| Abstract: | Interannual variability of regional climate was investigated on a seasonal basis. Observations and two global climate model (GCM) simulations were intercompared to identify model biases and climate change signals due to the enhanced greenhouse effect. Observed record length varies from 40 to 100 years, while the model output comes from two 100-year equilibrium climate simulations corresponding to atmospheric greenhouse gas concentrations at observed 1990 and projected 2050 levels. The GCM includes an atmosphere based on the NCAR CCM1 with the addition of the radiative effects of CH4, N2O and CFCs, a bulk layer land surface and a mixed-layer ocean with thermodynamic sea-ice and fixed meridional oceanic heat transport.Because comparisons of interannual variability are sensitive to the time period chosen, a climate ensemble technique has been developed. This technique provides comparisons between variance ratios of two time series for all possible contiguous sub-periods of a fixed length. The time autocorrelation is thus preserved within each sub-period. The optimal sub-period length was found to be 30 years, based on which robust statistics of the ensemble were obtained to identify substantial differences in interannual variability that are both physically important and statistically significant.Several aspects of observed interannual variability were reproduced by the GCM. These include: global surface air temperature; Arctic sea-ice extent; and regional variability of surface air temperature, sea level pressure and 500 mb height over about one quarter of the observed data domains. Substantial biases, however, exist over broad regions, where strong seasonality and systematic links between variables were identified. For instance, during summer substantially greater model variability was found for both surface air temperature and sea-level pressure over land areas between 20–50°N, while this tendency was confined to 20–30°N in other seasons. When greenhouse gas concentrations increase, atmospheric moisture variability is substantially larger over areas that experience the greatest surface warming. This corresponds to an intensified hydrologic cycle and, hence, regional increases in precipitation variability. Surface air temperature variability increases where hydrologic processes vary greatly or where mean soil moisture is much reduced. In contrast, temperature variability decreases substantially where sea-ice melts completely. These results indicate that regional changes in interannual variability due to the enhanced greenhouse effect are associated with mechanisms that depend on the variable and season. |
