The Advanced Study Program 2013 Seminar Series third seminar is presented by Matt Newman, CIRES/CDC University of Colorado and NOAA/ESRL/PSD.
Climate variability is often characterized by a notable separation between the dominant time scales of interacting processes. For example, rapid weather forcing of the slow ocean can be approximated as white noise forcing of a damped integrator, or univariate red noise. A similar approximation can be applied to the more general case of anomalies representing many evolving regional patterns of climate variables, yielding multivariate red noise. The empirical technique determining multivariate red noise from observations is called linear inverse modeling (LIM). In this talk it is shown that LIMs make forecasts whose skill is competitive with current global forecast coupled GCMs. At all forecast time scales, for some seasons and regions LIM skill is actually higher on average than the CGCM. LIM can thus serve as a key forecast benchmark, and in particular can help to focus on where CGCM improvements should be targeted to yield the most significant forecast gains. The geographical and temporal variations of forecast skill are also generally similar between the LIM and CGCMs. This makes the much simpler LIM an attractive tool for assessing and diagnosing overall climate predictability as well as the predictability of climate modes such as the MJO, ENSO, PDO, and AMO.