Extended Range Graphical Output

Extended range products are available on ecCharts:

(all charts correspond to weekly means)

  • Anomaly:
    • Temperature of the earth's surface
    • 2m temperature
    • Total precipitation
    • MSLP
    • 500hPa height
    • Winds at 10m, 100m, 850hPa, 700hPa, 500hPa, 200hPa
    • Sunshine duration
    • 10hPa temperature
    • 10hPa wind
  • Absolute:
    • 500hPa height
  • Probability distribution:
    • Temperature of the earth's surface
    • 2m temperature
    • Total precipitation
    • MSLP
  • Anomaly probability (for various pre-defined quantiles):
    • Temperature of the earth's surface
    • 2m temperature
    • Total precipitation
    • MSLP
  • Extreme Forecast Index (EFI) and Shift of Tails (SOT):
    • Temperature at 2m 
    • Total precipitation



Extended range products are available on web open charts (ecCharts):

(all charts correspond to weekly means)

Weekly mean anomaly from model climate.

The weekly mean anomaly charts display the anomaly between the forecast weekly mean and the corresponding weekly mean in the ER-M-climate.

Charts available are: 2m temperature, surface temperature, total precipitation, mean sea level pressure, winds. 

Click the mouse over any location on the ensemble charts to produce probability information in diagrammatic form.  Diagrams available:

Click on the central small icon in the bottom right of the web frame to show the colour scale of values appropriate to each display.

Probability that weekly mean anomaly from model climate greater than zero

The weekly anomaly probability charts display the probability that the weekly mean anomalies are greater than zero.  The probability is calculated from the number of ensemble members which show an anomaly within the higher half of the the ER-M-climate distribution. 

Charts available are: 2m temperature, surface temperature, total precipitation, mean sea level pressure. 

Two equally probable domains can be defined: below normal and above normal.

  • The chart displays the probability of the forecast anomaly lying in the upper half of the ER-M-climate distribution (i.e. above normal; warmer or wetter etc than the mean of the ER-M-climate). 

White on the plot means that either:

  • the proportion of members above the ER-M-climate mean is between 40% and 60%,
  • or that that proportion, whatever it is, is not statistically significant.

This plot structure circumvents the fact that some ER-M-Climate distributions will be skewed (i.e. the climatological probability of seeing more than the mean is far from 50%).

Note: On precipitation charts the range of colours available for "below average" anomalies (brown colours) is constrained by the local climatology in the ER-M-climate.  For example if all ENS members showed a dry week, the mean (-ve) anomaly could be no larger in magnitude than the mean in the ER-M-climate.  So in some locations the strongest dry signal one can ever see will only be in the first or the second of the brown shades.

Click the mouse over any location on the ensemble charts to produce probability information in diagrammatic form.  Diagrams available:

Click on the central small icon in the bottom right of the web frame to show the colour scale of values appropriate to each display.

Probability of weekly mean anomaly from model climate (in terciles, quintiles, deciles)

The weekly anomaly probability charts display the probability that the weekly mean anomalies are in the lower or uppermost third (tercile), fifth (quintile) or tenth (decile) of the ER-M-climate distribution.

Charts available are: 2m temperature, surface temperature, total precipitation, mean sea level pressure.

The probability is calculated from the number of ensemble members which show an anomaly within the highest or lowest tercile, quintile or decile of the the ER-M-climate.

Terciles:  Three equally probable domains can be defined: below normal, normal and above normal.

  • The upper tercile chart displays the probability of the forecast anomaly lying in the upper tercile of the ER-M-climate distribution (i.e. warmer or wetter).
  • The lower tercile chart displays the probability of the forecast anomaly lying in the lower tercile of the ER-M-climate distribution (i.e. cooler or drier etc).

Quintiles:  Five equally probable domains can be defined:

  • The upper quintile chart displays the probability of the forecast anomaly lying in the highest quintile of the ER-M-climate distribution (i.e. warmest or wettest).
  • The lower quintile chart displays the probability of the forecast anomaly lying in the lowest tercile of the ER-M-climate distribution (i.e. coolest or driest).

Deciles:  Five equally probable domains can be defined.

  • The upper decile chart displays the probability of the forecast anomaly lying in the highest decile of the ER-M-climate distribution (i.e. very warmest or very wettest).
  • The lower decile chart displays the probability of the forecast anomaly lying in the lowest decile of the ER-M-climate distribution (i.e. very coolest or very driest).

Click the mouse over any location on the ensemble charts to produce probability information in diagrammatic form.  Diagrams available:

Click on the central small icon in the bottom right of the web frame to show the colour scale of values appropriate to each display.

Multiparameter outlook - Extended range forecast

These multi-parameter charts charts display the ensemble weekly mean:

  • 500hPa geopotential height.
  • anomaly of 2m temperature.
  • anomaly of 10m wind.
  • anomaly of sunshine duration; the fraction of time there is sunshine compared with mean sunshine duration in the ER-M-climate (e.g. +0.01 means 1% more sunshine than in the ER-M-climate; -0.05 means 5% less sunshine).

Click the mouse over any location on the ensemble charts to produce probability information in diagrammatic form.  Diagrams available:

Click on the central small icon in the bottom right of the web frame to show the colour scale of values appropriate to each display.

Hovmöller diagram or Time-Longitudes diagram - Extended range forecast 

Hovmöller or Time-Longitudes diagrams show the time evolution of the ensemble mean anomaly of a parameter.  The x-axis represents the longitude, the y-axis represents the time evolution (time increasing downwards).  Past results lie above the horizontal line and forecast results lie below.  

The northern mid-latitude Hovmöller diagrams show the time evolution of the ensemble mean anomaly of geopotential height at 500hPa or 1000hPa, averaged over the latitude band 35N-60N (Northern extra-tropics) or 25S-50S (Southern extra-tropics).  The anomaly has been computed by averaging all the members of the real-time forecast and subtracting the mean of the ER-M-climate.   Contours are plotted every 1.5dam.  Since it is an ensemble mean, the structures shown below the horizontal line are much more detailed in the first days of the forecast (top part) than in the last days (bottom part).   Shaded areas represent the ensemble spread and are displayed only when the amplitude of the anomaly exceeds 2dam.  On average spread will naturally increase with forecast lead time, though occasionally, when moving to longer lead times, there can be a reduction.

The Madden-Julian Oscillation (MJO) Hovmöller diagrams show the ensemble mean anomalies of outgoing long wave radiation, zonal wind at 850hPa, and velocity potential at 200hPa averaged over a tropical band (15N-15S).  

Large scale mean flow - Weekly mean anomalies 500hPa and 10hPa

The mean flow and anomaly charts show, at global or regional scales, the weekly:

  • mean 500hPa geopotential height and the height anomaly from the corresponding weekly mean 500hPa geopotential height of the ER-M-climate.
  • mean 10hPa geopotential height and the temperature anomaly from the corresponding weekly mean 10hPa temperature of the ER-M-climate.

Click the mouse over any location on the ensemble charts to produce probability information in diagrammatic form.  Diagrams available:

Click on the central small icon in the bottom right of the web frame to show the colour scale of values appropriate to each display.


Measures of skill of forecasts. 

The extended range products should always be used with historical skill metrics in mind.

Broad indications of ensemble performance and predictability in the extended range should be taken into account when considering the extended range output.  Users should consult verification information:

Relative Operating Characteristics diagrams

The Relative Operating Characteristics give a measure of the effectiveness of a forecast system to predict an event that actually happens balanced against forecast of an event that fails to occur.   ROC is derived from measuring the area beneath the results plotted on a ROC diagram.  ROC area values indicate:

  • ROC value 1.0 implies consistent Hit Rate and no False Alarms.
  • ROC value 0.5 implies no skill with as many Hits as False Alarms. 
  • ROC value 0.0 implies consistent False Alarm rate and no Hits.


Reliability diagrams

The Reliability diagrams give a measure of the tendency of the forecast system to over- or under-forecast and event.  The diagram plots the frequency of a forecast probability of an event against the frequency that the event occurs.   Ideally these should match and is shown by the diagonal line. Where the plot lies:

  • below the diagonal implies over-forecasting.  The forecast probability is greater than the observed frequency.
  • above the diagonal implies under-forecasting.  The forecast probability is less than the observed frequency.

An explanation of reliability diagrams and ROC diagrams is given in the annex to this guide.


ROCmap - map of Relative Operating Characteristics (ROC)

The ROCmap shows the ROC score computed over each grid point with a resolution of 2.5 degrees longitude by 2.5 degrees latitude.  The charts are anomalies derived from previous extended range ensemble forecasts compared with ER-M-climate. 

The map colours show values of the ROC score shaded according to the scale above the map:

  • Red: ROC > 0.5 - the monthly forecast has more skill than climatology (pink, reds, dark browns).
  • Blue: ROC < 0.5 - the monthly forecast has less skill than climatology (cyan, blues, dark blues).


Ranked Probability Skill Score (RPSS) chart  

The RPSS map shows the RPSS score computed over each grid point with a resolution of 2.5 degrees longitude by 2.5 degrees latitude.

The Ranked Probability Skill Score (RPSS) compares the Ranked Probability Score of a probabilistic forecast system with a reference of some sort, usually climatology.  

The Ranked Probability Skill Score  charts give a measure of the extended range ensemble forecasts and is a measure of how good forecasts are in matching observed outcomes.  Forecasts falling into the upper or lower terciles of ER-M-climate, are verified against the subsequent reanalysis or operational analysis (for precipitation, 24hr forecasts).  This is equivalent to the Brier skill score where upper and lower terciles are used.

The map colours show values of the RPSS score (equivalent to the Brier skill score) shaded according to the scale above the map:

  • RPSS = 1 the forecast has perfect skill compared to the reference (assumed to be climatology) - the forecast is beneficial;
  • RPSS = 0 the forecast has no skill compared to the reference - the forecast has no benefit over climatology;
  • RPSS = a negative value the forecast is less accurate than the reference  - the forecast is misleading.