I read an article entitled "Peak Phosphorus" a while ago and it is referenced below. My attention was drawn back to it again by a more recent one called "Biofuels and the fertilizer problem", also referenced at the end of this posting, to the effect that there may be insufficient phosphate fertilizer to produce biofuels. Phosphorus is an essential element in all living things, from plants to you and me, along with nitrogen and potassium - known collectively as, P, N, K, in the form of micronutrients that drive growth. Global demand for phosphate rock is predicted to rise at 2.3% per year, but this is likely to increase in order to produce biomass for biofuel production.
If the transition is made to cellulosic ethanol as a fuel, because whole plants are consumed in the process, not merely the seeds etc., yet more phosphorus will be required and less of the plant (the "chaff") will be available to be returned as plant rubble after the harvest, which is a traditional and natural provider of K and P.
Now, the original Peak Phosphorus article is very interesting, if a bit doom and gloom, but only because it attests to yet another declining resource, namely phosphate rock. Similarly to the well known Hubbert Peak analysis which predicts that individual oil wells or indeed the global production of oil reaches a maximum, beyond which it declines relentlessly, a similar function can be fitted to world phosphate production. The method can be adapted in terms of the Hubbert Linearization, which I described recently (in the posting "Coal Dearth Era"), and involves plotting the annual production (P) divided by total production to date (Q), i.e. the ratio, (P/Q), against total (cumulative) production to date (Q), yielding an intercept on the x-axis which corresponds to the ultimate recoverable reserve.
The result indicates that the peak for phosphate production happened in the US in 1988 and for the world in 1989. The really telling aspect of the article is the inclusion of a plot of world oil production versus world population, for which the two quantities can be seen to follow one another closely. The conclusion is that we literally eat oil, since it underpins almost all agriculture, certainly in the developed nations, but also N and P, as required by the Green Revolution, which has preserved us from a Malthusian die-off scenario - so far, at least. Population has only grown as it has because of cheap phosphate deposits and cheap energy to produce the mineral and to get it onto farms around the world.
In contrast to fossil fuels, say, phosphorus can be recycled, but if phosphorus is wasted, there is no substitute for it. The evidence is that the world is using up its relatively limited supplies of phosphates in concentrated form. In Asia, agriculture has been enabled through returning animal and human manure to the soil, for example in the form of sewage sludge, and it is suggested that by the use of composting toilets, urine diversion, more efficient ways of using fertilizer and more efficient technology, the potential problem of phosphorus depletion might be circumvented. It all seems to add up to the same thing, that we will need to use less and more efficiently, whether that be fossil resources, or food products, including our own human waste. We are all bound on this planet and depend mutually on the various provisions of her. There are now so many of us that we will be unable to maintain current profligacy. In the form of localised communities as the global village will devolve into by the inevitable reduction in transportation, such strategies would seem sensible to food production at the local level. "Small is beautiful" as Schumacher wrote those many years ago, emphasising a system of "economics as if people mattered".
 "Biofuels and the fertilizer problem", By Tom Philpot, Energy Bulletin, February 14th, 2008. http://www.energybulletin.net/print.php?id=40300
 "Peak phosphorus", By Patrick Dery and Bart Anderson, Energy Bulletin, August 13th, 2007. http://energybulletin.net/print.php?id=33164.