Typhoon forecasting, 1944, or, the making of a cynic

1. Forecasting: Typhoon I

In The Genesis Strategy, Steve Schneider (1976) tells about a conversation he and I had right after we had shared a White House staff briefing. It was June 1974 and full awareness of the climatic events of 1972 had still not quite gotten into the official skull. The Sahelian drought relief action had to be quickly put together even though the drought had been going on for several years. The Russian drought followed by their huge grain purchases of 1972 and 1973 had depleted U.S. reserves badly, and the purchases had been made at very low U.S. subsidized prices. Climatic problems had popped up in dozens of countries. While the National Oceanic and Atmospheric Administration pretended that if a problem really existed they had things under control, as did the U.S. Department of Agriculture, the intelligence agencies did not agree and had arranged a briefing.

At the briefing I had spoken first in my usual freshman class teaching style telling what my colleagues and I knew about climatic change and climatic variations. After all, I had been teaching that specific subject for 26 years, about as long as Steve had been alive. Steve had then spoken about the dynamics of climate, as he saw it, and the problems of modeling such a complex system. He spoke of how much we did not know. At any rate, while having a drink afterward Steve commented that he did not feel "the science" was ready to be as positive as I seemed to be about forecasting climate, and that it was dangerous to make a forecast with such uncertainties at its base.

I responded rather vehemently with a story from World War II, when I had to make forecasts for the 20th Air Force with a lot less weather data than we have in terms of climatic data today, and much less knowledge of the subject. I said one must judge by the hazard generated by making no forecast when the airplane and crew had to go anyhow. Did he mean dangerous to the user or dangerous to the reputation of the forecaster? I had to reach an accommodation with myself, back then, for a bad forecast cost lives and there would inevitably be bad forecasts. I had killed men with a bad forecast and he had not.

I guess I must have at least reached Steve with my conviction, because he did remember the incident well enough to mention it in his book. Actually the discussion could not have taken as long as it takes to read this description of it. He could not have possibly known the full depth of my feelings about the ethics of forecasting or the wartime traumas that produced those feelings.

The first of my real-world forecasting experiences dealt with a typhoon. By this I really mean the first actually operational forecast I was ever party to.

In the summer of 1944 I was assigned to the 21st Wing of the 20th Air Force. The particular group of officers that I was to be part of started to gather at the Pentagon around the end of June. Ultimately it became clear that we were headed for the Pacific to become the weather central of the 20th Air Force. It was an odd group. There was hardly an experienced forecaster among them, other than E. Brewster Buxton. Bux did have several years of experience forecasting for Pan American Airways in Latin America and the Pacific. It was during his period of work in Fiji that he acquired the nickname "Bathless," but that's another story.

The rest of us were relatively new products of the "meteorology cadet" training programs that had been set up at five universities under the overall direction of the now mythical Carl-Gustaf Rossby. Rossby himself probably never made an operational forecast, but his qualifications as a theoretician and operator were impeccable. I had gone from being a geology graduate student at the University of Wisconsin to being a "meteorology cadet" in the air corps program at the University of Chicago in early 1942. Upon commissioning in late 1942, I was made an instructor in the same program. In the fall of 1943 1 was sent to the Institute of Tropical Meteorology at the University of Puerto Rico. I was taught there mostly by those who had entered the field about the same time I had. After a couple of months I "graduated" to teaching the same program. With still no practical forecasting experience, I was assigned to the group that was to forecast for the 20th Air Force.

In July we headed west, individually and in small groups, gathering at Hamilton Field, California. There I made a forecast with a real payoff-but it was hardly the shape of things to come. Some of the group had been playing a particular slot machine until they ran out of nickels. I brashly stepped forward, telling them to step aside so that I could show them how to win. I put in one nickel, collected the jackpot, then put in two more and collected the other jackpot. I made a few bucks, but no friends.

In Hawaii we were delayed by the simple fact that our home to be, Saipan, had not yet been taken by the combat forces. So our colonel, Bill Stone, arranged for us to work at the Fleet Weather Central at Pearl Harbor. They at least had been practicing forecasting for some time. About the first morning I went there to work I found myself assigned, along with Capt. Bill Plumley, to working with a lieutenant junior grade (j.g.). who had been one of our students back at Chicago. The problem for the day was simply stated-make a forecast for a carrier-based strike against Marcus Island the following day.

We looked at the weather map, looked at each other, looked back at the map. There was really nothing to say. It was nearly a blank map. I think we felt a little like Magellan setting out into the western sea.

In those days a typical weather map had data for the western parts of North America, most of the United States-held islands (there were not many west of Hawaii), Alaska, a few low quality Russian and Chinese bits and pieces on the western rim of the map, and a scattering of ship, submarine, and aircraft reports. On the weather map that day there were seven areas larger than the United States that had no data whatsoever.

There was one thing about the weather that was clear that day. Off to the southeast of Marcus Island there was a typhoon. That great storm and its path during the next day was the problem to be solvedwith very little data to work with. The few bits of data showed that there was indeed a cyclonic whirl of winds and dense cloud several hundred miles across, feeding a vast layer of high clouds, but showed none of the details of the inside of the storm. How strong the winds were, how large the waves, and whether it had developed an eye could only be imagined. None of us, at the time, had been inside a typhoon.

Though the tropical cyclonic storms like that one have different names over different oceans, the names all mean "big wind"-the name "hurricane," used in the Atlantic, comes from an Arawak word; "typhoon," the North Pacific name, came from the Chinese; and the "willy-willy" of northern Australia was Aboriginal.

By convention, thanks to Admiral Beaufort of the British navy, the word hurricane also meant a specific wind force-"that which no canvas can withstand." The nineteenth-century wind force scale, which bears Beaufort's name, was translated in the days of steam into specific wind speeds. The canvas of 1874 could not withstand winds of 75 miles per hour or higher. Nowadays a tropical storm is not called a hurricane or typhoon until the winds near the center of the storm reach Beaufort's scale of 12 (or 73 mph). The winds in a typhoon may reach double that speed.

We did not know the wind speeds in that storm, but we did know that high winds and waves, coupled with dense cloud and heavy rain, made a carrier operation hazardous at best. The forecast really came down to where the storm was going to go.

Ordinarily, in the trade wind regions, the upper winds are also from the east and the storm is steered westward. But was this the case on that day? There were no upper-air data in the area. If an intense storm were crossing the northern part of the ocean, its trough might be deep enough for the typhoon to recurve northward toward Marcus Island and the fleet. But there were no data from the northern area.

Finally we decided we would have to construct an upper-air chart, somehow. There were not very many observations, but there were a few wind observations from near the 10 000-ft level. And we had the basic knowledge of atmospheric physics we had learned in school. If we knew the weight of the lowest 10 000 ft, we could subtract that from the surface pressure reported from ships, submarines, and islands and obtain the 10 000-ft pressure pattern, so we would then know which way the upper winds would steer the typhoon.

But what was the weight of the lowest 10 000 ft of air? All we had was the surface pressure, surface winds, surface temperature, and the cloud amount and type.

And the cloud amount and type! There was our key, for we knew that in tropical regions the cloud type depended on the lapse rate. If we knew the surface temperature and that rate, we could get our upper-level pressure map-or a reasonable facsimile.

Digging out all the data on file, we determined the average lapse rate for each cloud condition and calculated the upper pressures. Plotting the results on a map, a picture of the wind flow at 10 000 ft emerged. Our hearts sank. There was a trough approaching from the west, exactly in the wrong position. The typhoon would recurve to rendezvous with the fleet at about strike time.

We discussed. We recalculated. We considered and reconsidered. Finally, by late afternoon we were convinced, and the duty officer wrote out the forecast for the admiral, calling for the recurved typhoon to be in a most awkward position. A brief hesitation and look of "Well, here it goes!" before transmittal.

At that point the senior aerologist (as the navy called meteorologists) walked into the room, drunk as a lord. Naturally the duty officer, especially if a j.g., does not transmit a message to the admiral when the senior officer is present. Captain Lockhart blearily looked at the message. When his eyes focused he roared out, "Nonsense! Typhoons don't recurve at that longitude at this season, they move straight west! Change that-forecast."

"Yes, sir!" is all a junior officer can say to a navy captain when given a direct order. And two air corps captains, visitors at that, could only keep their mouths shut. The forecast was changed.

The typhoon recurved. The fleet encountered it. There were losses of planes and brave men. I never saw the j.g. again. I often wonder how he felt the next day. When we reported the incident to Colonel Stone he just said, "Guess we better not try to work there any more."

Our forecast had been correct, but not all subsequent ones were. Still the total disregard of a forecast based on a combination of good science and some data in favor of a totally subjective, off-the-cuff opinion made a strong and lasting impression. It was not to be the last such experience.

2. The Caine Mutiny typhoon: December 1944

The typhoons that I had worked on as a forecaster with the 20th Air Force had convinced me of how significant they were to all sorts of operations. The typhoon that I had encountered as my first real-life forecasting experience, where the result was destruction of aircraft and lives; the storm that simply held up the progress of operations; and the others that were threats but did not materialize made it very clear that we had to watch typhoons very carefully. So any little indication in the data that there might be a typhoon had to be analyzed and carefully watched to see whether anything really would materialize.

The available data were very sparse. For example, off to the east the nearest weather station was about a thousand miles away on the island of Eniwetok. We also had a few aircraft observations in between there and Saipan.

Once I noticed a little shift in the wind at Kwajalein Island, which was still farther to the east, and more than the usual amount of cloud there. This suggested that there was some kind of a low pressure center south of Kwajalein, so I marked it on my weather map as a "potential" typhoon. Then when I got a new set of observations, I looked for evidence of that particular potential typhoon developing. There would be an occasional bit of an indication in an aircraft observation-bigger than usual clouds, a little different than usual wind direction. If the trade winds got around to the southeast, that was unusual. I would watch for a change in the wind direction from being more northerly and stronger to more southerly-because that would suggest that something other than the usual rippling of the trade winds was going on.

When the next station to the east, or an aircraft observation, would show that pattern of a little shift of the wind to the north or a little more shift to the south and bigger than usual clouds, I would assume that perhaps some kind of an incipient cyclone just moved by to the south of that place, and I would mark it on the map.

For 10 days, I had watched the maps as bits and pieces of data suggested the westward movement of a tropical cyclone. Then the data from Guam, which was a good weather station, started to show a rather strong northerly component to the wind and the weather there clouded up. We could see a veil of high-level cloudsa streaky veiling of cirrus clouds off in the east and southeast. These fingers of clouds from far off to the east had been known to mariners for centuries as the forerunner of a typhoon.

When I saw those, I felt quite certain there was a typhoon developing. As the weather at Guam turned quite bad and the wind shifted abruptly to the southeast, I became convinced that a typhoon had passed just to the south of Guam.

I knew that the fleet was assembled off to the west and northwest, between Saipan, the Philippines, and Okinawa. There were a lot of aircraft, carriers, battleships, cruisers, destroyers, oilers, etc., out there. It was really very important to know whether that typhoon south of Guam was going to continue on west toward the Philippines or recurve and turn northward toward the fleet.

I was particularly aware that once before I had been involved in forecasting a typhoon that did recurve rather than going straight west.

As far as I could tell from the weather map on 17 December, there was a good chance of a trough of low pressure approaching from the west that could cause the typhoon to recurve, so I sent a reconnaissance aircraft out to find the center of that typhoon. The mission was to see if it was indeed there, and if so, to find out how intense it was and its exact location. With an exact location, I could tell whether it had started to recurve or not.

Fortunately, we had the authority to ask for reconnaissance aircraft, and the reconnaissance aircraft took off almost immediately. After some hours we got a radio message back from them, uncoded, saying that they had located the eye of the typhoon. They gave its latitude and longitude, and estimated the wind speed at 140 kt at the surface.

This was nearly twice as strong as the minimum winds needed to call it a typhoon, and strong enough that not only could no canvas withstand it but many ships could not withstand it-even steel ships driven by modern power plants.

I immediately went to the Teletype connection we had with the Fleet Weather Central and typed a message to the effect that we had observed a typhoon at such and such latitude and longitude, with an estimated wind speed of 140 kt near the center. Previous position suggested that it was recurving to the northwest. I assumed that the fleet knew where the fleet was and that this would be on a track toward the fleet.

At the other end, some unknown navy aerologist typed back, "We don't believe you."

I was shocked despite my previous experience with the navy and typhoons. So I typed back that it was not a guess-that we actually had a reconnaissance aircraft out there and that the radioman had just reported that they were in the eye of the storm and that it was really there.

The reply came back, "We still don't believe you, but we'll watch."

Well, the next day that storm and the fleet did rendezvous. I have talked to several people since who were with the fleet at the time. I asked some of them about their experiences in the storm. Commander Rex told me that the serologists of the fleet eventually decided on their own that they were close to a typhoon. They sent a message to Admiral Halsey saying that in their joint estimation, the fleet was headed toward a typhoon and would rendezvous with it at a certain time. Halsey's alleged reply was, "I don't believe any serologist. Maintain present course."

Halsey was wrong. The serologists were right. Our observation was right. The fleet did rendezvous with the typhoon (Calhoun 1981). As I recall, 250 aircraft were destroyed, and 4 destroyers were sunk. Nine hundred people died because the admiral did not believe the forecast of his own serologists. There has been a novel written, probably based on that storm, called The Caine Mutiny: A Novel of World War lI (Wouk 1951).

That disaster could have been avoided. It was not something that just popped up without any observations. I had followed it for 10 days. We had sent a reconnaissance aircraft out to look at it. We found out where it was. We did tell the fleet. The fleet aerologists did tell the admiral, but he still continued to go right into the storm-with disastrous results.

About six years later, I told the story of this typhoon to my meteorology class with the idea of impressing on them that when they make operational forecasts they are dealing with life and death matters. They should not take it lightly.

Auditing the class at the time was a navy captainthen commandant of the naval ROTC at the University of Wisconsin. At the end of the class, the captain's face was as white as a sheet.

I said, "Captain, what's the problem?"

He said, "I commanded the Pittsburgh. I was in such a storm. That is why I'm not an admiral now. The front end of the Pittsburgh was torn off and they had to find a scapegoat."

I never saw him in class again. I think it was too painful for him to hear stories about something that was not his fault and could have been avoided.

I was sitting in my office one day, a year or so later, when the phone rang and a voice at the other end of the phone said, "Bryson?" "Yes."

"Are you the Bryson who was on Saipan in 1944?"

"Yes."

"Did you send a message to the Fleet Weather Central about a typhoon?"

"Yes."

"Well, I was the radio operator who received your message, and I passed it on to the fleet. I just wanted you to know that it didn't stop with me."

All these years, and it was still on his conscience that something could have been done about those dead seamen and was not. And it is still on mine too.

Somehow we could not get across that the presence of the typhoon was not a guess. It was not some vague anticipation of the future. It was a scientific observation on the spot. Ignoring science cost a lot of American lives and money. The same is true today. A bad forecast on a global scale might cost trillions of dollars and whole economies.

[Reference]

References

[Reference]

Calhoun, C. R., 1981: Typhoon: The Other Enemy. Naval Institute Press, 247 pp.

Schneider, S. H., 1976: The Genesis Strategy. Plenum Press, 419 pp.

Wouk, H., 1951: The Caine Mutiny: A Novel of World War II. Doubleday, xiv + 494 pp.

[Author Affiliation]

Reid A. Bryson

Center for Climatic Research, University of Wisconsin-Madison, Madison, Wisconsin

[Author Affiliation]

Corresponding author address: Dr. Reid A. Bryson, Center for Climatic Research, University of Wisconsin-Madison, 1225 West Dayton St., Madison, WI 53706.

E-mail: rabryson@facstaff.wisc.edu

In final form 2 May 2000.

2000 American Meteorological Society