One of the primary responsibilities of the National Agricultural Statistics Service (NASS) in the U. S. Department of Agriculture (USDA) is to provide forecasts and estimates of current and future supplies of production from the nation's farms. Cotton is a major commodity in the U.S. and the world, therefore, a considerable effort goes into providing current and timely information.

The schedule of reports about cotton production is designed primarily to provide early season projections of the size of the current year's crop. The first such measure is the March Prospective Plantings report which provides an estimate of how many acres growers expect to plant to cotton and other crops in the upcoming crop season. This is followed in June by estimates of the number of acres actually planted to cotton in the current season. The first survey based forecast of production is made in August. These production forecasts are updated each month through January. Shortly after cotton harvest begins, biweekly reports of bales ginned are issued based on a complete census of all cotton gins. These reports continue until the ginning season is finished and are the basis for the final estimates of cotton production.

The cotton production forecasts are determined by separately estimating the acres to be harvested and then forecasting a yield per acre. State-of-the art sample surveys are conducted to first determine the acres planted to cotton. One survey is based on a sample of growers who are asked to report the acreage of cotton and other crops that they have either already planted or intend to plant. The other survey is based on a personal enumeration of producers who operate land identified in scientifically selected random sections of land.

The yield forecasts are also based on two independent surveys. First, a sample of producers is interviewed each month to obtain the pounds of lint they expect to harvest from their operation. In addition, a random sample of cotton fields is selected to be visited each month during the growing season. Within these sample fields, random plots are located and counts of plants, blooms, and bolls in their various growth stages are obtained. These fruit counts are used in statistical models to forecast the number of bolls and boll weight. The official forecasts of yield are based on a combination of the grower reported expectations and the field measurements. When harvest begins, ginning data are also used to project the final crop production.

The following sections will provide more detail about the methods used to provide forecasts and estimates of the nation's cotton production.