Tropospheric Winds over West Sumatra - a Comparison between ERA-5 Reanalysis and Equatorial Atmospheric Radar

This study compares representation of horizontal and vertical wind over west Sumatra in ERA-5 reanalysis data and observations from the Equatorial Atmospheric Radar located there. The comparison focuses on mean horizontal and vertical wind profiles throughout an extended boreal winter (November – March), the sub-diurnal variation of all wind components as well as wind variability associated with assorted large scale phenomena (ENSO, QBO, MJO and CCKW). The analysis is performed based an analysis of 20-years of data. Overall, the average horizontal wind data observed by radar closely aligns with the reanalysis data, particularly in the lower troposphere. However, discrepancies emerge in the upper troposphere, where the reanalysis consistently shows greater magnitudes compared to radar measurements. This analysis has been augmented with horizontal winds data from upper sounding launched daily in Padang, West Sumatra, which show magnitudes higher than in ERA-5 throughout the troposphere. In contrast, the average vertical wind data from radar observations significantly diverge from the representation in the reanalysis, with substantial discrepancies between the two datasets. Namely, reanalysis fail in representation of mean downward motions in lower troposphere.

The analysis of diurnal variations for horizontal wind components shows general agreement between reanalysis and radar data in terms of phase and amplitude, with minor discrepancies in the upper troposphere (Figure 1). However, significant differences are observed for vertical wind, where EAR data show coherent evolution throughout a day, which can be related to diurnal evolution of convection over the region, while reanalysis data depict a relatively constant phase with height and significantly greater amplitude.

The study further explores the sub-diurnal variability of tropospheric winds, emphasizing the significance of diurnal and semi-diurnal cycles in daily wind variability. The analysis reveals that decomposition of diurnal variations into diurnal and semi-diurnal cycles explain a vast majority of observed variability, with correlations exceeding 0.95 for horizontal wind and over 0.85 for vertical wind, confirming the dominance of these modes in sub-diurnal variations in the investigated region.

Regarding variability of tropospheric winds in the West Sumatra related to the large scale circulation patterns, both reanalysis and radar data generally agree on the magnitude of horizontal wind variability. However, noticeable differences arise with QBO and MJO, where reanalysis indicates significantly greater variability, especially in the upper troposphere.

Furthermore, a detailed analysis of differences between EAR and ERA-5, focusing on the influence of the large-scale patterns on mean tropospheric wind and diurnal cycle amplitudes has been performed (Figure 2). Variability ranges for different wind components and diurnal cycle amplitudes are assessed across various tropospheric layers and large-scale modes. Notably, in the upper troposphere, ERA-5 consistently exhibits greater variability than EAR observations, particularly during a MJO.

The analysis extends to the diurnal amplitude of zonal and meridional winds. In the lowest layer, EAR data show greater variability in zonal wind amplitude than ERA-5, while differences in meridional wind variability are minimal. These patterns change with altitude, emphasizing the complex interplay of atmospheric dynamics.

Additionally, the study examines the amplitude of the diurnal wind cycle and its response to ENSO, revealing comparable representations by EAR and ERA-5 across atmospheric layers. However, notable contrasts emerge in the representation of MJO and CCKWs. ERA-5 consistently exhibits significantly greater variability in the diurnal amplitude of zonal wind in specific layers, indicating regional variations. Similarly, the amplitude of the diurnal vertical wind cycle during CCKW activity shows distinctive patterns in EAR and ERA-5, emphasizing the importance of considering regional variations in atmospheric processes.

In conclusion, while ERA-5 demonstrates effectiveness in representing average horizontal wind and its variability due to large-scale modes over Western Sumatra, significant discrepancies exist in capturing vertical wind patterns. Discrepancies between reanalysis and observations extend to the analysis of diurnal cycle variations influenced by ENSO, QBO, MJO or CCKW. These findings underscore the importance of wind observations in the Sumatra region and the broader Maritime Continent for a comprehensive understanding of atmospheric processes in the area.