This NREL map of mean annual PV solar radiation indicates that, at equal latitudes across the southern U.S., the energy available for photovoltaic systems varies by something like 40% because of variations in clouds, haze, and so forth. Photosynthesis should see a similar variation.
A rough estimate of the reduction in sunlight needed to offset warming of 1°C can be made by noting that 1°C is about 1/300 of the absolute temperature. Because of the T^4 scaling of thermal radiation with temperature, this is equivalent (keeping albedo and thermal emissivity constant) to a change of about 4/300 in the input and output radiated power -- about 1.3%. From a photovoltaic power perspective, this would be equivalent to reducing PV cells from (for example) 30% to 29.6% efficiency.
From this rough calculation, it seems that weather variability is a much larger effect. Words and ideas I offer here may be used freely and without attribution.
