Brewer‐Dobson circulation diagnosed from JRA‐55

We examine the Brewer‐Dobson circulation (BDC) in the lower stratosphere diagnosed from the Japanese 55 year Reanalysis (JRA‐55) data set and compare it with the BDCs diagnosed from the other reanalyses (i.e., ERA‐Interim and JRA‐25), and JRA‐55‐related data sets (i.e., JRA‐55C, created by assimilating only in situ observations, and JRA‐55AMIP, a simulation by a prediction model). The climatological mean seasonal change of the BDC in JRA‐55 is similar to that in ERA‐Interim but considerably different from that in JRA‐25. Dynamical and thermodynamical consistencies among the variables are greatly improved in the JRA‐55 data set. The interannual variations of the annual mean tropical upwellings in JRA‐55 are almost coincident with those in JRA‐55C. It suggests that the weakly increasing trend of the BDC found in JRA‐55 does not have anything to do with changes in the satellite observing system. The climatological mean tropical upwelling diagnosed from JRA‐55 is stronger than that from JRA‐55AMIP. This difference is presumed to partly link to the model's inability to simulate the quasi‐biennial oscillation (QBO). The JRA‐55AMIP data set, a simple simulation performed without data assimilation, does not simulate the QBO, whereas the JRA‐55 data set represents the QBO with the aid of data assimilation. The climatological mean zonal mean states of JRA‐55AMIP considerably deviate from those of JRA‐55 in the lower stratosphere. The deviation of JRA‐55AMIP is similar to the modulation pattern by the QBO in specific phase regions in which the BDC is rather weak. The simulated QBO might enhance the BDC and reduce the model biases in the tropical lower stratosphere.