Please excuse the rantings of a weather ignoramus, but it seems some aspects of the forecasts and models are being neglected.

1. I fail to see the objection to 5 day forecasts. Most negative comments are not about the forecast but the use of the forecast. Perhaps the problem is not the forecast (a set of 4-dimensional locations) but their presentation.

The NHC adds to the set of points line segments (most probable path?) and a "cone" (likely area to be within the eye?). Perhaps more useful additions to the points would be arcs to indicate distance the system might travel in a given time. In contrast to the current "cone" these arcs need not be symmetrical. The current presentation indicates an equal degree of uncertainty in either lateral direction.

Also, if the "cone" is eliminated, the points could become centers for an overlay of the predicted size of the system. The current graphic shows how large an area MIGHT be affected, but not how large an area WILL be affected. The arc through the center of this representation would show the uncertainty of where that area would be.

2. Most of the comments about the discrepancy between the forecast position and the actual position are 4-dimensional. My casual obsevation of the paths of storms shows greater correlation in 3 dimensions. While the pace of the storms has not been well predicted, the variations from the geographic paths have not been so problematic.

Since location in time is so difficult, perhaps another item could be added to the presentaion suggested above. In addition to the arc through the predicted location, perhaps another line segment along the anticipated path to show the 90 percentile for distance travelled. This segment will be much longer for points beyond 24 hours, but would better indicate the degree of uncertainty.

3. Finally, I have seen much discussion, comparison and critique of various models. I have seen little mention of what these models are working with. Each model is an attempt to reproduce a highly complex analog system with a very restricted set of digital data. It is my view the problems are not with the models, but with the absence of data.

Before any model can ever be relied upon, much more environmental data (historical and current) will be needed. I do not know how fine grain the data needs to be, but it certainly will have to not only be more frequent, but be much more closely spaced on the surface, below the surface, as well as in the atmosphere.

Even IF one of the current models was "perfect" we would never know it because we are unable to feed it the data necessary to correctly represent the external reality.