More stuff from NREL Solar Radiation BMS

Using hourly data from http://www.nrel.gov/midc/srrl_bms/ gives the possibility of checking the effect of cloud cover on upward and downward long wave infrared radiation (ULWIR and DLWIR), time of observation TOBs on temperature readings.

Cloud cover effect on D/ULWIR

Data from 2004 to present (hourly)

Using data at around dawn when solar heating is minimal (would be better to have data pre dawn but cloud coverage is not measured in the dark) it should be possible to see the effect of clouds on DLWIR.

This is the plot:

The effect of clouds is most noticeable up to 25% coverage but does continue increasing up to about 80%. In this plot temperature has been constrained to 16 to 22C and RH 20 to 45% for time from 0:00 to 8:00am.
Constraints cannot easily be made tighter else total data returned fall to zero.

Time of Observation (TOBS).

Assumed:
One measurement per day max and min calculated for 23 hours prior to last measurement. Constraints on cloud cover is tricky since no information is available over night so some expected perturbation may be seen at sunrise/sunset
Within the constraints noted in the chart header, the data for each hour over the record is split into 2 - 1st quartile and 3rd quartile (this lowers the effect of false max and min values). the hourly data returned is then averaged

Data is for all Augusts on record.
The day time cloud cover effect is noticeable (0 to 10% cover - night is o to 100%).

Removing the cloud constraint gives this plot:

It looks as if TOBs could change temperature measurements by +-1.5C

A couple of cloud coverage per hour plots January and July

The early and late drop offs in coverage may be an effect of the measurement method seems a bit consistent with both plots.

However it seems that cloud coverage in jan is costant with time of day wheras in july cloud coverage increases with time.

Effect of relative humidity on DLWIR.