Mean-ingless Measures?

by Howard Fienberg
February 1999

Global warming is apparently worse than ever. “1998 was hottest year on record,” according to the Associated Press (Jan. 12). Variations on the theme surfaced in newspapers from the Houston Chronicle to the Cincinnati Enquirer. The National Oceanic and Atmospheric Administration (NOAA) and National Air and Space Administration (NASA) both showed record high global mean temperatures, though each cited slightly different temperatures (NOAA at 58.1, NASA at 58.5 degrees Fahrenheit). But regardless of the numbers offered, problems remain in establishing the current temperature of the earth. How does one establish a climate “mean,” or average, for the entire globe?

NOAA and NASA said global warming since the mid-1970’s “exceeds that of any previous period of equal length since the collection of weather data began about a century ago.” But until recently, that data collection was primarily based in North America and Europe, leaving comparisons for the rest of the earth somewhat sketchy.

According to the Associated Press (Feb. 4), recording instruments are still not uniform, and “instruments for measuring such things as temperature, humidity and wind have changed constantly.” The AP was reporting a new National Research Council statement that “deficiencies in the accuracy, quality and continuity of the records [of climate change] place serious limitations on the confidence that can be placed in the research results.” According to Kevin Trenberth of the National Center for Atmospheric Research (NCAR), “It’s very clear we do not have a climate observing system... This may be a shock to many people who assume that we do know adequately what’s going on with the climate, but we don’t.”

According to one expert’s opinion, the reason we know so little is a lack of instrumentation. Dr. David E. Wojick, writing in The Electricity Daily, tried to calculate how many sensors would need to be distributed world-wide in order to arrive at an accurate global mean temperature. Based on a surface-area estimate for the earth of 200 million square miles, Wojick estimated that it would take “at least 8 billion temperature sensors to maybe get a one degree accuracy, especially since temperature can vary 50 degrees in one day, and 100 degrees in a year, in some places.” In contrast, the NOAA and NASA figures are based on approximately 7,000 surface sensors.

return to Howard Fienberg's page