Reconstructing streamflow in the Upper Rio Grande River Basin

The sustainability of water resources in Southwestern United States is challenged by increasing demands and changing water uses. The successful management of water demand growth can only be done through accurate water resource management planning. Hydrologic records are a source of fundamental data used in planning. This research uses tree-ring data coupled with stream gaging records to extend the hydrologic record and provide better planning information. The Upper Rio Grande River Basin has it\u27s headwaters in southwestern Colorado and extends south to Fort Quitman, Texas to its junction with the Rio Conchos at Presidio, Texas. There are a total of twenty-two counties located in the Basin area with six counties in Colorado, thirteen in New Mexico, and two in Texas. A small portion of the Basin is in Mexico. Streamflow records in the Upper Rio Grande River Basin are spatially limited and the existing records do not extend longer than 110 years within the region. This record is too short to fully represent the entire range of natural climate variability in the region. Using proxy records from tree rings, we can determine climate and streamflows over the past 730 years (longest reconstruction 1275-2005), giving a more comprehensive understanding of historical streamflow variability and allowing for more suitable water resource management plans. This is largely due to the governing factors linking tree growth, precipitation, evapotranspiration, and streamflow. Twenty-five Douglas Fir (Pseudotsuga menzeisii) and Piñon Pine (Pinus edulis) ring-width chronologies were used to reconstruct the mean summer flow of the upper Rio Grande River Basin at 6 gaging stations throughout New Mexico (Embudo, Pecos, Hondo, and Butte) and southern Colorado (Alamosa and Creede), for AD 1275-2000, as the longest chronology. The reconstructions incorporate 22 pre-existing tree-ring sites and newly collected data from 3 sites in the Upper Rio Grande River Basin, providing a more comprehensive representation of streamflow for water resource allocation. First, a multiple-linear regression model including annual lags of +2 years was used to investigate whether the tree-ring width indices of these species can be used as predictors for streamflow reconstruction and if so how strongly correlated tree-ring width is to streamflow. Results indicated that radial growth in low-elevation sites are positively influenced by an increase in late-summer streamflow due to a critical moisture source of surface recharge from streamflow. Correlations provide additional verification of the quality of the reconstructions and further indicate that the comprehensive series are related to conditions effecting streamflow and tree growth within the United States desert southwest. Based on the final modeled reconstructions, strong evidence suggests that the severe droughts other than ones already known (i.e. 1930s, 1950s, 1987-1989, and early 21st century) have occurred in the Basin over the period 1275-2005. The results suggest that the frequency of severe drought has been increasing over this period. As an example, the Embudo gaging station final model with the highest significant correlation (R2= 0.71) was ranked in order of lowest flow years. When the possible droughts are compared based on flows falling below the mean flow of the reconstruction for more than five consecutive years during the 730-year period, three of the lowest flows occur in the 20th century. Two of the ten worst occurred during the 1800s, one in the late 1700s, one in the mid 1600s, and one in the late 1500s, one in the late 1400s, and one in the late 1300s. If this trend should continue into the 21st century, the U.S. Southwest could be facing long periods of extended drought. (Abstract shortened by UMI.