Colorado River Response to Climate Change since the Late 19th Century

Marty Hoerling

NOAA/ESRL Physical Sciences Laboratory

Wednesday, Aug 22, 2018, 2:00 pm
DSRC Room 1D403

Abstract

Annual naturalized Lees Ferry streamflow, a measure of water availability in the Upper Colorado River Basin (UCRB),is estimated to have declined by 19% during 1896-2016, a period of substantial temperature rise. This talk examines causes for the declining flow, and will begin by addressing the paradox in current scientific understanding on this problem. On the one hand is the long-held view that Colorado Basin hydroclimate is highly variable, and that multi-decadal low flow epochs --- having greater severity than the post-2000 drought --- existed on the river. On the other hand is an emerging view of a Colorado River that is highly sensitive to temperature change. To resolve this paradox, results on the UCRB hydroclimate response to climate change forcing are presented using a new suite of large ensemble high resolution (~50km) global coupled land-atmosphere models (AGCMs. The experimental design employs an attribution method of conducting parallel AGCM simulations, one subjected to actual current boundary forcings and the other (so-called counterfactual runs) in which long-term global warming effects have been removed from boundary forcings. The simulations affirm that much of UCRB warming since the late 19th century has resulted from the basin’s response to climate change. A new finding is of a forced precipitation decline ranging from -2% to -5% among the three different AGCMs studied herein. A substantial portion of anthropogenically-forced flow declines since the late 19th Century is shown to result from this signal of precipitation decline, rather than from the temperature rise alone. This owes to a property of the semi-arid UCRB basin in which precipitation change has a magnified effect on streamflow change consistent with an elasticity of at least 2. In contrast, our AGCM experiments indicate an only modest temperature-related flow sensitivity of about 3%/°C, which is on the low end of empirical and land surface model estimates. In sum,the results affirm a climate change driven reduction in UCRB flow volumes over the past century that aligns with emergent views. Our results highlighting a key role for precipitation are, however, contrary to the notion that warming has been the cause for dwindling Lees Ferry flow. These different interpretations on the physical causes of climate change impacts are shown to have important implications for the debate on detectability of Lees Ferry flow change to date, and for predictions of future Colorado River water availability.