Fig. 2: The growth of different perturbations in FuXi and IFS for the forecast initialized at 0000 UTC Jul 20, 2021. | npj Climate and Atmospheric Science

Fig. 2: The growth of different perturbations in FuXi and IFS for the forecast initialized at 0000 UTC Jul 20, 2021.

From: A fast physics-based perturbation generator of machine learning weather model for efficient ensemble forecasts of tropical cyclone track

Fig. 2

Columns a–c correspond to perturbations, Gaussian noise in FuXi, and perturbations in IFS, respectively. Rows 1–6 correspond to the forecasts at 0 h, 24 h, 48 h, 72 h, 96 h, and 120 h, respectively; Contours in these figures indicate the 500-hPa geopotential height from the control forecast, colored shading is the differences in 500-hPa geopotential height between the perturbed forecast and the control forecast. Green stars represent locations of TC In-Fa. Row 7 gives the kinetic energy spectra of 500 hPa wind in control forecast (black line, mean spectra of 0–5-day forecast) and the kinetic energy spectra of 500 hPa wind in evolutionary perturbations (colored lines, colors represent different lead times). The kinetic energy spectra are evaluated between 0N and 70N.

Back to article page