Windchill Test Value / Graph Discrepancy

[Steve Schindler]

During the recent cold weather, I noticed a discrepancy between the windchill minimum value in the table and the value/time shown on the graph. For 12-24-22, the Low Wind Chill in the Since Midnight table is 23.6F at 07:31, but the graph doesn't reflect that value. It shows a low of about 27F at about 8:00. For 12-25-22, the temperature discrepancy is about the same but time difference is greater. See the screenshots below. All other values (temp, humidity, wind) agree between the table and the graphs.

I'm using a Davis VP2 with firmware version 1.82 so WeeWX is calculating the windchill value. 

Any thoughts on what's happening and how it can be corrected?

I'm in Florida so my troubleshooting/testing opportunities are limited to when a cold front passes through. However, I suspect that heat index has the same problem and, if so, I'll have plenty of time to work on that that this summer. :-)

12-24

12-25

[John Kline]

Tom and friends will know for sure, but I believe the low you are seeing (at 07:31) is from the archive_day_windchill table.

You likely have the following in your weewx.conf

    # Whether to include LOOP data in hi/low statistics

    loop_hilo = True

This tells weewx (well, it tells it to do exactly what the comment says).  Note: vantage loop packets do indeed include windchill.

The archive table (produced from vantage archive records) does not include this low and the graphs are produced from archive records.

I don’t think you are seeing anything *wrong*.

Note: these highs and lows in the daily stats get wiped out if you run wee_database —rebuild-daily (which is unfortunate).

On Dec 31, 2022, at 9:21 AM, Steve Schindler <sch...@gmail.com> wrote:

During the recent cold weather, I noticed a discrepancy between the windchill minimum value in the table and the value/time shown on the graph. For 12-24-22, the Low Wind Chill in the Since Midnight table is 23.6F at 07:31, but the graph doesn't reflect that value. It shows a low of about 27F at about 8:00. For 12-25-22, the temperature discrepancy is about the same but time difference is greater. See the screenshots below. All other values (temp, humidity, wind) agree between the table and the graphs.

<wx 2022-12-24 20:42:02.png>

12-25

<wx 2022-12-25 20:56:20.png>

--
You received this message because you are subscribed to the Google Groups "weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/weewx-user/183e065c-2fc6-4b2b-aaff-99ec35fa8ca1n%40googlegroups.com.

<wx 2022-12-25 20:56:20.png>

[Steve Schindler]

Thanks for the suggestion but I don't think that's the issue with my setup.  In a previous discussion (https://groups.google.com/g/weewx-user/c/roOG0u8YPX4/m/Tc4vWnznAgAJ), Tom and Gary confirmed that the VP2 hardware generated values for windchill are only available in the LOOP2 packets.

-------

tke...@gmail.com

Jul 24, 2020, 8:21:48 PM

to weewx-user

Geez, Gary, you're right! So, you would also have to specify LOOP2 packets. See the section [Vantage] in the User's Guide.

[Vantage]

  loop_request = 2

-tk

On Fri, Jul 24, 2020 at 5:13 PM gjr80 <gjrod...@gmail.com> wrote:

Windchill is only in included in the Davis LOOP2 packet, it’s not in the LOOP packet or archive record (the defaults) from the console. So unless you have set WeeWX to use the LOOP2 packet WeeWX will never see windchill from the console, irrespective of StdWXCalculate settings.

Gary

-------

And the User's Guide states "Not all stations support LOOP2 data. You need firmware version 1.90 or later."

So I don't think the loop_hilo parameter will affect the windchill value with my setup.

[John Kline]

> Tom and Gary confirmed that the VP2 hardware generated values for windchill are only available in the LOOP2 packets.

I’m not sure why you think that is relevant.  The question is, whether software or hardware generated, do the loop packets have windchill.  I confirmed they do contain windchill (for my VP2) before my initial reply.

On Jan 3, 2023, at 6:06 AM, Steve Schindler <sch...@gmail.com> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/weewx-user/70448a2d-6bfb-4e85-9e56-97f489dbf515n%40googlegroups.com.

[Steve Schindler]

My apologies. I probably could have phrased that better. You are correct that VP2s can generate windchill via hardware; however, an older unit like mine with older firmware doesn't send the windchill value in the LOOP packets.  

[gjr80]

Just to clarify, the linked post above dealt with an issue where the user noticed the wind chill value displayed on his Vantage2 console did not match the wind chill value displayed by WeeWX. The reason was the console displayed the value calculated by the console, but WeeWX displayed the wind chill value calculated by WeeWX (because LOOP1 packets were used by WeeWX the console windchill value was never seen by WeeWX so WeeWX calculated windchill - changing to LOOP2 packets sent the console calculated wind chill to WeeWX and WeeWX used this wind chill value rather than calculating wind chill). Different and unrelated issue to the problem outlined in this thread.

I agree with John's post; the issue is likely due to the loop_hilo setting and the source used for plots and reports. Native WeeWX plots are generated from archive table data. Archive records contain data for each observation that is representative of the archive period concerned, for temperatures this is typically the average value across the archive period. On the other hand, aggregates used in reports are based on the daily summaries which, by default, do record max/min values for each day based on loop packet max/mins. Consequently, it is not unusual (in fact I would say it happens frequently) for a day plot to not accurately show the day max/min. This is usually not an issue given the low resolution of the WeeWX plots (ie you just can't see it), however, I could see situations where there is a noticeable difference (as in this case in this thread). In fact, when it comes to derived obs such as wind chill that are dependent on a rapidly changing observation such as wind speed, it is quite possible there would be a noticeable difference between the day min value shown on a plot and the day min value shown in a report.

You could likely get rid of the discrepancy by changing loop_hilo to False, it's a personal preference, but you have to accept then your system will no longer record the max/min values seen in loop packets, rather it will record the max/min archive period average values. My preference is to have loop_hilo = True to record the loop highs and lows so my reporting is accurate and accept the occasional plot discrepancy.

Gary