Surface analysis has long been a key component to weather analysis and forecasting. Current surface analysis techniques are based upon the Norwegian conceptual model for cyclone development. Mass (1991) criticized current operational analysis techniques as deficient for the use of NCM frontal symbols to depict several different types of surface boundaries including: cyclone related frontal zones; shallow topographically induced boundaries; and shallow diabatically produced boundaries. In addition to these deficiencies, fronts are being defined through analysis of surface features that are sometimes, but not always, collocated with surface temperature boundaries. Sanders and Doswell (1995) list these features as: pressure troughs; pressure tendencies; wind shifts; dewpoint differences; and clouds and precipitation. Sanders and Doswell suggest that these are secondary and perhaps sometimes tertiary indications of the location of a surface thermal boundary. Uccellini et al. (1992) showed quite convincingly that even a group of noted research and operational meteorologists use different weighting of these secondary features to produce large differences in location of surface frontal features. Sanders and Doswell, therefore, suggest that surface analysis should focus on the thermal boundaries. In practice, however, the location of these secondary features, especially wind shifts, clouds and precipitation, may be important to forecasters.

In order to rectify the confusion in surface analysis related to the use of the NCM and the secondary frontal characteristics, Sanders (1999) proposed an alternate method of surface analysis. This method centers on the routine analysis of surface potential temperatures. The term "front" is reserved for those surface features that have: either a moderate or strong surface potential temperature gradient (defined to be 8 K 220 km-1 and 8 K 110 km-1, respectively) AND an associated pressure trough and wind shift. Surface features that have significant wind shifts or pressure troughs without a moderate or strong potential temperature gradients are labeled baroclinic troughs. All areas of strong potential temperature gradient not associated with significant pressure troughs are termed non-frontal baroclinic zones.

In order to assess the potential impact of Sanders' suggested analysis technique, maps of surface potential temperature and its gradient have been produced in the University at Albany maproom since November 1998 and can be found on the world wide web at: http://www.atmos.albany.edu/das/data.html. A comparison of these surface potential temperature maps to the corresponding surface analyses prepared by NCEP has been undertaken for the winter season (DJF)of 1999-2000 in order to examine the use of Sanders' alternative surface analysis method on a daily basis. Comparisons were made using twice daily analyses (00 and 12 UTC). The results show many fronts analyzed by NCEP do not meet Sanders' criteria. Cold fronts only met the criteria 52% of the time, while warm and stationary fronts met the criteria 27% and 23% of the time, respectively. Nearly 40% of all troughs analyzed by NCEP were in fact associated with moderate potential temperature gradients. In addition, 660 non-frontal baroclinic zones were observed over the three month period. These zones are often not analyzed on the NCEP maps. Case study examples will be presented to highlight the major findings of the statistical comparison.