quent (Gadon 1996). Health and safety training is neither required nor uniformly available in the industry; certification requirements are highly variable. Health effects of homeowner exposure are even more difficult to measure; despite the potential risk, particularly for homeowner applicators, virtually no statistics are available to allow a thorough evaluation of the problem.
For the purposes of the report, natural ecosystems include rangelands, forests, conservation holdings, rights of way, and aquatic systems. Preservation of target species to conserve biodiversity is the main goal of management efforts; as a consequence, nontarget impacts have greater importance in dictating management practices than in many other systems. Biological control, for example, can have more nontarget effects if the biocontrol agent is insufficiently host-specific and is capable of shifting onto native hosts (Louda et al 1997, Onstad and McManus 1996). Researchers and practitioners indicated that improved weed control is necessary for all these systems. Land managers can use an array of tactics—including hand-pulling weeds, biocontrol, and chemical pesticides—to protect native flora and fauna of natural parks, wildlands, and habitats preserved for conservation. In some cases, herbicides might be selected in preference to other tools, but this decision depends on site-specific conditions. Much herbicide use involves spot treatments with backpack sprayers, and overall quantities of pesticides used are generally low—an entire national park might require less than 1,000 gallons/year (John Randall, The Nature Conservancy, August 10, 1999, personal communication). With spot treatments, low-quantity use, and selection of lower-hazard and less-mobile herbicides, human health and environmental impacts are considered low. Nonetheless, public concerns about herbicide use in private and public forests remain high because of potential effects of these chemicals on water quality, species biodiversity and habitats, and other environmental characteristics.
Even a perfunctory examination of the diverse agroecosystems contributing to the US agricultural enterprise leads inevitably to the conclusion that such diversity precludes a simple formulaic approach to specifying which chemical products, if any, will play a role in the future. Evaluating alternative management approaches or alternative manage