Climate Change Monitoring and Detection
NOAA/CPO – Detection and Attribution of Climate Change Using Climate Indices for the United States
Funding Type: CIMMS Task III (Program Manager – Chris Miller, Climate Change Data and Detection)
Objectives
Evaluate US climate extremes indices from observational data and climate
model simulations; document the observed changes in climate extremes
in the U.S. over the 20th century; and attribute the observed changes
to specific climate forcings, where possible.
Accomplishments
Recent studies have used climate indices to detect changes in the climate
on continental and regional scales. This study utilizes a modified
version of the Climate Extremes Index (CEI) (Karl et al. 1996), which
is used by the National Climatic Data Center to operationally monitor
the tails of the distributions of various climate parameters within
the contiguous United States. Some issues with the operational calculation
of the CEI at NCDC were identified. Modifications to the operational
algorithm have been implemented at NCDC (Gleason et al. 2007) and the
new version of the CEI is available on the NCDC web site. The components
of this index use frequency-based statistics at each location and estimate
the fraction of the contiguous United States experiencing extreme values
during a given period. The components of the CEI are related to the
variations in extremes of mean maximum temperatures, mean minimum temperatures,
drought, heavy daily rainfall and the number of rain days. Extreme
is used to represent either much above normal (above the ninetieth
percentile) or much below normal (below the tenth percentile).
An assessment of the modified version of the CEI (mCEI) has been undertaken over the twentieth century using observations from the United States Historical Climatology Network (USHCN). The assessment encompasses annual, seasonal, and regional trends of the modified CEI as well as an evaluation of using daily temperature extremes over the period from 1910-2005. The observed variations have been compared with variations of the index using data from global climate model simulations for the twentieth century that include increasing concentrations of greenhouse gases and the effects of aerosols. Preindustrial control runs of the global climate models are also used to estimate the natural variability of the index with no greenhouse gas forcing to determine if observed trends are outside the range of natural climate variability. Model data from twentieth century experiment simulations with the NCAR PCM and CCSM3.0, GFDL CM2.0 and CM2.1, and CCCMA CGCM3.1 models, as well as the available preindustrial control runs for these models have been used.
Observed trends over the last thirty to fifty years in the temperature components of the mCEI for the whole United States, the western, central, and eastern U.S. are found to be statistically significant. Observed increases in the temperature components in the spring and winter are also found to be statistically significant. Daily warm temperature extremes are also shown to have large increases in areal coverage over the last thirty years, and the increase is unprecedented in the observed record. Model simulations for the twentieth century agree relatively well with these increases, and the increase is consistent with the expected response to anthropogenic increases in the concentrations of greenhouse gases. Therefore, the increased areal coverage of warm temperature extremes in the latter half of the twentieth century can be attributed to the anthropogenic increases in greenhouse gases.
Many of the other components of the mCEI do not show significant trends, with the exception of increases of the heavy precipitation component being marginally significant for the entire United States and the eastern portion of the United States. However, overall increases in many of the components, combined with the statistically significant increases in temperature extremes, yields significant increases in the modified CEI as a whole. Interpretation of the increases of the combined mCEI must be done with care by looking at the individual components before reaching a conclusion of the overall change in the observed climate.
Bryan Burkholder completed his MS thesis on this project in May 2007 and a paper arising from his research is being prepared for submission to the journal Science.
This project has been completed.
Publications
Gleason, K L., J. H. Lawrimore, D. H. Levinson, T. R. Karl, and D. J. Karoly,
2007: A revised U.S. climate extremes Index. J Climate, submitted.
Karoly, D. J., B. A. Burkholder, D. R. Easterling, K. L. Gleason and J. H. Lawrimore, 2007: Detection of a human influence on changes in climate extremes in the United States. To be submitted to Science.
Burkholder, B. A., and D. J. Karoly, 2007: Assessment of US climate variations using the US Climate Extremes Index. 19th Conf. on Climate Variability and Change, San Antonio TX, Amer. Meteor. Soc.
Gleason, K.L., J.H. Lawrimore, D. Levinson, and T.R. Karl, 2006: A revised US Climate Extremes Index. 18th Conference on Climate Variability and Change, Amer. Meteor. Soc., Atlanta, GA.
Karoly, D., A. Ruppert, D. Easterling, K. Gleason, and J. Lawrimore, 2006: Assessment of US climate variations using the US Climate Extremes Index. 18th Conference on Climate Variability and Change, Amer. Meteor. Soc., Atlanta, GA.
Index (mCEI) from observations for the US from NCDC (bold, solid), from climate model simulations for the twentieth century (light dash and dot patterns), and the 95% confidence interval for decadal variations of the mCEI (bold, dashed) based on long unforced model control runs. There is a significant increasing trend in the observed mCEI over the 20th century that is outside the range of natural variability and consistent with the model simulations that include changes in anthropogenic climate forcing factors