Date of Award


Degree Type


Degree Name

Doctor of Philosophy


Civil Engineering

Major Professor

Dr. Glenn Tootle

Committee Members

Dr. Henri Grissino-Mayer, Dr. Joshua Fu, Dr. John Schwartz


The Colorado River provides water to over 25 million people. Given the importance of this water supply, it is critical to understand the hydrologic variables in the Colorado River Basin. In this dissertation, I reconstructed hydrologic conditions (soil moisture, snowpack) in the Upper Colorado River Basin (UCRB) and examined different factors that influence water supply in the region (climate oscillations, oceanic-atmospheric variability).

Firstly, I reconstructed soil moisture in the UCRB. Principal components analysis (PCA) and k-Nearest Neighbor (k-NN) techniques were used to regionalize the gridded data. Correlated tree-ring chronologies (TRCs) were used as predictor variables in stepwise linear regression (SLR) to determine the optimal regression model with the highest skill. Soil moisture was successfully reconstructed for the various regions with R2 values ranging from 0.42 to 0.78. Reconstructions that used TRCs based on ponderosa pines or pinyon pines were more statistically skillful. Secondly, I reconstructed snowpack in the Upper Green River Basin (UGRB). I used PCA to regionalize the data. I used TRCs and climate signals (Southern Oscillation Index (SOI), Pacific Decadal Oscillation (PDO)) as predictor variables in SLR. The snowpack reconstruction using only TRCs had an R2 value of 0.21, while the reconstruction that included the SOI had an R2 value of 0.58, demonstrating the value of including climate oscillations in regional reconstruction efforts. Lastly, I reconstructed snowpack in the UGRB, including a regionally specific sea surface temperature (SST) index as a predictor variable. The SST index was identified using singular value decomposition (SVD) to determine a SST region that was teleconnected with UGRB snowpack. Using a teleconnected SST region increased reconstructive skill and resulted in an R2 value of 0.63.

The major contributions of this dissertation are the first soil moisture reconstruction in the United States, the first successful snowpack reconstruction in the UGRB, and conclusive evidence that using climate signals and regionally specific SST indices as predictor variables can augment reconstructive skill. Soil moisture data can be used to enhance understanding of paleoenvironments and help forecast future water availability. Reconstructed snowpack data for the UGRB will allow for a better understanding of hydrologic variability in the region.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."