Paleo-Dendrochronological "Tree-Ring" Hyrdoclimatic Reconstructions for Northern and Southern California River Basins

The purpose of this guidebook is to introduce water resource managers to extended records of streamflow and precipitation developed using tree-ring data, and to demonstrate how these data provide insights on drought risk. While streamflow and precipitation gage records show climate and hydrologic variability over the 20th and 21st centuries, the reconstructions document variability over a much longer period of time, from hundreds to thousands of years into the past.

This is a draft final report for “Klamath/San Joaquin/Sacramento Hydroclimatic Reconstructions”, which includes developing new and updated tree-ring chronologies, reconstructing specified streamflow and/or precipitation records in the Klamath, San Joaquin and Sacramento River Basins, and analyzing the time series properties of the reconstructions. These reconstructions allow assessment of hydrologic variability over centuries to millennia and give context for assessing recent drought events. Tasks on this project included field collections; laboratory work in sample preparation, dating and measurement; and statistical work in chronology development, reconstruction, and analysis of those reconstructions. The interpretation of the reconstructions includes five main components : 1) place instrumental-period flow statistics in a long-term context, 2) quantify droughts and wet periods, 3) identify cycles in wetness and dryness, 4) check consistency with other paleoclimatic data, and 5) assess reconstructed flow variations in the context of expected scenarios of climate. A total of 16 different hydroclimatic series – 11 flow records and 5 precipitation records, were reconstructed. This report summarizes the research project products and results. The Data and Methods section of this report describes the field collections and development of tree-ring chronologies, reconstruction modeling, and methods of analyzing reconstructions. The Results section focuses on the interpretation of selected key reconstructions for the Klamath, Sacramento, and San Joaquin basins. Annual time series of the 16 reconstructions are included as products in an appendix and in a digital spreadsheet.

The goal of this project was to develop tree-ring based reconstructions of streamflow and precipitation for southern California. These reconstructions, along with existing reconstructions for northern and central California and an updated reconstruction of the Colorado River, provide information about statewide and regional drought for the past millennium. For this project, six southern California records were reconstructed: four for total water year precipitation (Ojai, Lake Arrowhead, San Gabriel Dam, and Cuyamaca) and two for water year streamflow (Arroyo Seco and Santa Ana River), along with a reconstruction of Kern River streamflow in the southern Sierra Nevada. In addition, a reconstruction for Colorado River flow at Lees Ferry was updated. Two versions of the eight reconstructions were developed. One set emphasizes the best match with the observed records, called the most skillful set. This set of the reconstructions extends back to the early 1400s. The second set of reconstructions emphasizes length, and starts in the 1100s. The most skillful California reconstructions end in 2016, while the most skillful Colorado River reconstruction ends in 2015. The updated Colorado River reconstruction, most skillful version, is very similar to the reconstructions developed and published in 2006 and 2007 by the authors. The reconstructions provide a way to evaluate the characteristics of instrumental period streamflow and precipitation in the context of past centuries. Metrics such as average flow and variance for different intervals of time were used make this evaluation. One of the most notable findings in this assessment is that, over the 20th and 21st centuries, southern California precipitation and streamflow, along with Kern River streamflow, have been more variable with correspondingly less year-to-year persistence, than over the past six centuries. In contrast, average flow and precipitation are not markedly different over the instrumental period, compared to the past six-centuries, although there is a tendency for the instrumental period to be wetter than average compared to similar length periods in the past. The longer records also allow an assessment of the cyclic behavior that may not be detectable in the shorter instrumental records. A cycle of moisture variability at a periodicity of about 13 years is present in recent decades in both southern California and the Upper Colorado River basin, but is not a consistent feature over the multi-century time frame of the reconstructions. A cycle with an average period of about 23 years is present in the most skill streamflow reconstructions. In the longest reconstructions, a cycle slightly longer than 100 years is evident in the southern California and the Kern River but not in the Colorado River.