How Reliable is the Pacific Decadal Oscillation for Predicting Precipitation in Texas? Comparing the Precipitation Record for Two Stations in Texas between 1900 and 2013

Abstract:

The Pacific Decadal Oscillation Index (PDO), which has been correlated to weather patterns in North America, was compared to precipitation variations for two stations in Texas - Amarillo (mean annual precipitation= 509 mm) in the Texas Panhandle and San Antonio (mean= 742 mm) in south Texas. Negative values of the PDO have generally been correlated with drought periods in the southern United States, including Texas, and positive values of the PDO correlate with periods of above average precipitation. Using a smoothed time-averaged plot of the PDO index, the PDO record was divided into 7 time periods characterized by either positive or negative, as follows: 1900-1934 - neutral, 1935-1944 - positive, 1945-1977 - negative, 1978-2003 - positive, 2004-2007 - positive, and 2008 - 2013 - negative values. For each of these periods the mean annual precipitation was calculated and compared to the values of the PDO index. A correlation was considered positive if above average precipitation occurred during positive PDO values, or if below average precipitation occurred during a period when the PDO values were generally negative. For Amarillo, a positive correlation was noted for all periods except 1978-1997. For the San Antonio data, a positive correlation was noted for all periods except 1998-2003. Given the prevailing negative values of the PDO, 1998-2003 should have been a time of decreased rainfall, but instead the mean annual rainfall was more than 127 mm above the long-term mean. One possible explanation for the increased rainfall is that 1997-1998 was a time of the strongest recent El Nino and 2002-2003 was a time of moderate El Nino conditions - both of which tend to produce above average precipitation in Texas. Notable rainfall amounts in San Antonio during this period are the 459 mm recorded during the month of Oct. 1998 and 430 mm in July 2002. These results suggest that the PDO cannot be used , by itself, to predict precipitation trends in Texas.