Drought evaluation using tree-ring based reconstructed streamflows for rivers in New Mexico
The aim of this research is to extend the current hydrologic record by using tree-ring width data and gaging station information. Included in this study are the river basins of the Gila, Rio Grande, and Pecos Rivers, located in New Mexico that extend to the surrounding states of Arizona and Texas. Reconstructions using tree-ring data, were generated to estimate the streamflow for the past 1,875 years for the Gila River, 697 years for the Pecos River and 2,129 years for the Rio Grande. These estimates can give a larger picture of the streamflow variability for the region and allow for more suitable water management planning. To reconstruct the mean summer (May-October) flow for each river basin, 35 tree-ring chronologies, 22 Douglas Fir (Pseudotsuga menzeisii) and 13 Pinon Pine (Pinus edulis) were used at 12 gaging stations. Three gaging stations are located in Arizona (Kelvin, Clifton, and Calva) and the remaining nine in New Mexico (Virden, Gila, Glenwood, Cerro, Embudo, Taos, Albuquerque, Chico, and Santa Rosa). The longest record of gaging station data for the three river basins is as follows: Pecos 79 years, Rio Grande 82 years, and Gila 97 years. These instrument records show nothing of the long-term variability when compared to the length of estimated streamflows using tree-ring data. Reconstructions of the three rivers produce the following years of data, Rio Grande River 2,129 years, 697 years for the Pecos River, and 1,875 years for the Gila River. To reconstruct the streamflows for the three rivers, 33 pre-existing tree-ring sites and two newly collected sites were used. Based on previous studies, each of the newly collected sites could be used on any of the three rivers given that the river basins fall within 300 miles of the sampling sites. Multiple-linear regression with a +2 year lag was used to calibrate, verify and reconstruct the dependent variable of streamflow with the tree-ring width indices. The results show that this correlation analysis for the area prove tree-ring chronologies can be used to reconstruct past flows for the Rio Grande, Gila, and Pecos Rivers. The results for these 12 sites produces the following R2 values for the final reconstruction models, Kelvin R 2= .347, Clifton R2= .196, Virden R2= .442, Glenwood R2=.438, Gila R2=.391, Calva R2= .382, Santa Rosa R2 =.995 Chico R 2=.565, Albuquerque R2= .830, Taos R2= .635, Cerro R2=.724, and Embudo R2 =.725. The reconstructed streamflows can be used to study and evaluate drought patterns and help plan for future events. Possible applications of the reconstructions include analyzing possible periods of drought by evaluating flows that fall below the mean flow of the reconstruction for more than five consecutive years and ranking the periods. When looking at the three river basin reconstructions, the Gila River shows signs of having three droughts during the 1900's along with one in the 1800's followed by two in 1700's, three in 1600's and one in the 1500's when the top ten lowest flow periods are ranked. In ranking the top ten lowest flows for the Pecos River, the drought of 1950's shows up among the two droughts of the 20th century followed by two additional droughts for both the 1800's and 1700's, the remaining four lowest flows each occur in the 1600's, 1500's, 1400's, and 1300's. Finally the Rio Grande River reconstruction has one drought in each of the following centuries 20th, 19th and 13th and seven droughts after the 1200's. To put the drought analysis in perspective and understand what is currently happening to the rivers studied, it is important to compare the droughts not recorded to droughts occurring while streamflow data has been collected. While there are, intuitively, more droughts occurring before flows were records the droughts of the 20th and 19th century have been more severe. For the Gila River the drought analysis reveals that the 1950's drought ranked as the number one drought in terms of lowest streamflow. The Pecos River shows little difference ranking in the drought periods. The two driest periods occurring during instrumental record keeping are ranked as the 6th, the drought of the 1930's, and 8th being the most recent drought of the 1950's. While overall, the worse drought periods are occurring during current periods they are shorter in duration than "pregaging station" droughts. Further analysis should be conducted to analysis the patterns of drought and determine if current rivers are in for short periods of severe drought.
Foster, Nichole Michelle, "Drought evaluation using tree-ring based reconstructed streamflows for rivers in New Mexico" (2008). ETD Collection for University of Texas, El Paso. AAI1453860.