Timezone mismatch in Bird Clear Sky Model

[yhkee0404]

Timezone in bird_08_16_2012.xls file is ambiguous when interpreting DOY (Day of Year), HR (Hour of Day), and Decimal Time (Day of year and hour in decimal DD.DDDD), which is contained in Column B, C, and V, respectively. Hour Angle (Hour angle of the sun with respect to 0= azimuth of 180 degrees) in Column H considers the timezone as -7 in Cell A13 as its value gets nearest to zero whenever HR is about 12, or the local solar noon. However, timezone is thought as UTC rather than a local timezone in calculations of Dangle (Day Angle; representing position of Earth in orbit around the sun) in Column E, DEC (Solar Declination) in Column F, and EQT (Equation of Time for the sun) in Column G because their values are not local-specific but globally measured in its nature and dependent only on DOY, ignoring HR.

The issue was found in pvlib/tests/test_irradiance.py::test_bird and closed by pvlib/pvlib-python#2055 but then I came here since I thought the fundamental cause of the issue remains in the original bird_08_16_2012.xls file from NREL.

To unify its inconsistent assumption of timezone, I think we should first decide it between -7 and UTC. We can apply only one of the following two approaches but never both.

Timezone synchronization into -7

We could correct only Dangle from UTC to -7 to make any inconsistencies in the other columns like DEC, EQT, etc., be fixed accordingly. For example, Cell E3 contains this formula which results in UTC value:

    =6.283185*(B3-1)/365

The corrected formula for the localized value is:

    =6.283185*(OFFSET(B3, -$A$13, 0)-1)/365

Or more simply making Cell E3 refer to Cell B10 instead of B3:

    =6.283185*(B10-1)/365

Timezone synchronization into UTC

We could select only Hour Angle to change. The previous formula of Cell H3 considering local timezone is:

    =15*(C3-12.5)+($A$11)-($A$13)*15+G3/4

The new formula for UTC is simply:

    =15*(C3-12.5)+($A$11)+G3/4

which is closest to zero at UTC Hour 19, which is converted into Hour 12 in a local timezone of -7; 19 - 7 = 12.

---

I followed the explanation style of the following document. It seems the link to the error report is currently unavailable. Where is it now? I only managed to find some of its old versions like here.

SolarResourceModelsandTools/BirdClearSkyModel/README.md

Lines 10 to 30 in cfd99a9

	BEFORE 2101 GMT 08-16-2012
	Users who downloaded the BIRD CLEAR SKY model spreadsheet before 08/16/2012 should correct Column H, the HOUR ANGLE CALCULATION. Cell H3, and all others in column H, where Hour angle is calculated, needed re-arrangement of parameters in the equation in order to calculate solar noon correctly. [NOTE: the value 12.5 represents shift of solar geometry calculations to midpoint of the hour ending at hour indicated]
	The incorrect formula reads:
	=15*((C3-12.5)+(($A$13)*15-($A$11))*(4/60)+G3/60)
	The correct, simplified formula reads:
	=15*(C3-12.5)+($A$11)-($A$13)*15+G3/4
	and will produce HA results different from the previous version.
	The CORRECTED spreadsheet is available as of 8/16/2012; or users can make corrections by unprotecting the workbook and pasting the correct equation in column H.
	To correct a version downloaded before 18:00 GMT on 08/16/2012: UNPROTECT the sheet (Password BIRD); edit H3 by inserting the correct equation, and copying down the column. Then Re-protect the sheet (best to use the same password, BIRD).
	Our apologies for any inconvenience.
	Click [here](http://rredc.nrel.gov/solar/models/clearsky/error_reports.html) to see all error reports.

I submit a summary of this content to Site Feedback with the link to here, which is very brief due to space limit of 750 letters. Thank you for providing this invaluable data to people all around the world!