HR: 1340h
AN: H43B-0501    [Abstracts]
TI: Potential Predictability of Precipitation Anomalies in Regions of Varying Anomaly Regimes
AU: * Ferguson, I
EM:
AF: University of California Berkeley, 537 Davis Hall, Berkeley, ca 94566 United States
AU: Duffy, p
EM:
AF: Lawrence Livermore National Laboratory, L-103, PO Box 808, Livermore, ca 94550 United States
AU: liang, x
EM:
AF: University of California Berkeley, 537 Davis Hall, Berkeley, ca 94566 United States
AB: Precipitation regimes vary widely by region, in terms of frequency and intensity, as well as seasonality and mean annual totals. Moreover, fundamental differences in anomaly regimes-deviations from the long-term precipitation climatology-may suggest that the primary mechanisms forcing precipitation anomalies differ by region, and therefore that the potential predictability of drought may vary between regions. For example, anomalies strongly influenced by boundary over initial conditions indicates higher potential predictability, and vice versa. Analysis of precipitation anomaly regimes is essential to improving our understanding of the potential predictability of precipitation anomalies, which will have substantial implications for both seasonal-to-interannual precipitation forecasting and its application to water resources management. In this study, we analyze spatial variability in observed and simulated precipitation anomalies. Focusing on the U.S. Great Plains (95° - 105°W, 30° - 50°N) and the Western U.S. (115° - 125°W, 30° - 50°N), we show that observed monthly precipitation anomalies have marked spatial variability. Significant differences in the distribution and autocorrelation of regional anomalies suggest characteristic differences in anomaly regimes between these regions. Moreover, spectral analysis shows substantially higher variability at low frequencies (greater than ~6 years) than at higher frequencies (greater than ~2 years) over the Great Plains, while analysis of anomalies in the West shows relatively constant variability across the frequency domain. Using the mean anomaly correlation coefficient and signal-to-noise ratio, we then compare the potential predictability of precipitation anomalies over the Great Plains and the Western U.S. Potential predictability is evaluated in two ensembles of atmospheric GCM simulations, a 14-member ensemble of 20th century simulations using the NSIPP-1 AGCM and a 20-year, 15-member multi-model ensemble of simulations from the Atmospheric Model Intercomparison Project. In addition to capturing the characteristic differences in anomaly regimes, both ensembles show significantly higher potential predictability over the Great Plains, suggesting a stronger coupling between boundary conditions and precipitation in this region than in the Western U.S.
DE: 1616 Climate variability (1635, 3305, 3309, 4215, 4513)
DE: 1655 Water cycles (1836)
DE: 1807 Climate impacts
DE: 1812 Drought
SC: Hydrology [H]
MN: Fall Meeting 2005