The ESRL website will be unavailable for 24 hours starting Friday, March 6th at 5:00pm MT due to building maintenance.

Hoerling, M. P., and A. Kumar, 2003: Seasonal and interannual weather prediction. In Encyclopedia of Atmospheric Sciences, J. R. Holton, J. Pyle, and J. A. Curry (Eds.), Academic Press, 2562-2567.


Predictions of seasonal and interannual weather (climate) are predictions of the deviations from the normal march of the seasons. The predictions are of the statistics of the atmospheric states within a season. These are generally time-averaged, and a common prediction is of the seasonal mean anomaly together with a measure for its likelihood of occurrence. These are conventionally referred to as climate predictions, and are distinguished from short-range weather predictions that address the sequence of daily weather behavior.

The feasibility of seasonal and interannual climate prediction rests on the existence of slow, and predictable, variations in the Earth's boundary conditions. To be useful, these must exert a control on climate that is detectable among its random fluctuations. The expected skill of climate predictions depends on the relative strength of the detectable 'climate signal' due to boundary forcing and the 'climate noise' that is due to intrinsic random variations.

The leading boundary source of seasonal and interannual climate predictability is the ocean. Its thermal inertia produces long time scales of variations relative to those in the atmosphere, and it exerts an appreciable control on the atmosphere. Foremost among these are the tropical ocean states associated with El Niño/Southern Oscillation. Other sources may originate from air-sea interactions over the extratropical oceans and the impact of land surface factors such as soil moisture and snow cover.

The practice of seasonal and interannual climate prediction involves empirical techniques trained on historical observations, and dynamical techniques using general circulation models. At times these methods are comingled to take advantage of their respective strengths. Regardless of the procedure used, these predictions are probablistic rather than deterministic. This circumstance reflects the influence of random weather fluctuations that are not controlled by perturbations in boundary forcing. Seasonal predictions therefore express the shift in probability of the predictand, a common example being categorical forecasts of the tercile range of below-normal, near-normal, and above-normal seasonal mean temperature and precipitation.