Cooperative Institute for Research in Environmental Sciences

Abstracts: 10

NOAA-CIRES Joint Institute Research on Sub-Seasonal to Interannual Variability

P. Sardeshmukh, M. Alexander, J. Barsugli, G. Compo, T. Hamill, M. Hoerling, M. Newman, S. Peng, M. C. Penland, and J. Whitaker

A challenging research question concerns what aspects of climate variability within a season are potentially predictable and what useful information can be extracted from such predictions. Sub-seasonal variability is important not only because it accounts for a large fraction of the total atmospheric variability from synoptic to decadal scales, but also because of its association with winter and springtime floods, summertime droughts, likelihood of the occurrence of hurricanes, and other phenomena that have major societal consequences. Prediction problems within this time scale are particularly challenging because boundary conditions are becoming increasingly important, while the influence of initial conditions has not completely dissipated, and the chaos from unpredictable nonlinear interactions has nearly saturated. Joint institute research at CDC focuses on the variability and predictability of weekly averages using both modeling and diagnosis of the observed statistics out to 4-6 weeks in advance, and detailed investigations of NCEP's operational forecast ensembles for Week 2. A significant accomplishment has been to construct a low-dimensional linear empirical-dynamical model that successfully represents the statistics of weekly anomalies with forecast skill in Week 2 comparable to that of NCEP's operational ensemble model predictions, and superior to the operational model in Week 3. Much of the linear model's skill is due to processes that are not well represented in the NCEP or other numerical weather prediction models, such as intra-seasonal variations of tropical convection. On the other hand, much of the skill of the numerical models is likely due to processes that are not well represented in the empirical model, such as nonlinear baroclinic cyclogenesis and blocking development. Future research will explore how to systematically combine the empirical and numerical model forecasts to yield superior weekly to monthly forecasts.

On seasonal to interannual time scales, joint institute research conducted at CDC has demonstrated that several elements of the tropical Pacific ENSO phenomenon can be predicted two or more seasons in advance. Research with relatively simple models has increased understanding of fundamental ENSO dynamics, improving ENSO predictions and suggesting that at least some aspects of interannual variability might be predictable in other parts of the globe. Joint institute scientists have also provided important evidence that the predictable evolution of ENSO and of the associated remote teleconnections are governed largely by linear, low-dimensional dynamics, enabling simple empirical linear models to make seasonal predictions that are competitive with state-of-the-art GCMs. Research has identified situations in which the dynamics of ENSO are significantly nonlinear, clarified the extent to which the remote impacts can be nonlinear, and explored the extent to which those impacts might be sensitive to the details of the tropical SST anomaly patterns. Investigations have also explored how the nonlinear and higher-dimensional dynamics might affect the probability distributions of atmospheric variables on synoptic, subseasonal, and seasonal scales, and thus contributing to the development of quantitative risk estimates of extreme events associated with ENSO. Other research has assessed the predictability of SSTs in other oceans basins, both through "atmospheric bridge" teleconnections and through the year-long persistence, subsequent re-emergence of SST anomalies from seasonally-varying subsurface mixed layer, and the impact of such SSTs on the atmospheric circulation.