Evaluation of Hydrologically Relevant Pcm Climate Variables and Large-scale Variability over the Continental U.s

The ability of the Parallel Climate Model (PCM) to reproduce the mean and variability of hydrologically relevant climate variables was evaluated by comparing PCM historical climate runs with observations over temporal scales from sub-daily to annual. The domain was the continental U.S, and the model spatial resolution was T42 (about 2.8 degrees latitude by longitude). The climate variables evaluated include precipitation, surface air temperature, net surface solar radiation, soil moisture, and snow water equivalent. The results show that PCM has a winter dry bias in the Pacific Northwest and a summer wet bias in the central plains. The diurnal precipitation variation in summer is much stronger than observed, with an afternoon maximum in summer precipitation over much of the U.S. interior, in contrast with an observed nocturnal maximum in parts of the interior. PCM has a cold bias in annual mean temperature over most of the U.S., with deviations as large as –8 K. The PCM daily temperature range is lower than observed, especially in the central U.S. PCM generally overestimates the net solar radiation over most of the U.S, although the diurnal cycle is simulated well in spring, summer and winter. In autumn PCM has a pronounced noontime peak in solar radiation that differs by 5–10% from observations. PCM’s simulated soil moisture is less variable than that of a sophisticated land-surface hydrology model, especially in the interior of the country. PCM simulates the wetter conditions over the southeastern U.S. and California during warm (El Niño) events, but shifts the drier conditions in the Pacific Northwest northward and underestimates their magnitude. The temperature response to the North Pacific Oscillation is generally captured by PCM, but the amplitude of this response is overestimated by a factor of about two.