GHI calibration May 2015

The multiplots indicate that there was a sensor change ~ around 2015-05-05.


Elsa's logbook denotes the following:
- 13/5 2015 CM11 til globalstråling er kalibreret til 5.065 microV/W/m2 (tidligere 4.98 microV/W/m2).

Given the information in the logbook, it is chosen to correct the first five months of the 2015 using the ,calibration coefficient from May 2015. This decision is backed up by the plot of the closure equation which shows that the GHI is too high, which is consistent with the change in calibration coefficient.

The following correction has been applied:
> "'2015-03-11 00:00","2015-05-13 17:20","GHI","factor:4.98/5.065","Then the new calibration coefficient (4.98) for the GHI pyranometer was put in place on May 13 2015, the closure equation was much better. Therefore, this coefficient is backfilled for the first five months of the year."
```python
df_qc.loc['2015-03-11 00:00': '2015-05-13 17:20', 'GHI'] = 4.98/5.065 * \
df.loc['2015-03-11 00:00': '2015-05-13 17:20', 'GHI']
```
- [ ] See if more calibration sheets can be found for this pyranometer.

Start by locating the data-processing entry point containing the shown df_qc.loc correction and review how GHI calibration records are stored. Search the available calibration sheets and logbook material for this pyranometer, then document and apply any additional corrections supported by that evidence.

Written by the indexing model from the issue text.