In principle, rapid growth of forecast errors in convective or smaller scales places a severe constraint on the skill of forecasts for larger scales. Recent papers by the authors have demonstrated for the "surprise" snowstorm of 24-25 January 2000 that initial differences grew rapidly at scales below 200 km in the presence of moist processes, but numerous questions remain to be answered, including the generality of such rapid growth, the mechanisms involved, and the quantitative limits implied for larger-scale forecasts.

We have generalized the results from the previous case study through the investigation of error growth in an idealized baroclinic wave amplifying in a conditionally unstable atmosphere. Without the effects of moisture, there is little error growth in the short term (0-36 h) forecast error (starting from random noise), even though the basic jet used produces a rapidly growing synoptic-scale disturbance; the error is characterized by upscale grow, basically as found in the study of the "surprise" snowstorm. Results will be presented at the conference, together with ongoing work using both idealized systems, with a range of time and space scales, and high-resolution explicit simulations of moist convection within an idealized baroclinic wave using a multiply nested mesoscale model.