| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Soil and Water Science Dep., Univ. of Florida, P.O. Box 110510, Gainesville, FL 32611
b Dep. of Animal Science, Univ. of Florida, P.O. Box 110510, Gainesville, FL 32611
c USDA/ARS, Beltsville, MD 20705
d Dep. of Animal Sci., Washington State Univ., Pullman, WA 99164
e Dep. of Agronomy, Kansas State Univ., Manhattan, KS 66506
f Montgomery Watson, Lakewood, CO 80228
* Corresponding author (gao{at}ufl.edu)
Received for publication June 15, 2000.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONSources and Uses of...Molybdenum ToxicityMolybdenum Concentrations in...RISK ASSESSMENTSUMMARY AND CONCLUSIONSREFERENCES |
|---|
land plant animal pathway (Pathway 6). Following legal challenge, most Mo numerical standards were withdrawn, pending additional field-generated data using modern biosolids (Mo concentrations <75 mg kg-1) and a reassessment of this pathway. This paper presents a reevaluation of biosolids Mo data, refinement of the risk assessment algorithms, and a reassessment of Mo-induced hypocuprosis from land application of biosolids. Forage Mo uptake coefficients (UC) are derived from field studies, many of which used modern biosolids applied to numerous soil types, with varying soil pH values, and supporting various crops. Typical cattle diet scenarios are used to calculate a diet-weighted UC value that realistically represents forage Mo exposure to cattle. Recent biosolids use data are employed to estimate the fraction of animal forage (FC) likely to be affected by biosolids applications nationally. Field data are used to estimate long-term Mo leaching and a leaching correction factor (LC) is used to adjust cumulative biosolids application limits. The modified UC and new FC and LC factors are used in a new algorithm to calculate biosolids Mo Pathway 6 risk. The resulting numerical standards for Mo are cumulative limit (RPc) = 40 kg Mo ha-1, and alternate pollutant limit (APL) = 40 mg Mo kg-1. We regard the modifications to algorithms and parameters and calculations as conservative, and believe that the risk of Mo-induced hypocuprosis from biosolids Mo is small. Providing adequate Cu mineral supplements, standard procedure in proper herd management, would augment the conservatism of the new risk assessment.
Abbreviations: BC, background concentration of pollutant in forage • FC, fraction of animal forage likely to be affected by biosolids application • HEI, highly exposed individual • LC, leaching correction factor • RF, allowable Mo increment in plant tissue • RPc, cumulative biosolids application limit • TPI, threshold pollutant intake at which a toxic effect is noted in animals consuming the forage • UC, uptake coefficient
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONSources and Uses of...Molybdenum ToxicityMolybdenum Concentrations in...RISK ASSESSMENTSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Biosolids, however, also contain trace elements whose fate must be considered in pasture and rangeland improvement programs involving land application. In 1993, the U.S. Environmental Protection Agency (USEPA) promulgated regulations (40 Code of Federal Regulations [CFR] Part 503) that, coupled with state regulations, govern biosolids recycling (USEPA, 1994). The federal rule is risk-based (to protect against reasonably anticipated adverse effects), and assesses exposure of animals, humans, and the environment to biosolids-borne metals through 14 pathways. One of the assessment pathways pertinent to biosolids use on pastures and/or rangelands (Pathway 6) evaluates metal transfer from biosolids soil plants animals. Exposure to molybdenum (Mo) via Pathway 6 is critical because ruminants (especially cattle) grazing forage containing excessive Mo can develop a Mo-induced Cu deficiency known as molybdenosis. Pathway 6 was, by far, the limiting pathway for Mo in land application programs, and calculations of allowable biosolids Mo loads to soils from the pathway were used to set numerical standards for Mo in the federal rule. The next most limiting pathway was Pathway 3 (direct human consumption of soil), and yielded a Mo limit >20-fold greater than Pathway 6. Four tables in the Part 503 rule define various types of pollutant limits: Ceiling Concentrations in Table 1, Cumulative Pollutant Loading Rates in Table 2, Pollutant Concentrations in Table 3, and Annual Pollutant Loading Rates in Table 4. Molybdenum was included in all tables, with numerical values of 75 mg Mo kg-1 (Table 1), 18 kg Mo ha-1 (Table 2), 18 mg Mo kg-1 (Table 3), and 0.9 kg Mo ha-1 yr-1 (Table 4).
|
|
|
|
This document represents the reevaluation of biosolids Mo data and risk assessment. We begin with brief reviews of Mo sources and uses and Mo toxicity. We then address recent data for Mo in biosolids and, finally, consider Mo risk assessment. The latter effort includes a review of the initial risk algorithms and their parameters, and updated databases pertinent to their use. We then offer a new algorithm, and provide data for its use to calculate Mo numerical standards.
|
Sources and Uses of Molybdenum |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONSources and Uses of...Molybdenum ToxicityMolybdenum Concentrations in...RISK ASSESSMENTSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Besides Mo in food and feces, the most common source of Mo discharged to sewer systems is from comfort cooling towers, where water is used as a recirculating cooling medium in the towers. Chemicals are added to control corrosion, mineral deposition, scaling, and bacterial and algae growth (Bastain and Brobst, 1993).
|
Molybdenum Toxicity |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONSources and Uses of...Molybdenum ToxicityMolybdenum Concentrations in...RISK ASSESSMENTSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Cattle appear to be the most susceptible species to Mo toxicity, followed by sheep. Horses grazed the teart pastures of England with no clinical signs of Mo toxicity. Differences in susceptibility among species are usually interpreted to suggest that processes in the rumen enhance the toxicity of Mo by reducing the availability of Cu. However, ruminants like mule deer and goats tolerate up to 1000 mg Mo kg-1 diet, about the same as chickens, rabbits, and rats (Ward and Nagy, 1977; Anke et al., 1985).
No clear evidence of Mo toxicity has been reported in humans (Frieberg et al., 1975), but Ward (1994) reasoned that human tolerance would be expected to be much higher than for cattle or sheep, as is the case for all nonruminant species studied.
Copper intake is the primary interaction factor in Mo toxicity because sufficient Cu supplementation can counteract almost all disorders associated with high Mo intakes (Clawson et al., 1972). Ward (1994) identified dietary factors clearly related to Mo-induced hypocuprosis as Cu intake, Cu availability, S intake, Fe intake, and the physical form of the feed.
Dietary Cu is poorly absorbed in most animal species, although absorption is greater in young than mature animals and in Cu-deficient than Cu-sufficient animals. Mature sheep absorb less than 10% of the Cu ingested (Suttle, 1973). Often, only 1 to 3% of dietary Cu is absorbed in ruminants. The Cu availability in cereal grains may be 10 times greater than in forages (Suttle, 1986). This partially explains why Cu deficiency can be a problem with grazing bovines, but usually not with dairy cattle or finishing cattle that receive greater amounts of concentrates in their diets.
Copper bioavailability in forages is greatly influenced by forage levels of S and Mo and, to a lesser degree, by forage Fe, Zn, and Cd levels. In the presence of S, high intakes of Mo can induce a Cu deficiency due to formation of insoluble Cu–Mo–S complexes (e.g., thiomolybdates) in the digestive tract that reduce the absorption of Cu (Mason, 1986, 1990). Several pathways exist by which Cu x Mo x S interactions mediate Cu deficiency (Dick, 1956; Ryan et al., 1987). Sulfur also exerts an independent effect on the availability of Cu to ruminants, and the effect of S alone may be greater than the S-dependent effects of Mo (Underwood and Suttle, 1999). Sulfides react with molybdate in the reducing medium of the rumen to replace oxygen, producing thiomolybdates. These concepts of Mo–Cu antagonism in ruminants envisage that Mo acts, not by direct interaction with Cu, but as a secondary consequence of Mo affinity for sulfide generated within the rumen. Excessive quantities of dietary S (>3 to 4 g kg-1) as sulfate or elemental sulfur may cause toxic effects and, in extreme cases, can be fatal (Kandylis, 1984). The effects of soil ingestion and Fe excess on Cu absorption (Suttle et al., 1975; Suttle et al., 1984) are believed to result from Fe binding of sulfide in the rumen, with subsequent release of sulfide in the intestine that interferes with Cu absorption.
Most clinical signs attributed to the three-way interaction are the same as those produced by simple Cu deficiency and probably arise from impaired Cu metabolism. The tolerable risk threshold of Cu to Mo ratio in feed is not fixed, but declines from 5:1 to 2:1 as pasture Mo concentrations increase from 2 to 10 mg kg-1 (Suttle, 1991). Alloway (1973) suggests that the critical Cu to Mo ratio is 4:1, whereas Miltimore and Mason (1971) suggest a narrower ratio of 2:1. Inclusion of S in the interaction is preferable to use of only Cu to Mo ratios, but this consideration has not been reduced to a field-validated formula or ratio.
For grazing bovines, the problem of Cu deficiency due to low forage Cu or a conditioned Cu deficiency (e.g., high forage Mo and/or S) is restricted to the usual six-month season for grazing of green forages. The condition is rarely seen during the feeding of stored forages in either beef or dairy cattle. Copper deficiency can be highly detrimental for cattle grazing fresh forage in some regions, but when this same forage is dried as hay, there is no Cu deficiency (Huber et al., 1971; Allaway, 1977). These authors suggested that drying forage makes Cu more available for absorption and reduces the availability of Mo.
Suttle (1980) evaluated Cu bioavailability of grazed pastures, dried grass, hay, and silage by responses in plasma Cu during repletion of hypocupremic ewes. Copper in cut hay and grass was more bioavailable than Cu in fresh grass and silage from the same field. Copper absorption in fresh grass ranged from 0.5 to 2.8% in three of the four grasses. Copper absorption was 0.9 to 1.9% for grass silage, 3.1 to 4.9% for dried grass, and 5.2 to 7.2% for hay.
Bioavailability of Cu is affected by the genetics of ruminants as well as antagonists such as Mo and S. There are marked variations within breeds in the efficiency of absorption of minerals from the diet, varying from 2 to 10% for Cu in adult sheep (Field, 1981). When different breeds of sheep grazed certain pastures in Scotland, one breed exhibited signs of Cu poisoning, whereas another showed signs of Cu deficiency (Wiener and Field, 1969). Goonerante et al. (1989) reported that Cu deficiency in Simmental cattle from Canada was more frequent than in other breeds. Feeding high levels and combinations of Cu, Mo, and/or S resulted in greatly enhanced biliary Cu excretion in Simmental versus Angus cattle.
Underwood (1981) suggested that Mo is readily and rapidly absorbed from most diets. Hexavalent water-soluble forms, sodium and ammonium molybdate, and the Mo of high-Mo herbage, most of which is water soluble, are particularly well absorbed by cattle (Ferguson et al., 1943). Absorption of Mo from the disulfide (MoS2) is poor, owing to low solubility and the antagonistic effect that S has on Mo absorption. Molybdenum absorption depends on animal species, age, and level of Mo in the diet, but the average value is 20 to 30% based on experiments involving stable and radioactive isotopes of Mo (Georgievskii et al., 1981). Molybdenum is rapidly absorbed, but very rapidly excreted, mainly in the urine and in part through the bile. High Mo forages (11 to 32 mg kg-1) grown on soils containing reclaimed mine tailings in Canada had little effect on Cu status of grazing cattle, suggesting low bioavailability of Mo (Gardner et al., 1996).
Rates of Mo absorption, retention, and excretion are inversely related to the level of dietary S. In sheep, for instance, increasing the dietary S from 1 to 3 g kg-1 in a diet supplemented with 10 mg Mo per day decreased the Mo retention from 37 to 4%. A working hypothesis for the effect of S on Mo retention is that sulfate inhibits membrane transport of molybdate, thus decreasing absorption of Mo in the intestine and decreasing reabsorption of Mo by the renal tubules (Dick, 1956; Ryan et al., 1987). For sheep with a Mo intake of 0.3 mg d-1, total body Mo decreased from 92.9 to 16.8 mg when sulfate was increased from 0.9 to 6.3 g d-1 (Dick, 1956).
|
Molybdenum Concentrations in Biosolids |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONSources and Uses of...Molybdenum ToxicityMolybdenum Concentrations in...RISK ASSESSMENTSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Numerous studies (e.g., Logan, 1997) report that concentrations of most metals in modern biosolids are now lower than in biosolids sampled for the NSSS. An exception is Mo, the concentration of which increased until the early 1990s (Logan, 1997). Historical and modern Mo data compiled by Brobst (R.B. Brobst, personal communication, 2000) suggest an effect of national regulations on biosolids Mo concentrations. Sales of molybdates decreased 
25% following promulgation of Part 503 in 1993 as publicly owned treatment plants (POTWs) attempted to reduce biosolids Mo concentrations to meet the proposed Table 3 pollutant concentration value of 18 mg kg-1. Control of Mo discharges to sanitary sewers is achieved through local limits in the POTW pretreatment program. The concentration of Mo in biosolids in many regions of the USA began to decrease quickly to the 20 mg Mo kg-1 mean (and 75% tile) concentrations seen today.
|
RISK ASSESSMENT |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONSources and Uses of...Molybdenum ToxicityMolybdenum Concentrations in...RISK ASSESSMENTSUMMARY AND CONCLUSIONSREFERENCES |
|---|
soil plants animals). The original USEPA assessment of molybdenosis risk calculated the allowable, long-term biosolids Mo concentration in soil from the algorithm (USEPA, 1992):
 |
 |
The initial risk assessment assumed: BC = 2.08 mg kg-1, TPI = 10 mg kg-1, and UC = 0.423, resulting in RPc = 18 kg ha-1.
Background Concentration (BC)
The USEPA selected the BC value of 2.08 mg Mo kg-1 from the Pierzynski and Jacobs (1986) study, but this represents a very high background concentration for forages grown in low Mo (uncontaminated) soils. Numerous recent literature citations for forages worldwide (summarized in Kabata-Pendias and Pendias, 1991; Gupta, 1997a; O'Connor and McDowell, 1999) suggest that a more typical background value for Mo is almost always <1 mg kg-1. Plant Mo concentrations vary with plant species, stage of development, plant part, soil pH, soil drainage, soil Mo loads, etc., but uncontaminated (naturally or anthropogenically), well-drained, non-peat-like soils (normal organic matter contents) rarely support plants with Mo concentrations >1 mg kg-1, and frequently result in Mo concentrations <1 mg kg-1. Legumes (e.g., alfalfa [Medicago sativa L.], clover [Trifolium pratense L.], soybean [Glycine max (L.) Merr], etc.) can accumulate much greater concentrations (2 to 40 mg Mo kg-1) under natural conditions, but the literature (e.g., Miltimore and Mason, 1971, and references cited therein) suggests wide variations in legume forage Mo contents, many <1 mg kg-1, for areas as diverse as British Columbia and Nevada. High soil pHs and elevated natural soil Mo concentrations, or high organic matter (peat), poorly drained soils that accumulate Mo from leaching (Kubota et al., 1961) typically support vegetation with >1 mg Mo kg-1. Such soils have been mapped for the USA (Kabota, 1977), or are known historically to produce high-Mo crops and to require Cu supplementation of grazing animals (Allaway, 1977). Such soils are not likely candidates for additional Mo input via biosolids, unless special precautions are taken to address Mo issues. For most soils, we believe that BC can be realistically and conservatively estimated as 1 mg Mo kg-1, and that the USEPA should adopt the lower value.
Threshold Pollutant Intake (TPI)
The threshold pollutant intake (TPI) value used initially by the USEPA was based on guidance from National Research Council (1980), and was taken as 10 mg Mo kg-1 animal diet. Ruminants are by far the most susceptible herbivore to Mo-induced Cu deficiency, and cattle are the most sensitive ruminants (Ward, 1994; Suttle, 1991). Most references (Ward, 1994; Suttle, 1991; McDowell, 1997; Gupta, 1997a) report TPI values that range from a few (2 to 5) to tens (10 to 50) of mg kg-1, and stress the importance of other issues (e.g., forage Cu to Mo ratios, forage S content, age and condition of the forage, degree of mineral supplementation of the animals, etc.) as significant complicating factors. Unfortunately, attempts to predict Cu–Mo–S–other elemental interactions under field conditions have been largely unsuccessful (Suttle, 1991).
The expert committee report of the National Academy of Sciences (National Research Council, 1980) evaluated low-level, chronic Mo toxicity, and identified "5 to 10 mg Mo kg-1, which has been weakly associated with impaired bone development in young horses and cattle" as the critical level. However, substantially higher levels of Mo are tolerated in the presence of adequate Cu and inorganic sulfate. Forages grown on biosolids-amended soils (high cumulative biosolids loads) are typically normal (nondeficient) in Cu and sulfate, so the higher permissible Mo concentration recommendation of 10 mg Mo kg-1 seems appropriate. A large body of data on the toxicity of Mo in forages grown on soils naturally high in Mo (not Mo salt additions to animal diets) supports the use of 10 mg Mo kg-1 for forages with normal Cu concentrations in the diet (National Research Council, 1980 [Table 1]; Suttle, 1991).
The TPI used by the USEPA represents the best available data for domestic animals, and may be considered a lowest observed effect level (LOEL) rather than a no observed adverse effect level (NOAEL). This distinction is important, as other pathway analyses in Part 503 relied on NOAEL values (when available) to build conservatism into the risk assessment. Given the complexity of Cu–Mo–S interactions, however, focusing exclusively on Mo without reasonable and practical considerations of Cu and sulfate inputs in forage grown on biosolids-amended soil and from normal mineral supplements provided to grazing livestock seems unnecessarily restrictive. Ward (1994) concluded that it was nearly impossible to identify the equivalent of a NOAEL for Mo, as the interactions surrounding Mo toxicity are too great to establish the lower limit. He summarized numerous data sets to conclude that 100 mg Mo kg-1 is definitely toxic (molybdenosis, e.g., rapid scouring) to cattle, 25 to 50 mg kg-1 gives mixed results (sometimes no effect), and that the Mo effects attributed to feeds with <25 mg Mo kg-1 are often associated with very low, and poorly available, Cu. We conclude that using the higher end of the NRC critical range (10 mg Mo kg-1) as the TPI is justified.
The TPI is derived for domestic animals, as there are few data for wild (nondomesticated) animals. Mule deer reportedly tolerate up to 1000 mg Mo kg-1 in their diets (Ward and Nagy, 1977). Flynn et al. (1977) hypothesized a possible Cu x Mo x S interaction to explain abnormal hoof material in moose, but acknowledged that their overall data were consistent with the conclusions of Kubota (1974) and others that "available information does not suggest an existence of nutritional problems in moose due to imbalances of Mo and Cu in feed plants." Domestic cattle may not be the most sensitive animal to Mo toxicity, but there are no data to justify selecting another species for use in the risk assessment. Further, large, nondomesticated ruminant species typically graze large areas, and could be expected to receive much less biosolids Mo exposure than domestic species confined to biosolids-amended pastures.
Highly Exposed Individual
The risk assessment in Pathway 6 seeks to protect a highly exposed individual (HEI), defined as "the most sensitive/most exposed herbivorous livestock that consumes plants grown on biosolids-amended soil. It is assumed that 100% of the livestock diet consists of forage grown on sewage sludge–amended land, and that the animal is exposed to a background pollutant intake" (USEPA, 1994). Grazing ruminants may, indeed, be limited to forage growing on biosolids-amended land, but dietary intake of common mineral supplements containing Cu (that can obviate low Cu to Mo ratios of forage) is ignored in the risk assessment. Further, confined ruminants (finishing beef cattle or dairy cows) are usually fed a variety of forages, grains, and mineral supplements to maximize performance. How reasonable is it to assume that 100% of the ruminant diet is one kind of forage grown exclusively on biosolids-amended land? What mitigating effects do supplemental Cu and possible increased forage Cu (as a result of biosolids) to the diet have on the risk of Mo-induced hypocuprosis? We examine the diet assumptions used to assess Mo toxicity risk to ruminant livestock below.
Dietary Assumptions
Those livestock classified as ruminants include cattle, sheep, and goats. Given the greater sensitivity of cattle to Mo-induced hypocuprosis, we focus on cattle. Cattle can be separated into beef and dairy because of the differences in their feeding management. Risk assessment for sheep and goats can be considered similar to either dairy or beef cattle, depending upon their relative feeding management.
Dairy Cattle
Feeding management of dairy cattle can be divided into calves (0 to 3 mo), young heifers (3 to 6 mo), growing heifers (6 mo to breeding), pregnant heifers (15 to 20 mo), lactating cows (305 d), and nonlactating cows (60 d). Calves typically are born with high concentrations of Cu (>300 mg Cu kg-1 of dry matter) in their liver (Underwood, 1981) and these concentrations decline over several months to those typical of adult ruminants (Kincaid et al., 1986). Concentrates comprise most of the dry feed for calves and young heifers; thus, Mo-induced hypocuprosis is unlikely to occur in calves and heifers because of the endogenous Cu reserves, use of concentrates containing Cu supplements, and restricted intakes of fresh forages in diets of these animals. Growing heifers (6 mo to breeding) are typically fed concentrates along with their forage. Once heifers are pregnant, they are usually fed only forage and mineral supplements until 2 to 4 weeks prepartum. Lactating dairy cows are fed diets containing between 40 and 60% forage, and 60 to 40% concentrates (Ensminger et al., 1990). Concentrates for lactating cows consist of about 50% grain (e.g., corn [Zea mays L.] or barley [Hordeum vulgare L.]), 20% by-product feeds (e.g., whole cottonseeds, beet pulp, and wheat mill run), and 20% protein supplement (e.g., soybean meal), with the remainder (10%) being molasses, sodium bicarbonate, mineral supplements, vitamin supplements, and various other ingredients.
The forage portion of dairy cow diets varies among regions of the country (Mowrey and Spain, 1999). Often, the forage consists of near equal proportions (dry matter basis) of hay and silage during the nongrazing seasons, to as much as 100% fresh forage during the 4- to 6-month grazing season. The fresh forage can consist entirely of grasses and legumes, although mixed pastures are most common (Etgen et al., 1987). Regardless of the pasture forage species, the entire ration of the dairy cow rarely consists of more than 60% fresh legumes because of the need to incorporate other feed ingredients into diets to maximize milk production. Most large herds of dairy cattle in the USA do not graze pastures, and their diets remain fairly constant during all seasons of the year. Nonlactating (dry) dairy cows are fed roughages until 14 d prepartum, when limited amounts of concentrates are introduced into the diet to prepare the cows for the lactation ration fed after parturition. Nonlactating dairy cattle normally consume forages consisting of grasses or legume–grass mixes, but are not fed 100% fresh legumes because of possible health problems associated with excessive intakes of protein, energy, potassium, calcium, and increased incidence of bloat (Etgen et al., 1987; Ensminger et al., 1990). During the nonlactating period, cows are provided mineral supplements.
Beef Cattle
Beef cattle production can be divided into the cow–calf operation, stockers, and feedlot cattle. Because feedlot cattle are fed high-concentrate diets and slaughtered after about 120 d on feed, they are not considered at risk for Mo-induced hypocuprosis. Beef cattle are fed conserved roughages (hays, silages, and crop aftermath) during the nongrazing seasons. During the growing season, pastures for beef cows are dominated by grasses, although variations exist among regions in the USA. Legumes (alfalfa and clovers) are often incorporated into pastures, but they rarely constitute the entire ration of beef cows for extended periods. Mixed plant species in a pasture assist in lengthening the grazing periods and allow for plants that are best suited to the soil conditions within a pasture (Etgen et al., 1987). Young beef calves consume some fresh forage, and the amounts progressively increase as the calf grows and the milk production of the dam declines. Weaned beef calves (age > 259 d) may be fed only roughage until they enter the feedlot for fattening. The diets of these weaned calves (stockers) can include crop residues (corn stalks, cereal grain straws, and stubble), winter wheat (Triticum aestivum L.) (or other small grain) pastures, irrigated pastures, silages, and hay (Ensminger et al., 1990).
The cattle groups perhaps at greatest risk of Mo-induced Cu deficiency are beef cows, growing beef calves, and pregnant heifers because of the dominance of fresh forages in their diets. Attention to the concentrations of Mo, Cu, and S in their diets is needed and especially to the trace mineral supplementation programs provided for these animals when Mo problems are expected.
The risk of Mo-induced Cu deficiency is greatest during the period of active growing forage, which is only 5 to 6 months in many areas of the country. Molybdenum-induced Cu deficiency is not a problem for ruminants receiving stored forages, apparently because of increased availability of Cu and reduced availability of Mo in these feeds (Underwood and Suttle, 1999). Most important, Mo-induced Cu deficiency regions in the USA are well known and farmers have learned to compensate by providing Cu in mineral supplements (Allaway, 1977).
The use of a single TPI value of 10 mg Mo kg-1 is an oversimplification of general animal response to forage Mo exposure. Nevertheless, we believe that the TPI value chosen is reasonable for a national risk assessment for biosolids Mo when practical aspects of animal (HEI) management (e.g., mineral supplementation and total animal diet considerations) are recognized.
Allowable Molybdenum Increment (RF)
Based on the above discussion, the allowable Mo increment in plant tissue (RF) is:
 |
To complete the risk assessment using the USEPA's original algorithm, RF is divided by UC to arrive at RPc.
Uptake Coefficient (UC)
Problems associated with appropriate selection and/or calculation of UC values are many. Analytical problems with Mo (e.g., falsely high Mo analysis due to Fe and Al interference and high limits of detection with some methods; McBride et al., 2000) probably confound interpretation of even some modern literature. Further, many plant Mo uptake studies were conducted in the greenhouse, and probably suffered greenhouse effect errors (higher uptake slopes for contaminants compared with those found in the field at similar contaminant loads) discussed in the Part 503 Technical Support Document (USEPA, 1992).
Ideally, the UC value is obtained from field experiments where ranges of soil Mo loadings are evaluated. In this case, UC is calculated as the slope of the linear regression of plant tissue Mo concentrations (mg kg-1) versus biosolids Mo application (kg ha-1). The approach implies that the plant tissue Mo response to soil Mo load is linear and that the applied Mo remains in the root zone indefinitely. The linear model was assumed by the USEPA in the initial risk assessment for all metal uptake calculations, but data in the Part 503 Technical Support Document, and more recent data (e.g., Barbarick et al., 1995) suggest a plateau model as more appropriate for at least some metals or crops. Limited data presented by O'Connor and McDowell (1999) and Nguyen (1998) suggest that the plateau model may also be appropriate for biosolids Mo applied to a pasture grass. Using a linear model when a plateau model is appropriate overestimates crop Mo (and risk) at higher biosolids application rates and over extended periods of application. Given the incompleteness of the database, a linear model is assumed herein.
Soil Molybdenum Load
Obtaining the plant Mo concentrations for the UC calculations is relatively straightforward, but determining the soil Mo loading is less so. The most desirable situation is to have measured total soil Mo concentrations from a sampling time close to when the plants were grown. These concentrations can then be converted to soil Mo loadings (applied from biosolids) by multiplying by the appropriate factor, considering sampling depth and bulk density, and the background Mo concentration in similar nonamended soil. For a sampling depth of 15 cm and a bulk density of approximately 1.3 Mg m-3, one multiplies soil concentration by two to obtain soil Mo loading in kg ha-1. This approach is consistent with the assumption in the risk assessment that Mo is not a conservative element in the soil. That is, if some time has passed since the biosolids Mo was applied, leaching may have reduced the soil Mo concentration to levels lower than those predicted from the original Mo loading rate (biosolids rate x biosolids Mo concentration). Unfortunately, few published studies report total soil Mo concentrations for each of multiple cropping seasons, and the original Mo loading rate must be used in the UC calculation. In most studies, soil sampling and analysis is restricted to the 0- to 15-cm depth, and ignores possible contribution of deeper soil profile Mo to plant uptake. Thus, even if surface soil sampling occurs each year of a multiyear study, soil Mo accessible to deep-rooted plants may be underestimated. Another shortcoming of many studies is the minimal range or number of soil Mo loadings studied. Under these conditions, single-point estimates of UC must be made. This is done by subtracting the Mo concentration found in the control (no added Mo) plants from the concentration in plants grown on the biosolids-amended soil, and then dividing by the biosolids Mo load.
Variations in Uptake Coefficient with Time and Plant Part
Using point estimates of UC values is appropriate for whole plant estimates of Mo exposure when, for example, corn silage is fed, but not when perennial forages are consumed. In such cases, UC values change monthly (and widely) over an entire grazing season (e.g., Ferguson et al., 1943; Nguyen, 1998), with plant part actually consumed (grains vs. stover, e.g., Gupta, 1997b), and with forage condition (fresh vs. dried hay, e.g., Mills and Davis, 1987). O'Connor and McDowell (1999) recommended calculating (yield) weighted average Mo concentrations of grass grazed for 6 mo by cattle. The weighted average concentration was thought to more accurately represent the forage Mo concentration the cattle experienced throughout the growing season than simply calculating the average concentration of several cuttings. The weighted average Mo concentration was also thought to more reasonably represent the animal forage consumption (Mo exposure) that varied with grass yield throughout the grazing season.
Variations among Plant Species
Variations in plant species Mo concentrations—at the same soil Mo load—are widely acknowledged (e.g., Gupta, 1997b; Vlek and Lindsay, 1977; Kabata-Pendias and Pendias, 1991; Johansen et al., 1997) and, in fact, form the basis for management of cattle Mo intake on (known) high-Mo areas. Ranchers are advised to plant nonaccumulating grasses or grains, rather than accumulators such as legumes (e.g., alfalfa, clover, and soybean). Miltimore and Mason (1971), however, cite several data sets that show little difference in Mo concentrations in grass, grass–legume, legume, or corn silage cattle diets grown in areas as diverse as British Columbia, Canada, Kansas, Virginia, and Nevada. Soil pH, wetness, Mo concentration, and climate can all apparently play more important roles in determining forage Mo concentration than forage species alone. Nevertheless, legumes are particularly susceptible to excessive Mo accumulation (high uptake slopes), and databases used to assess Mo toxicity should include legume UC values.
Variations with Soil Properties
Variations in plant-available Mo with soil pH are well known (e.g., Gupta, 1997a,b; Kabata-Pendias and Pendias, 1991; Pierzynski and Jacobs, 1986; Williams and Gogna, 1981). High soil pH favors low Mo sorption and high plant availability. Molybdenum adsorption envelopes typically exhibit maximal Mo sorption at soil pH values <5, so near-neutral and calcareous soils tend to retain little Mo (e.g., Goldberg et al., 1996, 1998). Most molybdenosis problems occur on such high-pH soils (Allaway, 1977).
Variations in soil moisture and soil organic matter contents affect plant Mo concentrations. Kubota et al. (1963) showed that the same soil Mo concentration resulted in vastly different plant Mo concentrations, depending upon the soil moisture regime. Ready supply of Mo in moist (poorly drained) soils is well recognized (Kubota, 1977), and is one reason given for the abundant Mo in plants growing in wet, high-organic soils (e.g., peats) even at relatively low soil Mo concentrations (Allaway, 1977). In "normal" soils (well-drained, <50 g kg-1 soil organic matter), pH is usually the dominant factor determining Mo phytoavailability. Soils with pH values >6 frequently result in forage with excessive (10 to 20 mg kg-1) Mo concentrations if soil Mo concentrations are significantly above background levels (1 to 2 mg Mo kg-1) (Gupta, 1997a; Kabata-Pendias and Pendias, 1991; Barber, 1984; Allaway, 1977).
Biosolids Source Effects
Biosolids source effects, based on excessively high total biosolids Mo concentrations (e.g., Pierzynski and Jacobs, 1986), or other constituents (Fe and Al) in the biosolids (e.g., Soon and Bates, 1985; O'Connor and McDowell, 1999) can also lead to vastly different UC estimates. Such a biosolids effect was not directly considered in the original USEPA risk assessment, although the Technical Support Document (USEPA, 1992) includes reference to data of Soon and Bates (1985) showing lower Mo phytoavailability on soil amended with Fe or Al biosolids compared with Ca biosolids amendment. Ferric chloride and/or alum are routinely added in waste treatment processes to improve P removal, and would be expected to be similarly effective at immobilizing Mo. Lime-stabilized biosolids can raise soil pH and increase uptake slopes correspondingly (Soon and Bates, 1985; McBride, 1998).
The original USEPA risk assessment included the very high (1500 mg Mo kg-1) biosolids data of Pierzynski and Jacobs (1986), but data for such a biosolids (Mo concentration 20-fold greater than the 98th percentile of national biosolids) represent unique conditions. We believe that the USEPA should omit the Pierzynski and Jacobs (1986)-derived UC values when better data are available.
The ideal data set for purposes of risk assessment would be from field studies in which Mo was added at various rates from biosolids containing less than 75 mg Mo kg-1 (maximum biosolids Mo concentration allowed for land application). In addition, having UC values for all types of forages typically consumed by animals in the exposure pathway of concern would be desirable. Data for corn silage, stover, or grain would be preferable to corn leaf diagnostic tissue data, for example. As previously mentioned, total soil Mo concentrations that could be used to calculate Mo loadings would be preferable to calculated Mo loadings based on biosolids composition and application rates. Few studies satisfy all these criteria, and the selection of appropriate data sets frequently requires use of professional judgement. One essential requirement is that the UC values be generated from studies in which biosolids were the Mo source. Data from greenhouse pot studies would only be used if no other data were available for a particular forage type.
The USEPA attempted to resolve the UC value variation issues by calculating a geometric mean of UC estimates from a limited set of field- and greenhouse-generated data, where biosolids were the sources of Mo load. The original Mo data set used in the 1993 rule development was quite limited, but a few significant studies (Basta et al., 1999; O'Connor and McDowell, 1999; McBride et al., 2000; O'Connor et al., 2001a) that meet most of the criteria described above are now available. The data are included in Table 2, along with selected data from Tables C-39, C-40, and C-41 of the Technical Support Document (USEPA, 1992). Still underrepresented in the table are data from field studies where legumes are grown in biosolids-amended soils of high pH, and where the biosolids have less than the ceiling Mo concentration. A paper (O'Connor et al., 2001b) describing one such set of soybean grain data has been published; data are included in Table 2.
Uptake coefficients (UC values) of nonlegumes are low and almost always <0.5 (Table 2). The lone exception (UC = 4.6) in the database comes from a study involving a pasture grass grown in soil amended with one biosolids that was inadvertently overlimed in the second year of a 2-yr study (Nguyen, 1998). Another biosolids (slightly higher Fe and Al content) applied to the same soil yielded a more representative UC value for the grass of 0.39. The arithmetic average of the UC values for 29 nonlegume entries in Table 2 is 0.24. We conclude that a UC value for nonlegumes (vegetation or grain) is conservatively estimated as 0.5.
Field data for estimating UC values for legumes are limited, so greenhouse data (Pierzynski and Jacobs, 1986; McBride et al., 2000) are included in the database (Table 2). The arithmetic average of UC values for 24 legume entries is 2.2. Given the limited diversity of studies, we conclude that a UC value for legumes is conservatively estimated as 4.0. No adjustment of the UC value is made for the well-recognized decreased Mo availability to cattle in dried (vs. fresh) forages (Mills and Davis, 1987).
Based on the dietary considerations presented previously, we estimate that grazed fresh legumes will typically constitute no more than 50% of beef cattle or dairy cow animal diets, and frequently less. The remainder of the diets would typically consist of hay silage, roughage (grain by-products), and grains, plus mineral supplements. If an average UC value for fresh legumes is taken conservatively as 4, and an average UC value for all other plant-feedstuffs is taken very conservatively as 0.5, a diet-weighted UC value of [4(0.5)] + [0.5(0.5)] = 2.25 is obtained. We believe that such an UC value represents a worst-case estimate of typical forage Mo exposure, and note that it ignores normal mineral (Cu) supplementation recommended for good animal management.
Algorithm Modifications
Fraction of Diet Affected (FC)
Not all feed consumed by cattle is grown in the same place, nor is all the land used to grow animal feed likely to be biosolids-amended. It is particularly unlikely that the feed-producing land would be amended every year for a total application rate of 1000 Mg ha-1 (10 Mg ha-1 yr-1 x 100 yr), as was assumed in the risk assessment (USEPA, 1992). More likely, the feed-producing land would be amended at 5 to 10 Mg ha-1 (N-based application rates) only periodically (e.g., once every 3 to 5 yr). We examined biosolids production and use on crop and pasture land in each state to estimate the fraction of cattle diet (FC) likely to be affected by biosolids use.
Biosolids production is likely to remain the same or increase only slightly over the next several years. Wastewater treatment plant construction has slowed, and increases in biosolids production are related to improvements in existing treatment rather than new construction (Bastain, 1997). Estimates of biosolids production and final use and disposal were surveyed in 1996; the totals for the USA at that time were 6.8 x 106 dry Mg produced, with 54% of the total being land-applied (Bastain, 1997). Table 3 contains a summary of the quantities of biosolids produced and land-applied in each state.
Column B (Table 3) represents total acreage in farms and ranches held as private lands in 1997 (USDA, 1999). Column C represents land in crop rotation that, during the year of the survey, was in pasture for at least part of the year. Reviews of past USDA agricultural surveys suggest that acreages in this category differ little from year to year. Pasture and rangeland acreage (Column D) represents lands remaining in pasture for several years; generally considered permanent pasture. Past USDA agricultural surveys (back to 1987) suggest similar acreages. Column E values are the total acreages in pasture (Columns C + D), but do not include grazing lands leased from government agencies. Addition of these public lands would significantly increase the acreage, particularly in the 11 western states. Data for biosolids produced in each state (Column F) were summarized by Bastain (1997). The values do not account for intrastate transfer of biosolids, such as occurs from areas with small land bases to areas with large land bases. A biosolids load of 5 Mg ha-1 was chosen as a typical application rate to pasture land, and was used to calculate the acreage required to accommodate the load (Column G).
Column H represents land required for biosolids application (Column G) as a percentage of total land in pasture (Column E). The calculation included all biosolids produced in each state, regardless of use and/or disposal practice (e.g., land-applied, landfilled, or incinerated), and all biosolids were assumed to be applied to pastures, a conservative assumption. Brobst (R.B. Brobst, personal communication, 2000) conducted an informal survey of contemporary biosolids use in 15 states, from Georgia to Oregon and from Minnesota to Arizona. Twelve of the states applied <1% of their biosolids to pastures; Oklahoma and Georgia applied <10%, and Virginia applied 33%. Agricultural producers tended to use the nutrients in biosolids on croplands (rather than pastures) whenever possible to reduce production costs associated with commercial fertilizers.
The land application weighting factor (Column I) estimates the possible increases in biosolids production and the resulting fraction of land affected by land application. Column I values were derived by multiplying the percent of biosolids currently land-applied by 1.5, up to a maximum of 100% of the biosolids produced. Column J is the product of Column H and Column I, and represents a conservative estimate of pasture land (cattle feed) likely to be affected by biosolids (FC). Twenty-four states have FC values <1%, 15 have values between 1 and 10%, 5 between 10 and 20%, and 6 have FC values >20%. None of the six states (Pennsylvania, New Jersey, Connecticut, Delaware, Maryland, and Massachusetts) was among the top 28 beef-producing states in 1997 (USDA, 1999). We conclude that a reasonable estimate of the fraction of ruminant forage likely to be affected by biosolids (FC) is 0.1 to 0.2, and recommend using the more conservative value of FC = 0.2 for a national approach to rule making. The FC factor we propose may not be universally applicable, but is consistent with USEPA policy that estimates the fraction of HEI diet probably affected by food grown on biosolids-amended land in Pathways 1, 2, 4, and 5 (USEPA, 1992). Local conditions may warrant using a different FC, which would allow site-specific considerations in a biosolids land application program.
Leaching Correction (LC)
Maximum Mo loading considerations ignore the fact that Mo is not a conserved element when biosolids are applied to the same field for many years, especially to soils having high soil pH where Mo leaching, bioavailability, and risk are all maximized. Few studies have been conducted with the primary objective of directly characterizing Mo leaching dynamics and the parameters affecting leaching rates. Nonetheless, several studies provide indirect evidence of Mo leaching from biosolids and industrial Mo sources. These studies are summarized below, and demonstrate that Mo leaching from the plow layer occurs, sometimes at high rates. Additional studies of Mo leaching from biosolids sources are recommended to develop a more complete data set and to validate the approach used here.
Molybdenum is adsorbed by acidic soils, and exhibits maximal retention at soil pH values <5 (Goldberg et al., 1996, 1998). O'Connor and McDowell (1999) presented data for Mo sorption on biosolids-amended soils that confirm Mo behavior noted by Goldberg for unamended soils. Their data also confirm the dramatic reduction in Mo retention as soil pH increases, with sorption becoming negligible above pH 6. Thus, Mo availability to plants (and to leaching) increases dramatically at higher pH.
Artiola (J.F. Artiola, personal communication, 1999) conducted column studies with calcareous soils (pH = 7.8) amended with biosolids Mo. The biosolids application rate was about 200 Mg ha-1, and the New York biosolids Mo concentration was 6 g Mg-1 (Mo load = 1.2 kg Mo ha-1). Molybdenum moved through the columns essentially unimpeded, and reached low relative concentrations (C/Ci = 0.1) in drainage after only about two pore volumes of drainage. There was little difference in Mo leaching behavior whether the Mo was added as biosolids or Mo salts under such high-pH (7.8) conditions. O'Connor and McDowell (1999) and Brinton and O'Connor (2000) reported data that suggested different effects of biosolids source on Mo retention and/or release depending upon biosolids total Fe and Al contents, but soil pH effects dominated Mo behavior in two soils amended with the various biosolids. High soil pH reduces soil retardation of Mo movement and increases the likelihood that applied Mo will leach away from the zone of biosolids incorporation. The effect would be greatest under irrigated conditions, where net downward flow of water is maintained (positive leaching fraction) to control salinity. Chang and Page (2000) recently conducted an input–output balance of trace elements in soils of the San Joaquin Valley's West Side. Numerous inputs were considered, including Mo from biosolids additions. Despite the inputs from various sources, there was a net depletion of Mo from cropland on the West Side (Chang and Page, 2000). Amounts of Mo dissolved from soils and transported to tile drains exceeded amounts of Mo added from all input sources combined. Leaching of Mo from these near-neutral soils accounts for the net depletion of Mo, and is inevitable for sustainable irrigated crop production (Chang and Page, 2000).
Phillips and Meyer (1993) studied alfalfa growing in calcareous (soil pH > 7) fields with natively high soil Mo concentrations in Kern County, CA. An earlier (1950) survey of alfalfa from the county had identified average alfalfa tissue Mo concentrations of about 10 mg kg-1. Resampling the same fields 35 yr later revealed average alfalfa tissue Mo concentrations of about 3.5 mg kg-1. No quantification of soil Mo concentrations was attempted, but Phillips and Meyer concluded that the lowered tissue Mo concentrations were associated primarily with leaching of soluble salts (including Mo). Similarly, McBride et al. (1999) reported that nearly 80% of biosolids Mo applied at high rates to a calcareous silty clay loam (pH 7.0 to 7.3) had been lost from the topsoil under natural rainfall conditions over the nearly 20 yr since the biosolids were applied.
Leaching of Mo is not restricted to high-pH soils. Hemkes et al. (1980) applied moderate rates (6 to 18 Mg ha-1) of high-Mo biosolids (117 to 170 g Mo Mg-1) to a pH 5.9 sandy, permanent pasture soil in the Netherlands each year for 3 yr (total Mo loads = 2.1 to 9.2 kg Mo ha-1). Molybdenum concentrations in the surface (0–2.5 cm) soil reached 7.4 mg kg-1 at the greatest biosolids Mo load, and decreased with depth to background concentrations at >25 cm. They were unable to account for all of the Mo added, and hypothesized Mo leaching. Nguyen (1998) noted about 50% loss of biosolids Mo from the surface 0- to 15-cm depth of a limed acid (pH 5.5–6.9) Spodosol 3 to 4 yr after biosolids application. Thus, even slightly acid soils under high-rainfall conditions can be expected to allow leaching of biosolids Mo. Leaching would be promoted if the soils were limed or amended with high-pH (alkaline-treated) biosolids (McBride, 1998).
Leaching of Mo under nonirrigated, semiarid, or arid conditions is often difficult to predict. Infiltration below the root zone occurs when the magnitude and duration of moisture events (e.g., snow melt, rainfall) exceed evapotranspiration, runoff, soil moisture retention, and absorption by surface materials (e.g., biosolids). McCurry (1995) summarized infiltration data collected in several states. Data represented infiltration studies on biosolids projects in Colorado and New Mexico and basic infiltration studies conducted in New Mexico, Idaho, California, Texas, Oklahoma, and Utah. Infiltration data from the biosolids studies were consistent with data collected for nonbiosolids studies. Moisture infiltrated 40 to >270 cm depending upon location and type of excess moisture event (McCurry, 1995). Excess moisture in arid and semiarid regions is usually only available in the spring and early summer, following snow melt, and/or for briefs periods following intense summer thunderstorms. McCurry's data demonstrate that water (and dissolved ions) can move below the root zone during these periods. Additional measurements of solute movement through biosolids-amended soils under semiarid and arid conditions have been made (Janonis et al., 1996; Michalk, 1995; Moffet et al., 1995). Molybdenum was not specifically monitored in either study, but is expected to move as readily as other anions (e.g., PO4, and SO4) whose migration was detected. Thus, leaching of Mo can be expected in all but the driest of environments, and can be expected to be significant under irrigated or high-rainfall conditions. Under dryland conditions, salt accumulation typically limits biosolids application rather than N, P, or Mo loads.
Leaching loss of Mo is expected to be greatest under high soil pH conditions, where Mo bioavailability is problematic. Not accounting for Mo losses to leaching, even over the long periods (>100 yr) needed to attain limiting cumulative soil Mo loads, is unrealistic.
Pollutant leaching can be characterized using rigorous quantitative approaches or more qualitative approaches. The USEPA, for example, used two transport models (quantitative approach) to predict ground water effects of land-applied biosolids-pollutants in the Pathway 12 and 14 (drinking water) risk assessment. The VADOF subroutine of the RUSTIC model (USEPA, 1989) was used to estimate flow and pollutant transport through the unsaturated zone. A second model (AT123D; Yeh, 1981) was then used to estimate transport in the saturated zone (aquifer). The combined model outputs estimated pollutant transport to a receptor, a drinking water supply well. Results of the simulations were used to back-calculate reference pollutant application rates that result in threshold ground water effects at the receptor (USEPA, 1994). This rigorously quantitative approach is data- and assumption-intensive, and is deemed inappropriate for Pathway 6 risk assessment for two reasons. First, the receptor of interest in Pathway 6 (the most sensitive pathway of risk for Mo) is a ruminant grazing biosolids-amended forage, not a human drinking water from a downstream well. Second, model parameters were set to maximize pollutant leaching in Pathways 12 and 14 (USEPA, 1992), so risk in Pathway 6 is simultaneously underestimated (less Mo remaining in the soil for plant uptake). For these reasons, we chose a simplified, but conservative, model that uses a first-order leaching algorithm and empirical data for pollutant leaching characterization. The algorithm describes leaching that annually removes Mo from the root zone in proportion to the total Mo concentration in the soil (Eq. [1]):
 | [1] |
The model assumes relatively constant climatic conditions over, for example, the maximum assumed 100-yr biosolids application period. Further, one assumes a constant pollutant addition rate and a constant (average) pollutant leaching rate (k). If no leaching is assumed (k = 0), Mo accumulates in the soil in direct proportion to the addition rate. When k > 0, the model projects that the quantity of Mo leached will eventually equal the quantity added, resulting in a steady-state (plateau) soil Mo concentration (load). The plateau concentration depends on the annual Mo addition rate and the assumed annual Mo leaching rate. The critical condition in the risk assessment for Pathway 6 is that the plateau Mo concentration be less than the soil Mo concentration associated with forage Mo effect on cattle.
Average (long-term) leaching rates do not characterize pollutant leaching in the short term. Leaching is typically nonlinear, with high initial rates, followed by progressively slower rates in subsequent years as pollutant mass decreases. We regarded constant linear rates as appropriate for two reasons. First, biosolids additions are incremental over the assumed 100-yr site life, which better fits a constant, linear loss per year assumption. If the entire cumulative biosolids load were applied in one year, a nonlinear leaching model would be more appropriate. Second, selection of an average (long-term) annual loss rate results in less leaching estimation in the short term (when soil Mo concentrations are high), which leaves more Mo available for plant uptake. Overall, we believe such assumptions are more protective of the Pathway 6 HEI.
We estimated long-term Mo leaching rates (k values) from various literature sources mentioned previously. Few published studies quantified annual Mo leaching; rather, most reported changes in soil Mo concentrations many years after initial Mo contamination. The field data are presented in Table 4, and do not include the detailed laboratory data of Artiola (J.F. Artiola, personal communication, 1999), which suggest very high biosolids Mo leaching rates under uniform, short-term leaching conditions. Data in Table 4 represent estimated annual Mo leaching rates of biosolids Mo (McBride et al., 1999; Nguyen, 1998) and natural (Phillips and Meyer, 1993) or anthropogenic (Hornick et al., 1977) Mo sources. The data represent leaching periods of 3 to 35 yr. The average leaching rate (k) for data in Table 4 is about 5% yr-1, with a range of 1.8 to 12.5% yr-1. The USEPA (1992) used a leaching rate of 12% yr-1 for As, an anion expected to exhibit soil mobility similar to Mo.
Net soil Mo load (kg Mo ha-1) is calculated as a function of time, assuming an application of 10 Mg ha-1 yr-1 of biosolids containing 40 and 75 mg Mo kg-1 at 1.8% yr-1 leaching (Fig. 1) and 5% yr-1 leaching (Fig. 2). Figure 1 predicts that a biosolids containing 40 mg Mo kg-1 can be safely applied for 100 yr without exceeding the 20 kg Mo ha-1 diet-weighted (FC-corrected) criterion when the smallest annual leaching rate (k = 1.8% yr-1) is assumed. Figure 2 predicts that even a biosolids containing the ceiling concentration of 75 mg Mo kg-1 can be applied for 100 yr without exceeding a plateau Mo load of 16 kg ha-1 when k = 5% yr-1. A biosolids containing 40 mg Mo kg-1 could be applied for 100 yr (at 5% leaching yr-1) and never exceed 8 kg Mo ha-1, well below the soil load associated with Mo-induced hypocuprosis, even when the dietary factor (FC) is ignored. High soil pH exacerbates plant Mo uptake, but also promotes Mo leaching, so long-term Mo risk is minimized.
|
|
 |
 |
We believe that a conservative correction for leaching (LC = 0.45) makes the risk assessment appropriately realistic, but recognize that a different LC value could be necessary as additional field-scale leaching data become available.
Incorporating Mo leaching considerations into the Pathway 6 algorithm yields the following:
 |
Following the USEPA policy of rounding down critical limits, we derive an RPc value of 40 kg Mo ha-1, and a corresponding alternate pollutant limit (APL) of 40 mg Mo kg-1 for Part 503 Table 3.
Leaching of Mo at a particular site will be influenced by local climatic and management (irrigation leaching fraction) factors, but should not be less than the 1.8% yr-1 assumed here except under acid soil pHs where molybdenum availability is low and Mo toxicity is unlikely. At high soil pH, where Mo availability and toxicity can be problematic, leaching should be greater than assumed here, and ignoring its influence on permissible Mo loads is unnecessarily and inappropriately conservative.
Numerical Standards for Molybdenum
Based on the above considerations and the modified algorithm presented, we suggest the following numerical standards for Mo in the land application of biosolids:
 |
Additional Considerations
Animal and Pasture Management
Copper deficiencies in forage are common worldwide (Gartrell, 1981), and copper deficiency significantly affects ruminant livestock production in large areas of the world (McDowell, 1992, 1997). Similarly, Mo-induced hypocuprosis has been widely recognized as a problem in selected areas of the USA, Canada, Europe, and Australia for decades, and livestock owners have had corrective management practices recommended to them for >50 yr (e.g., Lewis, 1943; Cameron and Goss, 1948). The best primary management practice is to supplement cattle feed with added Cu; typically, 10 mg Cu kg-1 of diet, but increasing to 25 mg kg-1 during pregnancy, and to as much as 50 mg kg-1 if "animals have access to high sulfate water" (Gooneratne et al., 1989). Allaway (1977) concluded that "stockmen in the USA have met the problems of Mo toxicity by treating the animals, rather than measures applied to the soil or crop. The general utility of Cu therapy to prevent Mo toxicity has minimized efforts to meet the problem through soil or plant management."
Thus, Mo-induced Cu deficiency problems are well known, easily recognized, and easily corrected by ranchers operating in areas prone to Mo toxicities. Ranchers and farmers, who know from decades of experience that their soils produce crops prone to Mo accumulation, are unlikely to intentionally add high Mo loads via biosolids, or any other soil amendment, unless this is done in conjunction with Cu supplementation. Ranchers or dairymen receiving forage from elsewhere are not likely to feed cattle excessive amounts of Mo-accumulating plants (e.g., legumes) indiscriminately. Imported forage is typically dry hay, which has a much lower risk of Mo-induced hypocuprosis (e.g., Suttle, 1983; Mill and Davis, 1987; Underwood and Suttle, 1999). Large animal feeding operations typically keep close watch on animal condition and diet, and would logically address a potential Mo problem quickly with additional Cu supplements, or alternative feeds. The common sense and practicality of good pasture and animal management should add significant confidence to the risk assessment presented here.
Uncertainties in the Risk Assessment
A weakness in the original USEPA risk assessment was the limited animal (HEI) response database, especially animals consuming forages grown on biosolids-amended land. Much of the animal response data used to justify low (<10 mg Mo kg-1) threshold intake (TPI) values, for example, were based on Mo salt additions to diets of confined animals. There are very few studies involving biosolids-Mo fertilized forage grazed by ruminants (e.g., O'Connor and McDowell, 1999; K. Broersma and W.C. Gardner, personal communication, 1999). Gardner et al. (1996) recently reported minimal responses of cattle to grazed forages with total Mo concentrations of 20 to 40 mg Mo kg-1 and forage Cu to Mo ratios <1. Forage (including legumes) was grown on mine spoils with high pH (calcareous). Molybdenum concentrations in animal plasma and livers increased, but there was no evidence of molybdenosis or detrimental effects in the animals. Cattle supplemented with Cu had the same adequate Cu in plasma and livers, and the same overall health, as nonsupplemented cattle. Forage S concentrations were apparently low (but normal: 2 g kg-1) in their study, and may have minimized thiomolybdate (or direct cupric sulfide) formation. Also, Loneragan et al. (1998) noted that high Mo could decrease ruminal production of sulfide. Thus, there may actually have been too much Mo in the forage to promote adequate sulfide, and thiomolybdate, formation in the Gardner et al. (1996) study. Canadian researchers (K. Broersma and W.C. Gardner, personal communication, 1999) hypothesize that forage Mo is somehow unavailable for reaction in the ruminants, an effect similar to reduced Mo availability when forage is dried (Allaway, 1977). Research into the form of Mo in the forage and the effects of increased S in the diets is underway. Animals are being grazed on leguminous forages grown on biosolids-amended land (K. Broersma and W.C. Gardner, personal communication, 1999).
Forages grown on biosolids-amended soil frequently have increased S contents (Nguyen, 1998; O'Connor and McDowell, 1999; McBride et al., 2000). Low-maintenance pastures and rangelands are often underfertilized, and respond well to nutrient inputs provided in biosolids. Bahiagrass (Paspalum notatum flugge) (Nguyen, 1998; O'Connor and McDowell, 1999) was especially responsive to biosolids S, and the response was seemingly independent of biosolids source or rate of application (1 to 8 times agronomic rate). Forage S concentrations in all biosolids treatments reached potentially problematic levels (>4 g kg-1). Thus, additional attention to Pathway 6 risk assessment may be warranted aside from Mo issues. Special vigilance for adequate Cu supplementation of animals grazing biosolids-amended land to address possible biosolids S effects may be required.
