My guess is that Apple will do nothing.
I recall reading some university study that showed that the loss of solar cell efficiency was about five or seven percent from a soiled panel (dirt, dust, pollen, smoke particles, etc.). I'm not convinced that factor is enough to merit deploying the resources (cherry picker, water, labor) to clean the arrays.
In this part of California, precipitation is cyclical, with the first rains coming in October and stopping around May. On a per panel basis, they would see a loss of efficiency in the summer months/early autumn, but those are also peak production days with the typically clear skies. Much of the debris would wash off at the first couple of rains.
It is likely that Apple has based their electrical generation estimates on a forecast taking into consideration soiling and season, not peak production from a brand-new, immaculately clean panel array under the bright sun on summer solstice.
A similar consideration would be panel failure. It is highly unlikely that Apple will replace the panels as they fail individually. Again, I'm sure Apple has already figured in a certain amount of panel failure into their energy output calculations. Heck, with the size of this development and the number of solar panels, I wouldn't be surprised if there isn't a panel failure every other week.
A probable scenario is that Apple (or more likely its solar panel contractor) would replace defective units once a year unless a particular panel poses a danger.