Detection and Attribution of Climate Change Using Offline Model Simulations with Applications to Runoff in the United States and Streamflow in the Columbia River Basin

Detection and attribution analysis of climate change is the processes of statistically detecting a change in a particular climate variable or variable affected by climate and then confidently attributing the change to effects from external forcings such as greenhouse gases, aerosols, and solar-volcanic. The variables studied here are annual and seasonal runoff in the contiguous United States and streamflow in the Columbia River Basin for the period 1950 – 2010 and 1950 – 2008, respectively. For forcings, the effects of climate change and variability, CO₂ [carbon dioxide] concentration, nitrogen deposition, and land use and land cover change are used in both studies. Monthly observations of runoff were provided by WaterWatch from the United States Geological Survey, and an ensemble of semi-factorial land surface model simulations were used to quantify the effects due to external forcings. The two limitations of the study conducted on runoff in the United States were: the inclusion of human regulation and irrigation withdrawals within the observations and not in the model simulations and a dry bias within the model simulations due to the precipitation driver. These limitations were overcome in the streamflow study for the Columbia River Basin due to the availability of a naturalized streamflow dataset and a new ensemble of semi-factorial land surface model simulations which were driven by less biased precipitation.United States runoff had significant and insignificant increases in the east, north, and south, and a strong significant decrease in the west. These changes were detected in the effects of climate change and variability but could not be attributed due to the dry bias in the precipitation driver leading to underestimation in the model simulations. However, for the Columbia River Basin, the changes in annual total, center of timing of, and summer mean streamflow were attributed to climate change and variability. The most significant changes were the declines in the June – October months. On average, these months account for approximately 49% of the annual total flow. More specifically, the greatest decline was 28% for June which comprised approximately 22% [22 percent] of the total annual flow.