Following is some information examining some disagreement of altimeter readings from station AT008/N3PB, its closest neighbors AS660, and KBCB, and the CWOP QA analysis for AT008. Both AT008 and AS660 are Davis VantagePro2 with 24-hour fan aspirated radiation shield. AT008 is running wview on Linux and AS660 is running Davis's WeatherLink on MS Windows.
The symptom of the problem is that the altimeter readings for AT008 tend to overshoot the peaks of the analysis line 1-2mb. This problem seems worse when there is a large temperature swing during a day. Example: Daily Weather Quality Charts for AT008: 2008-Sep-18.
The Vantage Pro2 reports sea-level pressure (SLP) on its display and in data downloaded from it. It calculates SLP by taking the station pressure (SP) reading and correcting it for elevation, 12-hour average temperature, and humidity. The calculations are explained in Davis's Application Note 28: Derived Variables in Davis Weather Products.
The pressure reported to CWOP is altimeter. Altimeter is calculated from station pressure by correcting it for elevation only. Any software that needs to report altimeter must convert the reported sea-level pressure back to station pressure and then convert that to altimeter. Wview converts from SLP back to SP with a formula that uses the current outside temperature and not the 12-hour average. That can clearly result in some differences, but how significant are they? Fortunately, the VP2 can report the current reduction factor (R) it is using and some other data used for pressure conversion with the BARDATA command. The main formula of interest is SP = SLP/R which gives you the current station pressure from the current sea-level pressure.
The key factor to test assumptions about why the data are different is to get the station pressure from the VP2. The BARDATA command provides the necessary parameter (R) and wview was modified to log the results of the command when archive records are saved. Station pressure is then calculated from the SLP and R, converted to altimeter with wview, and logged. This gives what I'm calling the VP2 Altimeter. These changes log data only, no changes to what wview reports or archives have been made.
This first graph shows the altimeter as calculated by wview,
the altimeter calculated based on the VP2's R, and the
CWOP analysis line.
Here is a plot of the altimeter error (as determined from the
CWOP analysis) versus outside temperature.
Here is a plot of the outside temperature versus the VP2's
12-hour average.
