Use of tree ring indices for reconstructing streamflows with applications on drought analysis

Dr. Jose D. Salas, Ph.D. Professor, Department of Civil and Environmental Engineering
Colorado State University, Fort Collins, Colorado

The planning and management studies of water resources systems generally take into account the temporal and spatial variability of hydrological processes such as precipitation and streamflows, particularly periods of extreme floods and droughts. One may obtain some relevant information about these processes from the historical records that may be available in the river basin of interest. However, the information that can be obtained from the historical (systematic) records is quite limited because the records are generally short. Consequently the estimates of drought properties that can be obtained from the historical records have a large degree of uncertainty and sometimes it is not even possible to estimate them. An alternative is using paleo-hydrologic data (proxy data) such as tree ring indices obtained from trees located in the basin of interest for reconstructing the hydrologic records and extending the database. A multivariate model (MREXTN) that includes an autoregressive vector, a vector of explanatory variables (tree ring indices,) and a random (noise) vector, was developed, which enables reproducing the variances, autocorrelations, and cross-correlations of the underlying variables (e.g. streamflows.) The vector noise term has the disadvantage of yielding multiple solutions, i.e. multiple reconstructions and extensions. Thus a procedure was developed for selecting a single reconstructed flow sequence at each site from the multiple sequences. The proposed model was tested against competing models by using computer simulation experiments. It is concluded that the MREXTN model compares favorably versus alternative existing models. We will also show results obtained for reconstructing streamflows at 29 sites of the Colorado River system and their applications for estimating critical droughts and the return period of droughts in the Colorado River system.