Abstract. The simulation of intraseasonal precipitation variability over China in extended
summer (May–October) is evaluated based on six climate simulations of the
Met Office Unified Model. Two simulations use the Global Atmosphere 6.0 (GA6)
and four the Global Coupled 2.0 (GC2) configuration. Model biases are large
such that mean precipitation and intraseasonal variability reach twice their
observed values, particularly in southern China. To test the impact of
air–sea coupling and horizontal resolution, GA6 and GC2 at horizontal
resolutions corresponding to ∼25, 60, and 135 km at 50∘ N are
analyzed. Increasing the horizontal resolution and adding air–sea coupling
have little effect on these biases. Pre-monsoon rainfall in the Yangtze River
basin is too strong in all simulations. Simulated rainfall amounts in June
are too high along the southern coast and persist in the coastal region
through July, with only a weak northward progression. The observed northward
propagation of the Meiyu–Baiu–Changma rainband from spring to late summer
is poor in all GA6 and GC2 simulations. To assess how well the MetUM
simulates spatial patterns of temporally coherent precipitation, empirical
orthogonal teleconnection (EOT) analysis is applied to pentad-mean
precipitation. Patterns are connected to large-scale processes by regressing
atmospheric fields onto the EOT pentad time series. Most observed patterns of
intraseasonal rainfall variability are found in all simulations, including
the associated observed mechanisms. This suggests that GA6 and GC2 may
provide useful predictions of summer intraseasonal variability despite their
substantial biases in mean precipitation and overall intraseasonal variance.