Every ecologist knows that in a healthy ecosystem, species need a diverse gene pool to withstand environmental stresses. Drought, disease, pests, climate etc. can all wreak havoc on a population of any organism. In a diverse genetic population, some plants will have natural resistance and will fare better than others, saving the group as a whole from catastrophe.
Thanks to this diversity, a farmer can choose a crop variety that responds well to local environmental conditions. A farmer in California may require a variety that is much different than a farmer in Wisconsin to be successful.
GMO crops have extremely narrow genetic bases. Some are developed from a single genetic line. Because of this, a single disease, pest, or environmental stressor could wipe out an entire crop of the virtually identical GMO plants. The only way to combat this vulnerability is to prop up the crop with increasingly expensive chemicals.
Under ideal conditions, or with substantial inputs, GMO crops have good production potential. But the realities of field conditions, climatic variation, pests and diseases are far from ideal. It is folly to expect the same plant to do well in many conditions.
DNA and gene interaction is a poorly understood subject. While scientists can now isolate the effects of single genes or small groups of genes, there is little understanding of how they interact with each other. Some genes have secondary effects under certain conditions. Thus artificially inserted genes can have unintended—and undesirable—side effects.
A good example is stem splitting in soybeans. For example, the stems of Roundup Ready soybeans tend to split in hot conditions. It is believed that this condition is a secondary effect of the inserted gene—perhaps the interruption of a DNA sequence that prevented the splitting. As a result, Roundup Ready beans have underperformed on yields when compared to conventional varieties in the same conditions.
This is a fledgling science, and mistakes will be made. Forecasting the implications of those mistakes is impossible given the limited understanding of what we are working with.
We have been told again and again that to feed the world we must rely upon technology, and economies of scale—that "bigger is better." This message is so ubiquitous that few actually challenge the notion—but those that do find a much different truth.
Studies comparing large mechanized farms to small farms have shown that small farms doing multiple and succession plantings were significantly more productive than the monoculture plantings used in large mechanized farms. Diversified produce/livestock farms were the most productive of all and are exactly the sustainable models that would work in developing countries to empower and feed those in need.