ENVIRONMENTAL ASSESSMENT
for Management of Noxious Weeds and Hazardous Vegetation on Public Roads on National Forest System Lands

Appendix A - Risk Assessment for Humans and Non-Target Species

Introduction

The potential effects on humans and non-target species are somewhat complicated to analyze when synthetic herbicides are considered for use to control noxious weeds and invasive plants on rangelands and in forests and riparian areas. A method commonly used to analyze such effects is known as a risk assessment, which is an analytical tool that attempts to quantify the long-term risks from an action utilizing standards of safety generally accepted by the scientific and health communities. The Southwestern Region has analyzed the risk of the use of 21 herbicides and 4 carriers (USDA 1992). In addition, specific risk assessments are available for 14 of the herbicides being considered (http://www.fs.fed.us/foresthealth/pesticide/risk_assessments). A comparison of the 1992 risk assessment and the updated risk assessments indicate that the conclusions are essentially the same. All of these risk assessments are incorporated by reference and included in the project file.

“The Assessment for the Southwestern Region (1992)” displays estimated risks to the public and applicators when selected herbicides are used. In addition, estimated risks to non-target species of mammals, fish, birds, reptiles, amphibians, and invertebrates are displayed. These estimates are based on a comparison of laboratory toxicity studies with estimated exposures of representative species. The assessments display risks from “routine typical” and “routine extreme” cases. Routine typical cases represent risks to workers, the public, and other organisms that may occur as a result of routine operations. The routine extreme approach is used to estimate doses that would occur under conditions of maximum use and maximum exposure.

The risk assessment has three parts:

Exposure Analysis: This analysis estimates the range of possible doses to workers, the general public, aquatic organisms, etc. A variety of scenarios and exposure pathways are examined that could result in dermal and oral exposures.

Hazard Analysis: Tests and data related to the toxicity of herbicides are reviewed under this analysis. Data are reviewed to indicate the doses at which toxic effects occur and, conversely, levels at which no toxic effects are seen. Of particular interest is a value known as the “No Observed Effect Level” or NOEL. NOEL is the highest dose at which no adverse effects were noted in test animals. The NOEL, in combination with the various safety factors, is a partial basis for determining the safety of human doses and is useful for determining the possible noncarcinogenic effects of herbicides, such as effects on liver or kidney functions. The hazard analysis also reviews data on the possible carcinogenicity of the chemical.

Risk Analysis: Under this analysis, the dose levels calculated in the exposure analysis are compared to the NOEL levels to determine the noncarcinogenic effects of herbicides. Because NOEL levels are based on animal tests, it is assumed that the NOEL should be at least 100 times greater than the doses to establish a margin of safety (MOS). This risk analysis also indicates the probability of developing cancer based on a projection of the lifetime doses received from Forest Service spraying.

Proposed Action

The Forest Service proposes to authorize the Arizona Department of Transportation and other Public Road Authorities to treat about 5,000 acres of noxious weeds, invasive plants, and hazardous vegetation annually along public roadways that pass through the Apache-Sitgreaves, Coconino, Coronado, Kaibab, Prescott, and Tonto National Forests. The acreage of treatment each year may vary depending on infestation levels, new introductions, funding, and weather conditions. Even at the maximum level of annual treatments, only a small percentage of the rights-of-way acreage on the forests, less than 3 percent, would be treated with herbicides. All applications would be done by ground-based equipment, including low-pressure systems on pickup trucks or other vehicles, backpacks, and other hand-held devices.

Table 13 is a list of herbicides being considered for use through this document, including trade and common names. This is not an exclusive list and additional products may be registered.

Human Health Risk

A considerable body of information from tests on laboratory animals is available for the herbicides considered for possible use in controlling noxious and invasive weeds and hazardous species. Most of these tests were conducted as a requirement of the U.S. Environmental Protection Agency (EPA) for the registration process. Only those herbicides approved by the EPA will be considered for use. In addition, all of the herbicides proposed for use have been subjected to long-term feeding studies that test for general systemic effects, such as kidney and liver damage. Also, tests of the effects on reproductive and developmental toxicity (birth defects), mutagenicity (permanent transmissible change in genetic material), neurotoxicity (destructive or poisonous effect upon nerve tissue), carcinogenicity (ability or tendency to produce cancer), and immunotoxicity (poisonous to components of the entire immune system) have been conducted. NOELs are available for most types of these tests.

Extrapolating a NOEL from an animal study to humans is an uncertain process. No one can predict a safe exposure to any substance, natural or synthetic, unless the specific situation or context of exposure and dose are known. In other words, the risk or probability of harm from any substance or activity is never zero, but it can be so low as to be negligible. The EPA compensates for the uncertainty by dividing NOELs from test animals by a safety factor, typically 100, to derive a Reference Dose (RfD). Thus, the RfD is defined as the dose to humans at which there is a reasonable certainty of no harm. The factor of 100 is a risk management device that allows extrapolation of the data from animals to humans under the assumption that animals are less sensitive than humans. The factor also allows the data to be applicable to the most vulnerable members of the population — children and senior citizens. Because the NOEL is mostly based on animal lifetime exposure tests, the RfD actually represents the tolerable daily exposure over a lifetime (assumed to be 70 years for humans).

To evaluate the possible risk to humans, certain baseline criteria were set. Cancer risk is set at a benchmark value of 1 in 1 million. This benchmark is commonly accepted by the scientific community as a negligible addition to the current U.S. cancer rate. Evaluation of systemic and reproductive health risk will be based on a “no observed effect level” (NOEL), which is a long-term dose that does not result in apparent adverse effects.

In evaluating the potential impact of herbicides, it must be kept in mind the small amount that is typically used on National Forest System lands. This is normally less than 2 pounds per acre. Some products are applied at an ounce per acre.

Direct effects for workers are those that may occur from direct contact (dermal exposure) with an herbicide. Potential applications will be by backpack and ground based mechanical methods, and the area treated per day will be dependent on the specific site and type of application. It is determined that the proposed noxious weed, invasive plant, and hazardous species treatments fall within the typical scenario for herbicide use considering the proposed application rates (Table III-B-1, page III-B-3) and acres treated per day per worker (Table III-D-8, page III-D-23) in the 1992 risk assessment. It is determined that it is very unlikely that a project would include all of the conditions that exist in the routine extreme scenario (Table III-D-6, page III-D-20; Table III-B-2, page III-B-4; Table III-D-8, page III-D-23, 1992 risk assessment). The conditions of herbicide application will affect the exposure; thus, implementation of the mitigation measures and Best Management Practices, covered in Chapter 2 (page 28), will reduce possible exposures. Also, using personal protective equipment, as covered in the Safety and Spill Plan (Appendix B, page 83) will lower exposure of workers by as much as 68 percent, since most application exposure is through the skin and not through the lungs by breathing vapors (Monnig 1988). Proper training and certification of applicators on mixing, loading, and application is essential to reduce the risks to workers.

For the herbicides being considered for use, 2,4-D and triclopyr pose a moderate risk of systemic effects for backpack applicators and ground mechanical applicator/mixer loader (Table III-E-13, page III-E-17, 1992 risk assessment). In addition, 2,4-D, dicamba, and tebuthiuron have a moderate risk for reproductive effects. These risks would be mitigated by measures covered in the preceding paragraph and by limiting maximum exposure to these herbicides. Worker doses for the remaining herbicides proposed for use are likely to be well below the RfD if reasonable safety precautions are followed.

There is the possibility that workers could receive dermal exposures from the spill of a herbicide concentrate and/or the spill of a herbicide mixture, including carriers. Table III-E-14 (page III-E-18), 1992 risk assessment, for right-of-way sites, displays the risks associated with accidents (assuming a 2,000-gallon tank spill). The risk to workers associated with accidental spills is expected to be negligible if they are trained, use required protective clothing and equipment, and follow steps outlined in the Safety and Spill Plan (Appendix B).

Concern has been raised about the collection and consumption of native herbs, medicinal plants, berries, etc., that could be inadvertently sprayed. The main concern appears to center on the increased risk of cancer that could result from exposure to low levels of an herbicide. All of the herbicides being considered for use have undergone testing for cancer. Clopyralid and dicamba tests have shown no evidence of cancer initiation or promotion. The evidence for 2,4-D and picloram have been debated. Nevertheless, the 1992 risk assessment assumes that the various herbicides are carcinogens. The analyses also assume that any dose of a carcinogen could cause cancer and the probability of cancer increases with increased doses. Estimates of the probability of developing cancer from exposure to these compounds are based on a conservative extrapolation from cancer rates in animals subjected to the chemical for a lifetime. The projected cancer rates are highest for workers since their dose could be higher. Even for the workers, the risks seem relatively low compared to other commonly encountered risks. For example, one round-trip transcontinental aircraft trip carries with it an increased risk of cancer from cosmic rays in the order of one in a million. Smoking two cigarettes increases the risk of cancer by one in a million as does eating six pounds of peanut butter due to aflaxtoxin. Cancer probabilities would increase by one in a million after spraying 2,4-D for 137 days or spraying picloram for about 11,000 days. Since the average American has about a one in four chance of developing cancer in his or her lifetime, the cumulative impact from spraying herbicides at the proposed rates is considered to be insignificant. Nevertheless, studies by the California Environmental Protection Agency, Department of Pesticide Regulation, for tribal people who gather plant materials for food, medicinal, ceremonial, or basketry purposes show that herbicides were no longer detectable or plant materials were no longer available after 80 weeks (California Environmental Protection Agency, May 2001). As a result, if and when treatments are done, information on the timing and location of spraying will be provided upon request to individuals who want to avoid these areas (1-800-546-6591).

There is the possibility that a small percentage of the population in Arizona will be hypersensitive or allergic to any one or more of the herbicides proposed for use. Well-known allergenic substances include common foods, pollen, bacterial and fungal toxins, insect bites and stings, etc. Less frequent are hypersensitivities to certain fragrances and solvents. Allergies and hypersensitivities are atypical reactions exhibited by very few individuals in any population (Felsot 2001). Typical allergic symptoms include runny nose, watery eyes, swelling, and hives. Symptoms exhibited by allergic individuals are caused by specific immunological reactions of the body that are triggered by exposure to very low doses of allergens. Allergic reactions result when the body’s normal immune system defenses overproduce antibodies to specific foreign substances. Allergenic and hypersensitive reactions occur by different mechanisms than toxicity. Toxic reactions result when chemical doses become high enough to interfere with normal physiological functions of cells and tissues. Individuals who have allergic reactions or hypersensitivity are generally aware of their sensitivities and such people would not be permitted to work on spray crews. In addition, a toll free number (1-800-546-6591) is available to allow concerned members of the public to avoid the possibility of exposure from proposed herbicide applications conducted by the Arizona Department of Transportation. Other Public Road Authorities would need to be contacted directly regarding herbicide operations.

In summary, the risk or probability of harm to humans is not zero, but it is reasonable to expect that the human health impacts from the proposed herbicide applications would be insignificantly small.

Effects on Aquatic Resources

The potential impact of herbicides proposed for use on fish and other aquatic organisms is a function of three factors: 1) toxic characteristics of the active ingredient; 2) amount of the active ingredient in the water where aquatic organisms live, and 3) length of time an organism is exposed to the active ingredient.

Whether an organism is affected by an herbicide is generally measured in a laboratory using a “LC50” test. The LC50 is the herbicide concentration that is lethal to 50 percent of the organisms exposed to the active ingredient for a given time. Although the LC50 is frequently used as a toxicity standard, 50 percent mortality of fish or other aquatic organisms would not be acceptable under any circumstance on a national forest. For this reason, biologists calculate a “No Obsered Effect Level” (NOEL). This is the amount of active ingredient that would have no measurable effects on test organisms after several days of exposure.

The herbicides proposed for use are all characterized by relatively low aquatic toxicity under typical case water concentrations (Table III-H-6. page III-H-13. 1992 risk assessment). The only exceptions are for triclopyr and limonene, which may present a high risk for trout in streams and a moderate risk for trout in lakes. Picloram, dicamba, and 2,4-D may present a moderate risk under extreme water concentration, but this case seems highly unlikely under the conditions of proposed application. Clopyralid, dicamba, and glyphosate are roughly 1/5 to 1/50 as toxic to various aquatic organisms.

In regard to the risk to endangered and threatened (T&E) or sensitive aquatic organisms, triclopyr products not labeled for aquatic use may present an unacceptable risk to T&E cold water fish under the typical case scenario. Likewise, 2,4-D not labeled for aquatic use may present an unacceptable risk to T&E aquatic invertebrates. It must be noted that the assessment was made using aerial application as the treatment approach. A ground-based application would reduce the risk. Also, it does not appear that any proposed applications will occur where these organisms are present; however, to mitigate the concern, triclopyr products not labeled for aquatic use will not be sprayed within the high water zone of any stream or water course were cold water T&E or sensitive fish are present. In addition, 2,4-D products not labeled for aquatic use will not be sprayed in any location where there are T&E or sensitive aquatic invertebrate species.

The majority of herbicide applications near water will be by hand backpack or truck mounted hand wand applications, and this will result in minimal risk to contamination of surface water. Leaching of herbicides through soil is not a significant process. Herbicides do have the potential for overland flow during heavy rainstorms, but the likelihood of such movement on infiltration-dominated sites makes water contamination unlikely. Mitigation measures and Best Management Practices will serve to reduce the potential for possible adverse effects to aquatic organisms.

Non-Target Animal Species

A short list of management indicator species (MIS) for the Southwestern Region (Region 3) were identified in the 1992 risk assessment (Table III-H-10, page III-H-2) and these species can serve as general indicators for the proposed program’s effect on non-target animal species. To analyze the program’s potential risk to MIS, the various species were paired with the most closely related representative species used in the non-target species analysis. The results of the non-target species risk analysis were then extrapolated to the indicator species by assuming that the doses received by the representative species also apply to the indicator species.

There currently are 66 species listed as Management Indicator Species (MIS) that occur on National Forest System lands in Arizona. Population and habitat trends for each species can be found in MIS reports for the Apache-Sitgreaves, Coconino, Coronado, Kaibab, Prescott, and Tonto National Forests. Copies of the MIS reports are maintained at the forest supervisor’s office for each national forest in Arizona. A list of the MIS species follows:

Birds
Goshawk
Pygmy nuthatch
Merriam’s turkey
Yellow-bellied sapsucker (red-naped sapsucker)
Mexican owl
Plain titmouse (juniper titmouse)
Hairy woodpecker
Lincoln’s sparrow
Lucy’s warbler
Yellow-breasted chat
cinnamon teal
coppery-tailed trogon
sulpher-bellied flycatcher
gray hawk
blue-throated hymmingbird
rose-throated becard
thick-billed kingbird
northern beardless tyrannulet
Bell’s vireo
buff-breasted flycatcher
Mearn’s quail
Baird’s sparrow
five-spotted sparrow
peregrine falcon
Gould’s turkey
rufous-sided towhee
violet green swallow western bluebird
ash-throated flycatcher
fray vireo
Townsend’s solitaire
common flicker
black-chinned sparrow
savannah sparrow
horned lark
black-throated sparrow
brown towhee
Birds (continued)
bald eagle
summer tanager
hooded oriole
warbling vireo
western wood pewee
black hawk

Mammals
Abert’s squirrel
elk
mule deer
antelope
red squirrel
black bear
white-tailed deer
Mt Graham spruce squirrel
Arizona gray squirrel

Macro-invertebrates
Macro-invertebrates

Herptiles
desert massassauga
twin-spotted rattlesnake
Arizona ridge-nosed rattlesnake
Huachuca tiger salamander
Tarahumara frog
western barking frog
Arizona tree frog

Fish
Mexican stoneroller
Arizona trout
Gila topminnow
Gila chub
Sonora chub
Spikedace

Pages III-H-1 through III-H-9 and Table III-H-2 (page III-H-7), 1992 risk assessment, address possible effects on representative species. Under the typical case, all species are in the low risk category given the materials proposed for use. Although the 66 MIS listed above are not specifically addressed in either the 1992 risk assessment or the risk assessments for specific herbicides that are incorporated by reference (page 13), no additional information is available to assess the potential risks from the proposed use of the herbicides. Because the proposed herbicides have low toxicity to animals and any contact with herbicides would be very infrequent, it was concluded that the proposed application of the herbicides is not likely to have a greater affect on MIS than for threatened, endangered, sensitive, or other species.

An analysis of the potential effects of the proposed use of herbicides on MIS population trends and habitats was completed and included in the process record, and it was determined that the potential affects would not be significant. Linear roadways that pass through National Forest System lands in Arizona are disturbed sites that would not provide suitable habitat for MIS. Significantly, it is estimated that the majority (70 percent or more) of the proposed herbicide applications would occur within 5 feet of roadway edges. In addition, the proposed applications of herbicides would involve spot treatments and no more than 3 percent of the total area within easements could be treated. Thus, modification of habitats for individual species would be very small and the potential affect on MIS populations is expected to be negligible.