Jason Paul Giovannettone, PhD, PE
As many locations throughout the United States have recently experienced periods of extreme wet and dry conditions, an attempt is made to better understand the relationships between long-term precipitation and climate variability. The correlations between extended periods of total precipitation and low-frequency oscillations of global hydro-climate indices are analyzed using a method referred to as “long-window” correlations analysis. Long-term precipitation data (up to five years) for over 1,000 sites throughout the United States are analyzed and correlated to low frequencies of several hydro-climate indices using lead times ranging from 12 to 48 months. The strength and significance of each relationship are assessed using the Pearson’s correlation coefficient and a Monte-Carlo approach, respectively. Strong correlations (Pearson coefficients > 0.80) that exhibit high significance were found at several locations; in particular, low-frequency oscillations of the MJO and ENSO revealed strong links to annual and longer-term precipitation within multiple regions. A few individual sites are analyzed in further detail by constructing future predictions of total precipitation and comparing these to historic observations; average errors below 5% were found in some cases. The final results of this study allow a greater understanding of the climate mechanisms responsible for long-term variability in precipitation throughout the United States, leading to improved predictions of the onset and persistence of future droughts and flooding events at longer lead times.
The asset you are trying to access is locked. Please enter your access key to unlock.