Severe bug when using sunshine duration as input

[pschmidtwalter]

There is a bad issue with the function calc_globrad() that was discovered when running a longer simulation. In run_LWFB90(), calc_globrad() derives global radiation from sunshine duration hours when it is called with options_b90$fornetrad = 'sunhours'.

The bug was discovered from suspicious annual sums of potential transpiration (output$ptran):

Global radiation, as calculated from sunshine duration hour using calc_globrad() caused this weird course of potential transpiration, it seems an annual course of global radiation is somehow projected to entire timeseries:

I will fix this as soon as possible, and try to find out in what cases simulations are concerned.

[pschmidtwalter]

When merging data.frames of extraterrestrial radiation with sunshine duration hours, ordering by date chained to the merge() had no effect. So, the timeseries was ordered by day-of-year, explaining the suspicious course of global radiation seen above. The bug concerns basicly all simulations that used sunshine duration hours as input. Now this is fixed.