The Current Drought (1999-2003) in Historical Perspective

Julio L. Betancourt, Desert Laboratory, U.S. Geological Survey & University of Arizona, jlbetanc@usgs.gov

In the southwestern U.S., the past century has seen several large swings in climate. These swings are defined by dry spells during 1898-1904, 1946-1972, and wet periods during 1905-1924 and 1976-1998. Since 1999, the southwestern U.S., the southern and central Rockies and the western Great Plains have been gripped by persistent drought, particularly in 2002. The water year (October-September) of 2002 was the driest of the last century (45% of the normal from 1895-2002) in Arizona and second driest for the Southwest (AZ, NM, CO, UT). Still, the four-year average from 1999-2002 (77.8% of normal) did not exceed that for 1953-1956 (76.6%) or 1901-1904 (71.9%). Regardless of ranking, the ongoing drought has produced remarkable phenomena on the southwestern landscape, from a half-a-million-acre fire on the Mogollon Rim to more than a million acres of pinyon and ponderosa tree dieoffs in Arizona and New Mexico alone. Warmer and increasingly longer growing seasons cast troubling uncertainty about plant succession and movement in the wake of large-scale disturbances, synchronized by drought from New Mexico to Montana. How unusual is the current drought, what caused it, and how long can it continue?

Instrumental records of climate, spanning only the last 100 years, are inadequate for assessing the potential severity, persistence and frequency of drought, and for examining the sources of low-frequency climatic variability underlying short-term precipitation trends. As a proxy for instrumental records, ring widths from climatically sensitive trees provide a robust method for evaluating the long-term context for drought during the last millennium. Tree rings yield continuous, annually-resolved climatic reconstructions that are usually highly replicable. With proper sampling and analysis, tree rings can capture both high- and low-frequency signals of climate variability at local to subcontinental scales. In the long-term context provided by the rapidly-growing network of tree-ring chronologies, there are a few, perhaps relevant observations.

1. Drought frequency and magnitude varies considerably over time and space. During the 20th century, New Mexico experienced more frequent and severe droughts than Arizona. The reason may be that Arizona receives precipitation from a greater mix of storm types than New Mexico, and Arizona climate is therefore less susceptible to multiseasonal drought. Another explanation is that the Arizona-New Mexico contrast is peculiar to the 20th century, and Arizona perhaps has been drier than New Mexico at other times in the past. Tree rings suggest that the 20th century may or may not be representative of extreme climatic variability, depending on the region. For example, in southern New Mexico, 1950-1956 was the third driest period, while 1976-1998 represents the third wettest spell, in the last millennium. In the upper Colorado River Basin, none of the dry spells in the 20th century approach the severity or duration of extreme droughts in the tree-ring record, and the last 100 years stand out as the wettest century of the millennium. Hence, in New Mexico the 1950s drought is an appropriate, worst-case scenario for water and resource managers; in the upper Colorado River Basin, it is not.

2. There is conspicuous decadal (20-30 yrs) to multidecadal-scale (30-70 yrs) variability in the tree-ring record indicating large-scale persistence of climatic anomalies. Over a few centuries in a few tree-ring records (primarily in New Mexico), this variability appears cyclical. There is some optimism in the climate community that, once understood, cyclic behavior in the climate of western North America (and ocean temperatures as the likely source) could help extend severe drought prediction beyond seasonal forecasts. A more complete assessment of the regional network for tree-ring reconstructed precipitation, however, indicates that these low-frequency oscillations are not cyclic; instead they are unstable in time and space. This apparent nonstationarity (change in means and moments of the distribution through time) in the tree-ring record means that extended droughts (and extended wet periods) are generally unpredictable. Regional phasing of the most extreme droughts across large areas of North America, however, strongly implicate abrupt shifts and persistent anomalies in ocean climate.

3. A barrage of recent studies on decadal-scale climate variability linked to slow ocean processes holds some promise that the forecasting window for drought can be widened from a few months to a few years. This is what all the commotion has been about the so-called Pacific Decadal Oscillation or PDO, the fact that Pacific sea surface temperatures (SST) and associated climate tend to vary on a 20-30 year 'cycle' with abrupt regime shifts. For example, in 1946 the tropical Pacific went cold and the Southwestern U.S. went dry, producing the 1950s drought. And in 1976, the tropical Pacific got warm and the Southwest experienced a prolonged series of wet winters. More recent studies indicate that the relationship between ENSO, PDO and warm season (May-September) rainfall across North America is modulated by low-frequency variability in North Atlantic SST's (Atlantic Multidecadal Index or AMO). The North Atlantic was anomalously warm between 1930 and 1960, encompassing summer drought in both the 1930s and 1950s. Could we have predicted the 1950's drought, involving dry winters and summers, or the moist period after 1976, defined by wet winters and summers, and all of their consequences for the Southwest? In other words were 1946 and 1976 diagnostic of years to follow? Are there precursor states in oceanic climate, similar in fact to the way we now use Tropical Atmosphere Ocean (TAO) moorings in the tropical Pacific to predict and monitor El Niño or La Niña, which might allow us to anticipate catastrophic drought?

4. Beginning in 1995, the North Atlantic turned warm and in 1998, the tropical Pacific became cold, prompting many climatologists across the country to unofficially speculate that catastrophic and perhaps multi-year to multidecadal drought was right around the bend. To many, the dryness in 1999-2002 was no surprise. We can't say how severe or how many years this drought will last, but based on the slow turning of the ocean, the odds are that it will. It is always prudent to be conservative about water and other resources vulnerable to drought; it is particularly prudent now. Finally, there are some added complications. Societal demands on surface water are on the increase. Leaf area and tree densities are higher than ever, while the growing season has increased by as much as two weeks in the spring, raising both heat load and evapotranspiration. Clearly, the context for long-term drought is different now than it has been in the past.