DOL's Bureau of Labor Statistics released employment projections for the 2012-22 decade in December 2013 (www.bls.gov/opub/mlr/2013/article/overview-of-projections-to-2022-1.htm).

BLS projected that US employment will grow from 145 million in 2012 to 161 million in 2022, with a third of the employment growth in the health care and social assistance industry. Nonfarm wage and salary employment is projected to grow from 134 million to 150 million.

The labor force is expected to increase from 155 million to 163.5 million and to be older. The median age of US workers, which was 37 in 1992 and 40 in 2002, is expected to be almost 43 in 2022, when a quarter of workers are expected to be 55 or older.

Nine of the 10 fastest-growing occupations are near the low-wage end of the wage spectrum, including personal care aides, retail salespersons, home health aides and food preparation and serving workers. One major exception to fast growth and low wages is registered nurses, where wages averaged $65,000 in 2010.

The number of agricultural workers, including self-employed farm operators, unpaid family workers, and wage and salary workers, is projected to decline from 2.1 million in 2012 to 1.9 million in 2022. Most of this decline is expected to be among farm operators and unpaid family workers, whose average employment is projected to fall from 806,000 to 607,000.

The employment of agricultural wage and salary workers, 1.3 million in 2012, is projected to fall by only 25,000 over the decade to 2022.

Mechanization. The share of workers employed in agriculture typically falls with economic growth and development. All countries with more than half of their workers employed in agriculture are classified by the World Bank as poor, and all countries with fewer than five percent of workers employed in agriculture are classified by the World Bank as rich.

Production of most field crops such as corn and soybeans, and many livestock-related operations, have been mechanized. Four commodities account for two-thirds of the labor expenditures reported by farm employers to the Census of Agriculture: fruits and nuts, 25 percent; nursery crops, 20 percent; vegetables, 12 percent; and dairy, 10 percent.

Labor-saving mechanization in fruits, vegetables and nurseries has been slowed by ample supplies of farm workers over most of the past three decades. However, the slowdown in Mexico-US migration since 2008 has prompted rising farm wages which, coupled with the falling cost of robotics, is encouraging another wave of labor-saving farm mechanization.

Nurseries. Firms such as Harvest Automation (www.harvestai.com) are developing labor-saving machines for nurseries, which must often move plants in containers. The $30,000 two-foot square and 90-pound Harvest Vehicle-100 or Harvey robot moves the 20-pound containers in which nursery plants grow so that they get enough sunlight to continue growing. An estimated one to two billion containers with plants are moved every year in US nurseries, often multiple times.

At $10 an hour, a 2,000 hour-a-year nursery worker earns $20,000 a year. Harvest Automation says that, because its robot can work almost all of the 8,760 hours a year, with down time only for recharging batteries, the HV-100 can repay its cost in three years or less.

Robots can move plant containers and put them in the desired pattern, which is often rectangular. The HV-100 picks up containers and follows sensors that tell it where to move containers. To save costs, the HV-100 does not use GPS, relying instead on sensors and boundary markers. Europe's container plant market is larger than the US market, and wages for nursery workers are often higher, so the HV-100 is expected to be sold first in Europe.

Many robots are designed for unique conditions, such as battlefields and space, where performance is more important than costs. Agriculture is different in the sense that cost is often more important than performance, meaning that robots must be simple and stand up to often harsh weather and other conditions, but can make some mistakes if the result is lower costs.

Some speculate that the success of container-moving robots such as the HV-100 might encourage farmers to experiment with food plants in containers that could be moved to achieve optimal growing conditions.

Fields and Orchards. It is much harder to develop robots to work in settings such as a field or orchard, where branches, vines and leaves create complexity. There are two broad approaches to mechanization in these more complex settings: change plants to promote uniform ripening and use once-over harvesters, or use two robots, one to "map" the fruit or vegetable and the other to harvest it.

An example of uniform ripening and mechanization is the processing tomato harvest. The California industry that today produces most of the small red tomatoes that are turned into catsup and other tomato products evolved in the 1950s in California when Mexican Bracero workers were available to pick the tomatoes. When Congress considered ending the Bracero program in the early 1960s, farmers and processors complained that, without Braceros, the processing tomato would have to move to Mexico to get harvest workers.

Instead, scientists developed a uniformly ripening tomato and engineers made a machine that cut the tomato plant, shook of the ripe tomatoes, and conveyed them to trucks that delivered 25 tons to processors. California today produces five times more processing tomatoes at a fraction of the cost of the early 1960s, and with fewer and better paid workers.

Uniform ripening and once-over machines may be possible in other crops. Most spinach and baby lettuce is cut by machine, and these machines may spread to iceberg lettuce, where three-fourths is typically harvested in the first pass through the field. In response to rising wages, most farmers harvest vegetable fields only twice rather than three or four times.

Picking fruit crops from trees is harder because of uneven ripening and the need to minimize damage to trees. Tree nuts such as almonds ripen uniformly and are harvested by a machine that grasps the trunk and shakes the nuts off the tree branches and into a skirt or on the ground for later sweeping into rows and collection.

Tree fruits such as apples and pears are more fragile and ripen less uniformly, making mechanization more difficult. Experiments that use one robot to collect data on where ripe fruit is located, and another robot to harvest the fruit identified by the first robot, are in the experimental stage.

Joseph Jones, Harvey: A Working Robot for Container Crops, October 23, 2013, Robohub. http://robohub.org/harvey-a-working-robot-for-container-crops/