The effect of atmospheric nudging on the stratospheric residual circulation in chemistry–climate models

Abstract

Abstract. We perform the first multi-model intercomparison of the impact of nudged meteorology on the stratospheric residual circulation using hindcast simulations from the Chemistry–Climate Model Initiative (CCMI). We examine simulations over the period 1980–2009 from seven models in which the meteorological fields are nudged towards a reanalysis dataset and compare these with their equivalent free-running simulations and the reanalyses themselves. We show that for the current implementations, nudging meteorology does not constrain the mean strength of the stratospheric residual circulation and that the inter-model spread is similar, or even larger, than in the free-running simulations. The nudged models generally show slightly stronger upwelling in the tropical lower stratosphere compared to the free-running versions and exhibit marked differences compared to the directly estimated residual circulation from the reanalysis dataset they are nudged towards. Downward control calculations applied to the nudged simulations reveal substantial differences between the climatological lower-stratospheric tropical upward mass flux (TUMF) computed from the modelled wave forcing and that calculated directly from the residual circulation. This explicitly shows that nudging decouples the wave forcing and the residual circulation so that the divergence of the angular momentum flux due to the mean motion is not balanced by eddy motions, as would typically be expected in the time mean. Overall, nudging meteorological fields leads to increased inter-model spread for most of the measures of the mean climatological stratospheric residual circulation assessed in this study. In contrast, the nudged simulations show a high degree of consistency in the inter-annual variability in the TUMF in the lower stratosphere, which is primarily related to the contribution to variability from the resolved wave forcing. The more consistent inter-annual variability in TUMF in the nudged models also compares more closely with the variability found in the reanalyses, particularly in boreal winter. We apply a multiple linear regression (MLR) model to separate the drivers of inter-annual and long-term variations in the simulated TUMF; this explains up to ∼75 % of the variance in TUMF in the nudged simulations. The MLR model reveals a statistically significant positive trend in TUMF for most models over the period 1980–2009. The TUMF trend magnitude is generally larger in the nudged models compared to their free-running counterparts, but the intermodel range of trends doubles from around a factor of 2 to a factor of 4 due to nudging. Furthermore, the nudged models generally do not match the TUMF trends in the reanalysis they are nudged towards for trends over different periods in the interval 1980–2009. Hence, we conclude that nudging does not strongly constrain long-term trends simulated by the chemistry–climate model (CCM) in the residual circulation. Our findings show that while nudged simulations may, by construction, produce accurate temperatures and realistic representations of fast horizontal transport, this is not typically the case for the slower zonal mean vertical transport in the stratosphere. Consequently, caution is required when using nudged simulations to interpret the behaviour of stratospheric tracers that are affected by the residual circulation.

Description and Evaluation of the specified-dynamics experiment in the Chemistry-Climate Model Initiative

“…2b) approach ∼ 0.0008 m s −1 (also ∼ 100 % the ensemble mean climatological value). Chrysanthou et al (2019) noted similarly large differences in w * among the REF-C1SD simulations, although they examined a slightly different region in the stratosphere (10 and 70 hPa vs. the 30 and 80 hPa pressure levels examined here).…”

“…In addition to the seasonal cycle we also assess how well the simulations covary with each other on interannual timescales over years 1980-2009. As such, our assessment of interannual variability, which evaluates only the degree of correlation between time series, differs from previous studies (Chrysanthou et al, 2019), in which time series were further decomposed in terms of different modes of interannual variability (i.e., the El Niño-Southern Oscillation, the Quasi-Biennial Oscillation, etc.). More precisely, for each member within the SD ensemble we identify a given variable χ for which we first remove the linear trend and then calculate the correlation coefficient between the annual mean time series corresponding to that ensemble member i and the annual mean time series of its corresponding ensemble mean.…”

“…Finally, we conclude our analysis by comparing trends of the various measures, evaluated here simply as the linear fit to the deseasonalized time series. While previous studies have already cautioned about the suitability of inferring trends from either reanalyses (Abalos et al, 2015) or nudged experiments (Chrysanthou et al, 2019), those studies focused on the lower stratosphere (specifically, the transformed Eulerian mean circulation), and it is not clear to what extent their conclusions about trends apply to other dynamical and constituent fields.…”

“…Via (2) our goal is to present a more comprehensive evaluation of the large-scale flow than presented in Orbe et al (2018) and Yang et al (2019) and to extend our analysis to the stratosphere, which has been evaluated in CCMI models primarily using the free-running REF-C1 experiment (Dietmüller et al, 2018). Note that, while Chrysanthou et al (2019) presented the first comparison of the stratospheric residual circulation among the nudged CCMI hindcast runs, our analysis, which complements the findings presented in that study, has a broader scope by focusing on both the troposphere and the stratosphere and including discussions of large-scale transport.…”

A Regime Perspective on the North Atlantic Eddy-Driven Jet Response to Sudden Stratospheric Warmings

“…This means the nudged simulation may not lead to similar wave driving or stratospheric residual circulation as occurred during the actual SSW event simulated in the free-running experiment. The wave forcing in the free-running experiment that causes the SSW corresponds to sources and sinks of angular momentum within the stratosphere; these sources and sinks must be produced in the nudged simulation in order to constrain the zonal mean flow leading to nonconservation on angular momentum (e.g., Shepherd and Shaw 2004;Chrysanthou et al 2019). HS14 showed there are marked differences in the stratospheric Eliassen-Palm flux divergence between the SSW generated in the free-running simulation and the SSW nudged ensemble.…”