HR: 1340h
AN: A53B-0880    [Abstracts]
TI: Stochastic modeling of daily summertime rainfall over the southwestern US and its relation to interannual variability
AU: * Wang, J
EM: wjy@bu.edu
AF: Geography Department, Boston University, 675 Commonwealth Ave, Boston, MA 02215
AU: Anderson, B
EM: brucea@bu.edu
AF: Geography Department, Boston University, 675 Commonwealth Ave, Boston, MA 02215
AU: Salvucci, G
EM: gdsalvuc@bu.edu
AF: Geography Department, Boston University, 675 Commonwealth Ave, Boston, MA 02215
AB: A large fraction of annual precipitation over the Southwestern US is concentrated in the summer season and attributable to the North American Monsoon System (NAMS). In addition, previous studies suggest strong relationships between the NAMS precipitation and rainfall over other regions, such as the Great Plains and the East Coast, suggesting regional rainfall over the southwest is also related to the larger-scale hydrologic cycle over North American continent. As such, the NAMS precipitation pattern has tremendous impact on the local hydrologic and ecological system in this region and may provide potential predictability for interannual variations in the seasonal rainfall during the summer. Here, the interannual variance in this summertime seasonal precipitation over 78 southwestern US stations is studied using Markov Chain models and empirical intensity distributions. Modeling results suggest that a 2-order Markov Chain can optimally portray the temporal structure of the summer daily precipitation process over the southwestern US. The 2-order Markov Chain model with stationary event frequency and intensity characteristics, in turn, can explain approximately 75 % of the interannual variance in the seasonal number of wet days and 85 % of the interannual variance in the total seasonal precipitation. In addition, only a small fraction of anomalous years at any given station (generally smaller than 20 %) show significant changes in either of these characteristics. However, relatively higher fractions of the anomalous years are observed in regions north of 37 N suggesting possible spatial heterogeneity in the frequency of non-stationary behavior of rainfall over the domain. In general, at a given station the anomalous years are related to both anomalous numbers of seasonal total wet days and anomalous light/heavy rain intensity distributions. Studies investigating variance explained by non-stationary variations in the occurrence and intensity characteristics indicate they display similar significance in capturing the remaining 15 % of interannual variance of seasonal total precipitation. However, numerical tests suggest that these two low frequency variations are not independent variables for the NAMS precipitation over the southwestern US. Complex covariance that cannot be described with stochastic statistical models may exist between those two variations. Further study with dynamic models and cluster analyses is needed to examine the low frequency variation processes and their influence on potentially predictable precipitation in this region.
DE: 1836 Hydrologic budget (1655)
SC: Atmospheric Sciences [A]
MN: 2004 AGU Fall Meeting