Cotton lint yields in the United States increased an average of 8.45 kg/ha/yr from 1920 to 1963. However, from 1964 to 1980 yields declines at the rate of 2.24 kg/ha/yr. This decline occurred in all the major cotton-growing states and ranged from 11 69 kg/ha/yr in Mississippi to 0.37 kg/ha/yr in South Carolina. To determine the possible causes of these yield declines, the cotton simulation GOSSYM was used to study the effects of various factors on cotton production. Data sets for the simulation analysis were used from the cotton variety trials conducted at six different geographic locations: Florence, SC; Stoneville, MS; College Station, TX; Lubbock, TX; Phoenix, AZ; and Fresno, CA. The analysis of weather variability, increased atmospheric carbon dioxide and ozone concentrations, soil compaction, herbicides, and insect pest management indicated that the yield decline was caused by different factors in different regions of the U.S. cotton belt.

In the South Carolina Coastal Plains, no measurable yield decline occurred (0.37 kg/ha/yr), and most of the decline or the yield plateau was due to increased ozone levels and changes in insect pest management. In the Mississippi Delta, there were large declines in yields (11.69 kg/ha/yr). This yield decline occurred in spite of increased fertilizer inputs. The decline appeared to be due to changes in pest management strategy that resulted in increased bollworm infestation in some years. A small part of the decline in the Mississippi Delta can also be attributed to increased ozone levels.

The yield declines at the Brazos River alluvial plains and the Texas High Plains were mainly due to reduced water and nitrogen inputs. In the Brazos River area, insect problems also contributed to yield decline. For the Texas High Plains, lint yields were estimated to decrease an average of 161 kg/ha/yr when nitrogen was decreased from 45 to 33 kg/ha/yr and irrigations were reduced from a level of preplant plus one summer irrigation to a level of preplant to only irrigation.

The Salt River area of Arizona had no yield decline but a slower rate of lint yield increase due to increases in ozone and insect pressures. The cotton crops in the San Joaquin Valley experienced the highest ozone levels of any location in the cotton belt. These high ozone levels caused Most of the yield decline at this location. The other important factor that contributed to declining yields in this area was pest management

The analysis of the weather data for the yield decline period and the simulation analysis indicated that although weather caused large yearly yield fluctuations, the weather effect throughout the yield decline period was neutral. During this period, given enough nitrogen, increased atmospheric carbon dioxide would probably have resulted in a 10% increase in lint yields anywhere in the cotton belt. However, most of the variety trial locations experienced nitrogen stress during the growing season and could not fully utilize the increased carbon dioxide

Starting in 1981, average yields in the U.S. appeared to show an upward trend, probably due to a small decrease in ozone levels and better insect control in certain areas of the cotton belt. A reversal of the yield decline trend is a reasonable expectation. The development of properly focused technology is necessary to maintain or improve cotton yields with combinations of greater inputs of water and fertilizer, improved insect control, proper application of herbicides, and better control of air pollutants like ozone.