Simulating the atmospheric response to the 11-year solar cycle forcing with the UM-UKCA model: the role of detection method and natural variability

Abstract

Abstract. The 11-year solar cycle forcing is recognised as an important atmospheric\nforcing; however, there remain uncertainties in characterising the effects of\nsolar variability on the atmosphere from observations and models. Here we\npresent the first detailed assessment of the atmospheric response to the\n11-year solar cycle in the UM-UKCA (Unified Model coupled to the United Kingdom Chemistry and\nAerosol model) chemistry–climate model (CCM) using a three-member ensemble over the recent past (1966–2010). Comparison of the model\nsimulations is made with satellite observations and reanalysis datasets. The\nUM-UKCA model produces a statistically significant response to the 11-year\nsolar cycle in stratospheric temperatures, ozone and zonal winds. However,\nthere are also differences in magnitude, spatial structure and timing of the\nsignals compared to observational and reanalysis estimates. This could be due\nto deficiencies in the model performance, and so we include a critical\ndiscussion of the model limitations, and/or uncertainties in the current\nobservational estimates of the solar cycle signals. Importantly, in contrast\nto many previous studies of the solar cycle impacts, we pay particular\nattention to the role of the chosen analysis method in UM-UKCA by comparing\nthe model composite and a multiple linear regression (MLR) results. We show that\nthe stratospheric solar responses diagnosed using both techniques largely\nagree with each other within the associated uncertainties; however, the\nresults show that apparently different signals can be identified by the\nmethods in the troposphere and in the tropical lower stratosphere. Lastly, we\nexamine how internal atmospheric variability affects the detection of the\n11-year solar responses in the model by comparing the results diagnosed from\nthe three individual ensemble members (as opposed to those diagnosed from the\nfull ensemble). We show overall agreement between the responses diagnosed for\nthe ensemble members in the tropical and mid-latitude\nmid-stratosphere to lower mesosphere but larger apparent differences at Northern Hemisphere\n(NH) high latitudes during the dynamically active season. Our UM-UKCA results\nsuggest the need for long data sets for confident detection of solar cycle\nimpacts in the atmosphere, as well as for more research on possible\ninterdependence of the solar cycle forcing with other atmospheric forcings\nand processes (e.g. Quasi-Biennial Oscillation, QBO; El Nino–Southern\nOscillation, ENSO).