1: That's far too short a time-period to be drawing any major inferences, given the level of noise. I'd be very surprised if analysis showed any statistically significant trend in that OLR data at all.
ii. The atmosphere clearly plays a much bigger role in ENSO than I'd appreciated. Look at the big dip in OLR coinciding with each big El Niño temperature spike.
(c) In the absence of increasing greenhouse gas concentrations, an increasing surface temperature SHOULD coincide with an increase in OLR. Specifically, radiated power scales with temperature to the fourth power - so if I've done the maths right a surface increase of 1 degree C should cause an increase of about 11W/m2. Anything less than that, and you know something is trapping heat.
IV: whether OLR is increasing or decreasing depends on whether the surface temperature is "catching up" to our influence (i.e. it's warming faster than the rate at which we're locking additional future warming in, so the rate of warming is slowing down) or falling further behind. As far as I'm aware, all but the most pessimistic models say it's the former, which is consistent with a very gradual increase in OLR back to equilibrium.
5: some care should be taken in looking at how OLR is calculated, and from what measurements. Remember that increasing GHG concentrations will reduce OLR only in the frequency bands at which they absorb - in other bands the radiated energy will increase with increasing temperature. It's also very highly dependent on temperature and, of course, water vapour content. So integrating both over the frequency spectrum and over the Earth's surface to get a single figure is decidedly non-trivial. Any subtle biases in sampling (e.g. dry/wet, hot/cold, day/night) could throw the result out substantially.
