One of the most important dynamic processes in the middle and upper atmosphere, gravity waves (GWs) play a key role in determining global atmospheric circulation. Gravity wave potential energy (GW <i>E</i><sub>p</sub>) is an important parameter that characterizes GW intensity, so it is critical to understand its global distribution. In this paper, a deep learning algorithm (DeepLab V3+) is used to estimate the stratospheric GW <i>E</i><sub>p</sub>. The deep learning model inputs are ERA5 reanalysis datasets and GMTED2010 terrain data. GW <i>E</i><sub>p</sub> averaged over 20−30 km from 60°S−60°N, calculated by COSMIC radio occultation (RO) data, is used as the measured value corresponding to the model output. The results show that (1) this method can effectively estimate the zonal trend of GW <i>E</i><sub>p</sub>. However, the errors between the estimated and measured value of <i>E</i><sub>p</sub> are larger in low-latitude regions than in mid-latitude regions, possibly due to the large number of convolution operations used in the deep learning model. Additionally, the measured <i>E</i><sub>p</sub> has errors associated with interpolation to the grid; this tends to be amplified in low-latitude regions because the GW <i>E</i><sub>p</sub> is larger and the RO data are relatively sparse, affecting the training accuracy. (2) The estimated <i>E</i><sub>p</sub> shows seasonal variations, which are stronger in the winter hemisphere and weaker in the summer hemisphere. (3) The effect of quasi-biennial oscillation (QBO) can be clearly observed in the monthly variation of estimated GW <i>E</i><sub>p</sub>, and its QBO amplitude may be less than that of the measured<i> E</i><sub>p</sub>.
Read full abstract