Verification of the First 11 Years of IRI’s Seasonal Climate Forecasts

Abstract

This paper examines the quality of seasonal probabilistic forecasts of near-global temperature and precipitation issued by the International Research Institute for Climate and Society (IRI) from late 1997 through 2008, using mainly a two-tiered multimodel dynamical prediction system. Skill levels, while modest when globally averaged, depend markedly on season and location and average higher in the tropics than extratropics. To first order, seasons and regions of useful skill correspond to known direct effects as well as remote teleconnections from anomalies of tropical sea surface temperature in the Pacific Ocean (e.g., ENSO related) and in other tropical basins. This result is consistent with previous skill assessments by IRI and others and suggests skill levels beneficial to informed clients making climate risk management decisions for specific applications. Skill levels for temperature are generally higher, and less seasonally and regionally dependent, than those for precipitation, partly because of correct forecasts of enhanced probabilities for above-normal temperatures associated with warming trends. However, underforecasting of above-normal temperatures suggests that the dynamical forecast system could be improved through inclusion of time-varying greenhouse gas concentrations. Skills of the objective multimodel probability forecasts, used as the primary basis for the final forecaster-modified issued forecasts, are comparable to those of the final forecasts, but their probabilistic reliability is somewhat weaker. Automated recalibration of the multimodel output should permit improvements to their reliability, allowing them to be issued as is. IRI is currently developing single-tier prediction components.

Skill of ECMWF system‐4 ensemble seasonal climate forecasts for East Africa

“…For detail see (Buizza and Palmer, 1998;Mason and Graham, 1999;Hamill and Juras, 2006;Mason and Stephenson, 2008;Barnston et al, 2010;Diro et al, 2012). Appendix C. Ranked probability score and ranked probability skill score (RPSS) RPS is a squared measure that compares the cumulative density function (CDF) of a probabilistic forecast with the CDF of the corresponding observation over a given number of discrete probability categories (Epstein, 1969;Weigel et al, 2007b).…”

“…A 100% RPSS implies a perfect probabilistic forecast while a negative value of RPSS means the skill of the forecast probability is worse than the reference. For details, see Epstein (1969), Palmer (1998), Hersbach (2000), Kumar et al (2001), Doblas-Reyes et al (2003), Mason (2004), Müller et al (2005, Weigel et al (2007aWeigel et al ( , 2007b, Tippett and Barnston (2008), Barnston et al (2010), Diro et al (2012), Manzanas et al (2014).…”

A verification framework for interannual-to-decadal predictions experiments

“…For predictions of seasonal climate anomalies, for example, skill estimates are obtained from a set of hindcasts for which the observations are also available (Barnston et al 2010;Wang et al 2010). The use of the forecast information without reference to the skill of similar forecasts for past cases invites undue confidence in the forecast information, whereas information about how skillful past predictions or forecasts have been provides historical context that can assist users to incorporate real-time forecasts into their decision making process.…”

“…The ranked probability score is commonly used to assess probabilistic forecasts (e.g. Goddard et al 2003;Barnston et al 2010), but is typically used with categorical forecasts. Since the changing background climate subverts the usefulness of categorical forecasts, we wish to cast the hindcasts in terms of a continuous, quantitative, analytical distribution with a mean and standard deviation determined from the hindcast ensemble, although clearly both of these parameters are subject to substantial sampling errors with the small nominal ensemble sizes requested for CMIP5.…”

Should seasonal rainfall forecasts be used for flood preparedness?

“…To calculate seasonal total rainfall, we sum the daily rainfall estimates for each overlapping 3-month season (JFM, FMA, etc.) over a 2.5 • grid box, as this is the resolution of many seasonal forecasting products from the Global Producing Centres for Long-Range Forecasts (Barnston et al, 2003;WMO, 2017).…”