Skip to main content

File: el-nino-warm-events-accuracy-dynamical-statistical.png

File name: el-nino-warm-events-accuracy-dynamical-statistical.png
Original Resolution (pixels): 2400 × 1599
File Size: 128 KB (MIME Type: image/png)
Alternate Text: Line graph with red solid and dotted lines showing the accuracy at predicting El Nino onset by climate models. Solid lines representing dynamical models start with high accuracy before falling to 30% by the middle of their forecast. Statistical models show lower accuracy.
 View image full size
 Download file
Line graph with red solid and dotted lines showing the accuracy at predicting El Nino onset by climate models. Solid lines representing dynamical models start with high accuracy before falling to 30% by the middle of their forecast. Statistical models show lower accuracy.

Image caption

How often (percent) real-time climate model forecasts correctly predict the onset of El Nino (3-month average Nino3.4 values greater than or equal to 0.5C) at increasingly far-off lead times. Solid lines are the multi-model mean of dynamical model forecasts. Dotted lines are the multi-model mean of statistical model forecasts. Dynamical models have excellent accuracy at predicting the onset through the first three seasonal lead times before dropping to 30-40% accuracy from lead-4 onwards. Statistical models show less accuracy overall. Climate.gov image adapted from Ehsan et al., 2024.