time zones in SDK / pvwattsfunc

Experiencing a problem w.r.t. time and pvwattsfunc.

I have a weather data file from California (36 N, 116 W) who times are all in GMT. I presume, I should set the "tz" variable to 0. When I do this, I get a peak POA around solar noon at that location, GMT ~19-20ish which looks about right.

However, there appears to be a bug in which all output from morning hours is cut off completely. MAybe it is a wrap-around issue (or lack of wraparound)? For example, for a lat/lon of 36/-116, which corresponds to CA desert, the GMT hour 0 should be about 7 am local -- the sun should be up. However, the POA returned is zero regardless of what the beam and diffuse are set to.

I notice also that I can set the hour to 24,25,26, in which I get reasonable output -- but those inputs I would have hoped would have triggered error messages as the hour is supposed to be 0-23.

Anyway, two files attached:

1. example that shows the issues
2. picture of POA with fixed beam over a day. You can see that the 'sine wave' is cut off for the morning hours.

Hello,

The time stamp in the weather file should be in local time, and the time zone should be the time zone of the location.

There is a lot of error checking that the weather file could be doing, but isn't. That's because it was originally designed to work with the standard weather files from the National Solar Radiation Database, which have been through a quality control process.

Does changing the time stamps and time zone solve the problem?

Best regards,
Paul.