Biophysical limits of crop production

I just started reading Feeding the World by Vaclav Smil and it is excellent. He's a biophysicist by training and his pragmatic, scientifically driven, top-down approach reminds me of David Mackay's in "Sustainable Energy - Without the Hot Air" (free for download and also highly recommended, particularly for anyone with an interest in renewable energy).

Brief background - Thomas Malthus famously said that agricultural production could only grow linearly and thus could not support geometric population growth (some argue he recanted later in life, but this doesn't enter the popular perception). VERY briefly, the reason humankind has been able to refute this (up until now) has been huge improvements in agricultural productivity (per hectare of land cropped). Whether this improvement can continue is a huge topic for debate.

Smil brings the top-down biophysical perspective, and says that ag productivity per hectare basically boils down to the photosynthesis rate (how fast carbon is converted to plant matter) and the harvest index (what % of the plant matter is edible). Interestingly, centuries of cultivation - from ancient breeding to modern biotech and GM crops - has not improved the photosynthesis rate AT ALL. Recent yield improvements have been driven entirely by an increase in the harvest index - for example, traditional wheat had a harvest index of 20-30% at the beginning of the 20th century, but by the 1970s certain semidwarf strains were approaching 50%. The theoretical maximum for most cereals is 65% - you can only have so much edible "phytomass" and the rest needs to go to inedible structural components.

What does this mean? There is still room for yield improvements via higher harvest indices, but there is also an upper bound not too far off. Once we hit that asymptote, our remaining options to produce more food are:

1) figure out how to improve the photosynthetic rate
2) plant more land (obviously there is an upper bound on that too, although it too is a subject of fierce debate)
3) improve yields in poorer countries (e.g. Sub-Saharan Africa) to approach those in developed countries (e.g. the U.S.)