This Title All WIREs
How to cite this WIREs title:
WIREs Clim Change
Impact Factor: 5.124

Communicating probabilistic information from climate model ensembles—lessons from numerical weather prediction

Full article on Wiley Online Library:   HTML PDF

Can't access this content? Tell your librarian.

Climate model ensembles are widely heralded for their potential to quantify uncertainties and generate probabilistic climate projections. However, such technical improvements to modeling science will do little to deliver on their ultimate promise of improving climate policymaking and adaptation unless the insights they generate can be effectively communicated to decision makers. While some of these communicative challenges are unique to climate ensembles, others are common to hydrometeorological modeling more generally, and to the tensions arising between the imperatives for saliency, robustness, and richness in risk communication. The paper reviews emerging approaches to visualizing and communicating climate ensembles and compares them to the more established and thoroughly evaluated communication methods used in the numerical weather prediction domains of day‐to‐day weather forecasting (in particular probabilities of precipitation), hurricane and flood warning, and seasonal forecasting. This comparative analysis informs recommendations on best practice for climate modelers, as well as prompting some further thoughts on key research challenges to improve the future communication of climate change uncertainties. WIREs Clim Change 2012. doi: 10.1002/wcc.187

This article is categorized under:

  • Climate Models and Modeling > Knowledge Generation with Models
  • Perceptions, Behavior, and Communication of Climate Change > Communication
Figure 1.

The three imperatives for visualization.

[ Normal View | Magnified View ]
Figure 2.

Multimodel global means (solid lines) and ±1 standard deviation range of individual model annual averages (shading) of surface warming for the 20th century and SRESs A2, A1B, and B1. The gray bars at right indicate the best estimate (solid line within each bar) and the likely range assessed for six SRESs. (Reprinted with permission from Ref 1. Copyright 2007 Intergovernmental Panel on Climate Change; Cambridge University Press)

[ Normal View | Magnified View ]
Figure 3.

Probability density of the strength of the meridional overturning circulation through the 21st century. (Reprinted with permission from Ref 24. Copyright 2010 Oxford University Press; Oxford University Press)

[ Normal View | Magnified View ]
Figure 4.

Changes in 20‐year mean surface air temperature over the HadSM3 grid box corresponding to Wales, in March, in response to doubled CO2. Green histogram shows 280 perturbed physics simulations of HadSM3. Black ticks show corresponding changes simulated by 12 multimodel ensemble (MME) members. Red curve shows the distribution obtained by emulating responses across the full parameter space of surface and atmospheric processes in HadSM3. The red curve also includes the broadening effect of adding the variance (but not the mean) of discrepancy. Blue curve shows the effects of weighting the emulated responses according to observational constraints. Black curve shows the posterior distribution, which includes the shift arising from adding in the mean effect of discrepancy. (Reprinted with permission from Ref 86. Copyright 2009 UKCP09)

[ Normal View | Magnified View ]
Figure 5.

Relative changes in precipitation (%) for the period 2090–2099, relative to 1980–1999. Values are multimodel averages based on the SRES A1B for December to February. White areas are where less than 66% of the models agree in the sign of the change and stippled areas are where more than 90% of the models agree in the sign of the change. (Reprinted with permission from Ref 1. Copyright 2007 IPCC; Cambridge University Press)

[ Normal View | Magnified View ]
Figure 6.

New mapping technique illustrating change in precipitation (similarly to C4) with hues and percentage model agreement across the ensemble with saturation. (Reprinted with permission from Ref 27. Copyright 2011 Copernicus Publications)

[ Normal View | Magnified View ]
Figure 7.

Red arrows track what summers could feel like in the NYC Tri‐State region over the course of the century under the higher emission scenario. Yellow arrows track what summers in these states would feel like under a lower emission scenario. (Reprinted with permission from Ref 30. Copyright 2007 Union of Concerned Scientists)

[ Normal View | Magnified View ]
Figure 8.

These ‘thermometers’ show projected increases in regional average summer temperatures for three‐time periods: early, mid‐, and late‐ century. (Reprinted with permission from Ref. 30. Copyright 2007 Union of Concerned Scientists)

[ Normal View | Magnified View ]
Figure 9.

The roulette‐style spinning wheels depict the estimated probability, or likelihood, of potential temperature change (global average surface temperature) over the next 100 years. The face of each wheel is divided into colored slices, with the size of each slice representing the estimated probability of the temperature change in the year 2100 falling within that range. The top greenhouse gamble wheel is the ‘no policy’ or reference case, in which it is assumed that no action is taken trying to curb the global emissions of greenhouse gases. The bottom greenhouse gamble wheel is the ‘with policy’ case, which assumes that policies are enacted to limit cumulative emissions of greenhouse gases over the century to 4.2 trillion metric tons, measured in CO2 equivalent. Available at: http://globalchange.mit.edu/focus‐areas/uncertainty/gamble. (Accessed July 20, 2012). By permission of MIT Global Change Program.

[ Normal View | Magnified View ]
Figure 10.

Probability of precipitation (PoP) with no graphic, from wunderground.com forecast for Des Moines, IA, USA. Available at: http://www.wunderground.com/q/zmw:50301.1.99999.

[ Normal View | Magnified View ]
Figure 11.

Probability of precipitation (PoP) with probability bar graphic for Sydney from the Bureau of Meteorology, Australia. Available at: http://www.bom.gov.au/nsw/forecasts/sydney.shtml.

[ Normal View | Magnified View ]
Figure 12.

Probability of precipitation (PoP) with probability pie charts, University of Washington Probcast. Available at: http://probcast.washington.edu/.

[ Normal View | Magnified View ]
Figure 13.

Probability of precipitation (PoP) with probability bar, note how vertical bar and blue color might cause confusion with the amount of rain. Accuweather. Available at: http://www.accuweather.com/en/us/new‐york‐ny/10017/weather‐accupop/3712_pc.

[ Normal View | Magnified View ]
Figure 14.

Time series showing 50 and 80% probability range for temperature and precipitation amount, Norwegian Meteorological Institute. Available at: http://www.yr.no/place/Norway/Oslo/Oslo/Oslo/long.html.

[ Normal View | Magnified View ]
Figure 15.

Temperature range bar, showing 90% range for predicted maximum (and minimum in separate tab) temperatures. UK Met Office. Available at: http://www.metoffice.gov.uk/public/beta/ weather/forecast/?tab=fiveDay&dayIdx=0&locId=350610.

[ Normal View | Magnified View ]
Figure 16.

Spaghetti plot showing hurricane tracks. ABC Weather. Available at: http://www.wjla.com/blogs/weather/2011/08/hurricane‐irene‐path‐projections‐spaghetti‐style‐12544.html.

[ Normal View | Magnified View ]
Figure 17.

Hurricane Cone of Uncertainty, note that the estimated ‘best forecast track’ has now been removed to avoid confusion. National Hurricane Center (US). Available at: http://newsfeed.time.com/2011/08/26/hurricane‐irenes‐path‐how‐do‐forecasters‐predict‐the‐cone‐of‐uncertainty/.

[ Normal View | Magnified View ]
Figure 18.

Further tabular detail for a selected point in southern Poland from the EFAS forecast for 12:00 h UTC on May 17, 2010 when severe flooding resulted in 2.5 billion Euros in damages. The first two rows classify the various EFAS river flow forecasts produced for that point using the deterministic rainfall forecasts from DWD (Deutscher Wetterdienst, the German national meteorological service) and ECMWF, with purple indicating flows in excess of the EFAS Severe Alert Level (SAL) corresponding to a simulated flood event with a return period of >20 years, red indicating flows in excess of the EFAS High Alert Level (HAL) and yellow in excess of the Medium Alert Level. The numbers in the subsequent boxes indicate the number of EFAS ensemble members produced using the ECMWF ensemble (EUE) and the COSMO‐LEPS limited area ensemble (COS) that exceed the HAL and SAL. Courtesy of EFAS, Joint Research Centre, European Commission, Ispra, Italy.

[ Normal View | Magnified View ]
Figure 19.

The Flood Guidance Statement issued in England and Wales by the joint Met Office/Environment Agency Flood Forecasting Centre provides a simple cartographic display. This same risk matrix is now also used by the UK Met Office as part of its National Severe Weather Warning Service to communicate the likelihood and impact of severe weather events.

[ Normal View | Magnified View ]
Figure 20.

The Vigicrues flood risk used by the SCHAPI (Service Central d'Hydrométéorologie et d'Appui à la Prévision des Inondations/Central Service for Hydrometeorology and Flood Prediction Support) in France to communicate the risk of flooding over the next 24 h on main rivers. The Green, yellow, orange, and red pixels represent escalating levels of hazardousness that call for corresponding levels of vigilance in response to the threat. These color codes do not explicitly distinguish the probability of flooding from its magnitude, which can lead to confusion.

[ Normal View | Magnified View ]
Figure 21.

In Austria, emergency services personnel working in the Abteilung Feurwehr und Zivilschutz Landeswarnzentrale (the Fire Service and Civil Defence Early Warning Centre) have additional access to much richer EP outputs, including ‘spaghetti’ plots of the 51‐member ALADIN‐LAEF of convective rainfall, which are the light colored lines in this plot which also shows the deterministic forecast (Hauptlauf) in black and the observed in red.

[ Normal View | Magnified View ]
Figure 22.

In Austria, the public has access to simplified HEPS forecasts of streamflow, with the blue line showing observation, the green line a ‘best‐guess’ forecast, and the two gray lines the 10 and 90% confidence intervals.

[ Normal View | Magnified View ]
Figure 23.

The 1‐month and 3‐month UK outlook for temperature in the context of the observed climatology. UK Met Office Seasonal Outlook. Available at: http://www.metoffice.gov.uk/publicsector/contingency‐planners.

[ Normal View | Magnified View ]
Figure 24.

National Weather Service (NWS), monthly to seasonal outlooks. Available at: http://www.cpc.ncep.noaa.gov/products/predictions/long_range/lead01/off01_temp.gif.

[ Normal View | Magnified View ]

Browse by Topic

Perceptions, Behavior, and Communication of Climate Change > Communication
Climate Models and Modeling > Knowledge Generation with Models

Access to this WIREs title is by subscription only.

Recommend to Your
Librarian Now!

The latest WIREs articles in your inbox

Sign Up for Article Alerts