub­lished by Cam­bridge Uni­ver­si­ty Press in 1922.

Richard­son’s vision­ary dream has come true. His math­e­mat­i­cal the­o­ry is at the heart of present-day efforts of achieve reli­able long-range pre­dic­tion of the weath­er. A world­wide net­work of ground weath­er sta­tions and atmos­phere-sens­ing satel­lites pours rivers of data into the offices of nation­al weath­er ser­vices here and abroad. High-speed dig­i­tal com­put­ers crunch the num­bers through the math­e­mat­i­cal equa­tions and pre­dict (with grow­ing suc­cess) what the weath­er will be days and weeks in the future. The cur­rent gen­er­a­tion of super­com­put­ers accom­plish­es in min­utes what Richard­son’s imag­ined army of 64,000 math­e­mat­i­cal clerks might have done in a day.

Even my per­son­al com­put­er could keep up with Richard­son’s pen­cil-wield­ing army of clerks. It could, that is, if it weren’t on the fritz.