Abstract
Consistent with a southward bias in zonal winds in eighteen of the Intergovernmental Panel on Climate Change Fourth Assessment (IPCC AR4) simulations of the 20th century, model estimates of stormtrack location tend to cluster south of the observed stormtrack, particularly during March and April. There are two mechanisms by which a southward-displaced stormtrack could increase downstream precipitation. The first is by changing the latitudinal distribution of storms. Second, a southward-displaced stormtrack allows storms to develop over warmer sea surfaces, increasing the amount of water they hold. Although they capture the general structure and seasonality of precipitation over North America quite well, IPCC AR4 coupled model simulations of the late-20th century (1979-1999) typically overestimate winter (November to April) precipitation in western North America in comparison to values from the Global Precipitation Climatology Project version 2 (GPCPv2). While there are multiple controls on precipitation amount and distribution, we suspect that a southward bias in zonal wind speeds contributes to the precipitation bias observed in many of the models included in this study. Many of the models in this study show greater overestimates of precipitation to the south than to the north, consistent with a southward bias in stormtrack position. The generally positive bias in precipitation across western North America seen in many of the models suggests that sea-surface temperature may also play a role. As the modeling community moves toward coupled earth system models with dynamic vegetation, the precipitation bias may become a more significant problem. Vegetation types are typically determined by seasonal patterns in temperature and precipitation. Errors of even 25% in precipitation totals may contribute to significant changes in the simulated vegetation and carbon fluxes, particularly in arid and semi-arid regions like the western United States.
spacer