Understanding recent global hydroclimate change using multivariate detection & attribution techniques and GCM Experiments

Lead PI: Dr. Yochanan Kushnir, Katherine Marvel , Dr. Michela Biasutti

Unit Affiliation: Ocean and Climate Physics, Lamont-Doherty Earth Observatory (LDEO)

August 2015 - July 2019
Global ; New York City, NY ; New York
Project Type: Research

DESCRIPTION: The observed warming of the global average surface air temperature displays intervals of rapid increase alternating with intervals with a slow rate of change. Moreover, examining the regional surface air temperature trend reveal regions of the glob that cooled for decadal and longer time intervals as the global averaged temperature increase (for example, the cooling of the eastern tropical Pacific between 1998 and 2013 and the slightly negative trend in the North Atlantic subtropical gyre during most of the 20th century). Understanding the cause of such changes and their implication for the future, poses a challenge for climate researchers who use global climate models for future projection. These models fail to reproduce the observed leveling of the global average temperature (T) trend given the forcing prescribed in the historical CMIP simulations.

In order to enhance our confidence in these models it is imperative to differentiate between the contributions of external radiative forcing and internal climate variability. However, recent climate change is not fully characterized by a single variable: the slowdown in global temperature change has been accompanied by a decrease in global average precipitation and distinct regional precipitation patterns. The global precipitation (P) response to a temperature change is understood to depend on the particular external forcing that is dominant. Patterns of natural decadal variability of P and T are also distinct from those of radiatively forced change.

Thus, a multivariate analysis will be able to determine the contributions of different radiative forcing and natural variability to climate change in the late 20th and early 21st centuries. The goal of this project is to use advanced methods of detection and attribution (D&A) in order to distinguish between the influence of external forcing and internal climate interactions in recent global climate change. In particular, in the relative role of internal and external forces in the observed temperature change. The project also seeks to find out if climate models are able to reproduce the change signals detected in observations and if they can be used to investigate the underlying processes.

OUTCOMES: We have conduction the planned experiments. This involved moving our model experiments from the NCAR Yellowstone facility to the new Cheyenne facility. Preliminary results from the early experiments were incorporated in the student's doctoral thesis (published in the university thesis collection) and are now being incorporated into a peer-reviewed article currently in review.

• The characteristic fingerprints of external forcing in the seasonality of both precipitation and cloud cover were identified;
• A significant model-observation mismatch in tropical precipitation seasonality was presented. While all CMIP5 models predict a significant increase in the amplitude of the precipitation annual cycle (i.e., the range between wet and dry seasons), the observations indicate that large reductions have occurred over the satellite era. We linked this discrepancy to the early 21st century cooling of the eastern tropical Pacific which was not captured by the models;
• In collaboration with LLNL researchers we explore the effects of external forcing on global hydroclimate. We showed that the global precipitation response is not linear in multiple anthropogenic forcings due to interactions between different forcings not captured in standard climate models and developed a new framework to identify regions likely to experience future rainfall anomalies that are without precedent in the current climate (Bonfils et al 2015). Additionally, this research showed that, contrary to previous claims, observed changes to the amplitude of the precipitation annual cycle are inconsistent with model estimates of internal variability but not attributable to the model-predicted response to external forcing (Marvel et al 2017).
• We established that model failure to capture CO2-induced plant physiological effects may over estimate future drying (Bonfils et al 2017) and showed that a anthropogenic influence is already visible in tree-ring-derived records of early 20th century drought conditions (Marvel et al 2019).
• We showed that the seasonal cycle amplitude of total cloud fraction is indeed increasing in the tropics;
• We developed a new “emergent constraint” for climate sensitivity based on the width of the climatological tropical cloud belt.
• We investigated the physical processes responsible for the ""North Atlantic Warming Hole"" - an area at the center of the subpolar gyre where there ocean surface cooled slightly during the 20th century warming trend found in most ocean areas; We find that in the NCAR Community Earth System Model (CESM) future climate projections this region continues to lag in warming with respect to other oceans regions and that this is a result of the inflow of sea ice- melt from the Arctic (Gervais et al., 2018).


Department of Energy




Celine Bonfils, Karl Taylor


Center for Climate Systems Research (CCSR); NASA Goddard Institute for Space Studies


Livermore National Laboratory




Bonfils, C. J., Santer, B. D., Phillips, T. J., Marvel, K., Leung, L. R., Doutriaux, C., & Capotondi, A. (2015). Relative contributions of mean-state shifts and ENSO-driven variability to precipitation changes in a warming climate. Journal of Climate, 28(24), 9997-10013.,

Marvel, K., Schmidt, G. A., Shindell, D., Bonfils, C., LeGrande, A. N., Nazarenko, L., & Tsigaridis, K. (2015). Do responses to different anthropogenic forcings add linearly in climate models?. Environmental Research Letters, 10(10), 104010.,

Marvel, K., Biasutti, M., Bonfils, C., Taylor, K. E., Kushnir, Y., & Cook, B. I. (2017). Observed and projected changes to the precipitation annual cycle. Journal of Climate, 30(13), 4983-4995.,

Bonfils, C., Anderson, G., Santer, B. D., Phillips, T. J., Taylor, K. E., Cuntz, M., ... & Durack, P. J. (2017). Competing influences of anthropogenic warming, ENSO, and plant physiology on future terrestrial aridity. Journal of climate, 30(17), 6883-6904.,

Marvel, K., Cook, B. I., Bonfils, C. J., Durack, P. J., Smerdon, J. E., & Williams, A. P. (2019). Twentieth-century hydroclimate changes consistent with human influence. Nature, 569(7754), 59.


aerosols volcanic eruption climate fluctuation surface air temperature hydroclimate change detection techniques modeling climate change hydroclimate greenhouse gas climate models precipitation


Modeling and Adapting to Future Climate