The met data I had was for north facing, latitude inclination i.e. 38 deg. so this is what I used. I converted Mj to kWh (/3.6) and the figures appear similar to others for this part of the world i.e. Victoria. Like California we are among the worst bush fire prone areas in the world. Our scientists have predicted due to climate change bush fires will occur more often in the future.
After investigating this a little, I think the pattern you noticed in the mean PV array output simply results from the orientation of the PV array and the cloudiness pattern.
The graph below shows the effect of the PV slope on the monthly mean PV energy output. (I created this by doing a sensitivity analysis on the slope and exporting the results into a spreadsheet.) The zero-degree line tracks exactly with the mean monthly global radiation on the horizontal surface, as we expect. As the slope increases, the seasonal pattern in the solar radiation flattens and then inverts, again as we expect.
The only slightly unusual thing about your data is that the mean energy output rises so sharply in April at moderate slopes. But I think that’s just the results of the sharp increase in clearness index in April.