Results of weather forecasts

Results of weather forecasts

On this page the results of weather forecasts are shown, constantly updated, so that anyone can know the quality of the service offered.
Since the accuracy of a prediction can be understood in different ways also, depending on the users' goals, Lungimira indicates some reference intervals and considers proper or accurate a prediction when the distance from the observation is less than the interval set. A correct prediction is often not enough if it is not too precise and therefore useful in helping to make choices. Lungimira considers useful or precise a prediction when its variability interval does not exceed the range set for each of the three atmospheric variables.

Evaluation of: Precision
Accuracy Forecast: Likely (very precise) Unlikely (not precise)
Correct (very accurate) A B
Incorrect (not accurate) C D

Square correctness-utilities.

AReliable and correct forecast (precise and accurate): with a narrow range of variability and a small deviation from the recorded data.
BUnreliable but correct forecast (imprecise but accurate): with an excessive variability interval but a small deviation from the recorded data.
CReliable but incorrect forecast (precise but inaccurate): with a limited range of variability but excessive deviation from the measured data.
DUnreliable and incorrect forecast (neither precise nor accurate): with an excessive range of variability and a large deviation from the recorded data.

Lungimira took into account the majority of human activities to determine what are the sufficient conditions to define a good level of accuracy and precision.
The following list defines the ranges of accuracy and precision of a
forecast of a period of ten days for the three main atmospheric variables:

Variables Accuracy: Difference between observation and median expected Precision: Interval between percentile 5 and 95 expected
Temperature |ΔT| < 3 °C ΔT > 8 °C
Rainfall |ΔR| < 20 mm ΔR > 60 mm
Atm. Pressure. |ΔP| < 7 hPa ΔP > 15 hPa

 

From the statistics of the predictions of Lungimira, available at the end of the page, result the following mean statistical indices that describe the level of quality of the service for the three main atmospheric variables: (Preliminary values to be updated)

Reliable and accurate forecasts ([A] / [A+B+C+D]) Accuracy Index ([A+B] / [A+B+C+D]) Precision Index ([A+C] / [A+B+C+D]) Mean Error (ME) Mean Absolute Error (MAE) Mean Squared Error (MSE) Root Mean Squared Error (RMSE)
Temperature 76.0 % 80.4 % 92.1 % 0.06 °C 1.52 °C 4.27 °C^2 2.06 °C
Rainfall 68.3 % 71.5 % 91.9 % 0.10 mm 12.40 mm 474.83 mm^2 21.64 mm
Atmospheric pressure at sea level 69.7 % 75.2 % 93.1 % 0.51 hPa 3.80 hPa 29.23 hPa^2 5.32 hPa

 

From the following list you can check out the annual tables with percentages of correct predictions for the three atmospheric variables considered, estimated by means of a control sample. To view the tables select a year among those present, and will open another window. By clicking on the data in the table it appears the maps of the corresponding forecasts.

Index of years with available data: (in update)
2015 2016 2017

 

Share:
Email Facebook Twitter Google Plus Linkedin Pinterest

Analysis of recent events