Wednesday, 15 May 2002: 11:45 AM
Associations between the Palmer Drought Index and various atmospheric and oceanic indices
In order to improve seasonal predictions of United States drought conditions, an extensive variety of oceanic and atmospheric indices were studied to determine if useful correlations could be derived. The monthly Palmer Drought Indices (PDI) were used to represent drought conditions, and the correlations were determined for nine U.S. regions as well as selected states and climate divisions. Indices examined included the various measures of ENSO such as SSTs for Nino regions 3, 3.4, 4, and 1+2, and the Multivariate ENSO Index (MEI). SSTs in the North Pacific and several other parts of the world were also used. Atmospheric indices included the NAO, AO, and SOI. In the stratosphere, the Quasi-Bienniel Oscillation (QBO) at both 50 and 30 hpa, as well as the difference between both levels, was included. The 10.7 cm solar flux and sunspot numbers were used to represent solar activity. Indices were averaged for various time periods and correlations were calculated for a variety of lead times. Most correlations were based on the 1979-2001 period, as results for the most recent climate conditions were desired. One or more of the atmospheric/oceanic/solar indices exhibited statistically significant correlations with the PDI during most seasons of the year and for nearly all regions. Looking at all months of the year, it was found that various ENSO indicators had the highest correlations with drought across the South and Southeast, while the Trans-Nino Index (TNI) was important in the western regions. For the West North Central region, solar activity indicators showed the highest correlations. The solar indicators had the highest correlations with drought all seasons of the year in this region, and were especially high in the fall and winter. The ENSO indicators showed the highest correlations across the South and Southeast during winter and spring. In the summer, Pacific SSTs north of the equatorial region had the most significant correlations, with the PDO important in the South and Southwest, and SSTs in a region around Baja California important in the Central and Northeast states. In many cases, the correlations diminish considerably at several months lead times, making their use as forecast tools problematic. In other cases, indices with leads as much as 6 months show significant correlations with the PDI. It appears that knowledge of these relationships could result in improved drought forecasts in at least parts of the country during certain seasons.