Recent decades saw wetter continents.

Recent decades saw wetter continents

Measurements from weather stations show that Earth's land areas have gotten wetter over the last few decades, say atmospheric scientists who have compiled a century's worth of global data. But they say it is not yet clear whether natural climate fluctuations or the predicted "greenhouse" warming have precipitated this change for the wetter.

Data from 2,201 stations for the period 1890-1986 indicate that the mean annual precipitation falling on land areas in the Southern Hemisphere has increased since the 1940s, while the Northern Hemisphere has seen no significant change, report Henry F. Diaz from the Environmental Research Laboratory of the National Oceanic and Atmospheric Administration (NOAA) in Boulder, Colo., and his colleagues in the Jan. 20 JOURNAL OF GEOPHYSICAL RESEARCH. South of the equator, precipitation has increased mostly within the last 15 years.

When the researchers reported the Northern Hemisphere data two years ago, they had not yet compiled precipitation records for the Southern Hemisphere. Now, for the first time, those data enable scientists to estimate a worldwide total and make comparisons. The researchers used data from 1,545 northern and 656 southern weather stations, all of which have been recording since at least the mid 1920s. Observers have only recently begun recording precipitation over Antarctica, and the report includes no information from this continent.

Although the overall trend is toward more precipitation, many areas have become drier in recent years or experienced no overall change over the last century. South American stations, and to a lesser extent South Asian and Australian stations, report precipitation increases, but records from stations in southern Africa show no rise. Also, while the midlatitude regions on both sides of the equator have experienced more precipitation, the northern tropics have dried while the southern tropics have become wetter. The researchers say drought in Africa and the Caribbean accounts for recent dryness in the northern tropics.

Diaz and his colleagues say the changes in precipitation loosely match what is expected to occur during a greenhouse warming -- a general heating of Earth's climate due to increasing atmospheric concentrations of certain gases. But, he cautions, "at this point I would certainly not want to say that we are seeing any sort of greenhouse signal."

Part of the problem with identifying such a signal, he says, is that precipitation fluctuates radically from year to year and from decade to decade. During the years 1899-1920, for instance, precipitation levels over South America were much lower than the mean annual amounts during a reference period of 1921-1960. However, in the years since 1970 the same continent has experienced precipitation levels above the mean measurements for the reference period.

The researchers also say the record clearly shows large swings in precipitation caused by strong El Nino-Southern Oscillations. These seesaw shifts in the warm ocean water and high atmospheric pressures alter precipitation patterns across the Pacific and other areas.

If the expected climate warming has started to affect the levels of precipitation, this change would be superimposed on the large natural swings, making it difficult to detect the "fingerprint" of greenhouse warming in these historical records. "We have to be cautious when we try to ascribe a particular mechanism to the data," Diaz says.

The researchers also admit to other problems with the long-term precipitation records. Because certain portions of the continents have few or no weather stations, the researchers had to average the available measurements over large tracts of land. In many cases, a single station must cover hundreds or thousands of square kilometers. If such a station has measured a drop in precipitation, it is difficult to tell, whether the entire area surrounding the station suffered a drought at that time or whether the rain simply shifted slightly, leaving the station drier than nearby areas.

Moreover, the researchers say changes in instruments or procedures may have introduced large biases into the measurements.

Because of these and other caveats, some scientists are skeptical about using the historical precipitation records to test for climate change signals. Says Chester Ropelewski at the NOAA Climate Analysis Center in Camp Springs, Md., "We may have to write off the historical records."

Yet the worldwide data for precipitation are among the most detailed of all weather records, and many experts say they will become quite important in the near future to those who work with climate models. Computer specialists are beginning to run detailed simulations of how greenhouse warming will slowly change the climate. According to computer modeler Michael Schlesinger at Oregon State University in Corvallis, matching these predictions against the historical records for precipitation and other variables will help determine how the climate is changing and whether greenhouse gases are at the root of the change.