MSLP change versus 5appp group change

Hi...

This isn't a gempak question directly but relates to data used by gempak.

I was looking at surface METAR data today and came upon some perplexing discrepancies between observed 3-hour differences in reported mean sea-level pressure (e.g. P(12Z)-P(09Z)) versus the reported 3-hour tendencies (5appp groups reported in the METAR remarks sections). Most stations do not exhibit the discrepancy, but a few show significant differences. I first noticed this in some case data from last July and decided to see if it shows up in todays data, and it does. For example:

KDUJ 040856Z AUTO 27013KT 10SM BKN023 OVC030 M06/M10 A3016 RMK AO2
     SNB04E44 SLP234 P0000 60000 T10611100 58002

KDUJ 040956Z AUTO 26014KT 10SM BKN021 OVC026 M07/M11 A3017 RMK AO2
     SLP238 T10671106

KDUJ 041056Z AUTO 27013KT 10SM BKN021 OVC028 M07/M11 A3018 RMK AO2
     SLP242 T10671106

KDUJ 041156Z AUTO 26014KT 6SM -SN BKN016 BKN026 OVC037 M08/M11
     A3018 RMK AO2 SNB1059 SLP244 P0000 60000 T10781111 11039
     21078 51005

The 12Z-09Z SLP difference is 1024.4-1023.4 = +1.0 mb/3hrs, whereas the 12Z 5appp group is 51005 or +0.5 mb/3hrs, a 0.5 mb discrepancy.

Similarly,

METAR KAVP 040854Z 34009G14KT 10SM BKN037 OVC085 01/M04 A3006 RMK
AO2 PK WND 230243/0803 SLP182 T00061044 58010 $

METAR KAVP 040954Z 35013G22KT 10SM SCT029 BKN036 OVC046 00/M07 A3007
RMK AO2 SLP187 T00001067 $

METAR KAVP 041054Z 33012KT 10SM CLR M01/M09 A3007 RMK AO2 SLP186
T10111094 $

METAR KAVP 041154Z 33011KT 10SM BKN034 OVC055 M01/M08 A3009 RMK AO2
SLP192 T10111078 10006 21011 53007 $

The 12Z-09Z SLP difference is 1019.2-1018.2 = +1.0 mb/3hrs, whereas the
12Z 5appp group is 53007 or +0.7 mb/3hrs, a 0.3 mb discrepancy.

The case I was looking at from July had discrepancies as large as 0.9 mb.

I suspect this is some sort of "feature" of ASOS related to the "characteristic" of the changing pressure which tries to capture a more immediate tendency than a simple 3-hour difference, but if true, that's a departure from what weather forecasters are used to seeing. Does anyone know the real reason why the ASOS 5appp tendencies don't always agree with the change in the reported pressures? Are there any documentation links you could point me to?

                            Thanks...

                              Art

Arthur A. Person
Research Assistant, System Administrator
Penn State Department of Meteorology
email:  person@xxxxxxxxxxxxx, phone:  814-863-1563