A new Old Weather project

Old Weather – WW2 is now an active Zooniverse project. The purpose of this project is to recover hidden marine weather data recorded in U.S. Navy ships’ logbooks during World War II. Like all Old Weather projects, these data will be used to drive sophisticated computer models that help us understand and reconstruct weather and climate in extraordinary detail. But there is another goal that is just as vital – to uncover the source of a mysterious distortion in sea-surface temperature data collected during the war.

This distortion, which Chan & Huybers refer to as the World War II Warm Anomaly, may not be physical in nature, but, they argue, arises from earlier bias adjustments applied to correct for differences in sampling methods. The age-old technique of hauling up a bucket of seawater then taking its temperature on deck produces a different value compared to those collected from an engine-room cooling water intake, for example. There may even be differences in values because of the characteristics of particular ships, especially how deep they sit in the water (their draft). So knowing the fleet composition within the data set, and the specifics of how the measurements were made is crucial to teasing out an answer.

Because of the work we’ve been doing with the U.S. National Archives over the past eight years or so we are in a unique position to help investigate this question – we have digital images of many of the original U.S. Navy logbooks. The logs we are transcribing in this project were selected because the ships were often in the same place at the same time, even moored alongside each other in nests. All but two survived the entire war, 1941-1945. Twelve were based at Pearl Harbor in 1941, eighteen were in the Aleutian Islands in 1942-1943, and ten were caught in Typhoon Cobra in December 1944. This opens up many opportunities to investigate sources of bias, from factors associated with different ship types, the weather instruments in use at different times, or changes in methods required by wartime operations (such as blackout for example). It will also be possible to investigate how tropical and sub-polar environments may have influenced the data in different ways.

Moreover, with logbooks in hand, we can also better understand legacy issues associated with U.S. Navy data key-punched onto IBM cards in the early 1950s and subsequently migrated into today’s widely used International Comprehensive Ocean Atmosphere Data Set (ICOADS). Already we have found out that the fleet composition represented in ICOADS is tilted toward submarines, destroyers, and destroyer escorts. Aircraft carriers, with probably the best-equipped and best-trained weather personnel in the Navy, appear to be mostly neglected.

The World War II Warm Anomaly is large enough to appear in the long-term global mean sea-surface temperature record. This presents as an aspect of natural variability that may not in fact be real. If this turns out to be true, the corrected temperature record would appear to evolve upward more smoothly over time, and correspond more closely to model results as described by Chan & Huybers.