How much soil do cattle ingest? A review

Limited number of studies, none from Australia

First, we found that none of these studies was conducted under Australian conditions, with the research undertaken in England, France, New Zealand and the United States of America. A summaries of data from 11 studies are presented in Table 1 (dairy) and Table 2 (beef). For dairy cattle, the reported results of ingested weight ranged from 0.1 kg/day to 2.61 kg/day (maximum, New Zealand) and an average range of 0.33–1.22 kg/day. For beef cattle, the maximum ranged from 0.50 kg/day to 2.43 kg/day (maximum, England) and an average range was 0.20–0.99 kg/day.

The highest soil-ingestion values in cattle were recorded in a dairy herd in New Zealand (Healy 1968) and a beef herd in England (Thornton and Abrahams 1983; Abrahams and Thornton 1994). These values were at the extreme upper end of the range of measured soil ingestion for cattle, at 1 kg more than other results reported in the same study, and in other studies. The lowest value (0.1 kg/day) reported in the literature was for dairy cows held on a paved surface. Several studies noted that even under conditions of abundant pasture growth, reported/measured soil intake is never zero (e.g. Healy 1968).

Indirect measurements

None of the reviewed studies had directly measured soil ingestion in cattle. Of the 11 studies, all used modelling based on soil content in faeces and/or oesophageal samples to estimate soil ingestion. Samples were analysed either for titanium or acid-insoluble ash, to calculate proportion of soil ingested, either as kilogram per day or as a percentage of DMI. These methodologies were based on a level of uncertainty, relying on estimated input variables such as acid-insoluble ash concentrations in soil, feed digestibility, and feed intake to calculate soil ingestion. We discuss the limitations of this approach below. More recently, Jurjanz et al. (2012) (in France) measured all input values and reported average and maximum daily soil intakes in individual dairy cattle of 0.47 kg/day and 1.3 kg/day respectively. They concluded that cows ingested less than 0.25 kg/day under good pasture conditions with sward height above 5 cm.

High variability in results linked to many variables

Soil-ingestion studies reported that variability in soil ingestion in cattle was influenced by a range of factors, primarily seasonal conditions, but also soil type, stocking density, pasture availability and livestock management system, including supplementary feeding. The highest recorded ingestion rate occurred during winter in New Zealand (Healy 1968) and England (Thornton and Abrahams 1983; Abrahams and Thornton 1994). In winter, stocking rates were at their highest, pasture growth rate was low, soil was muddiest and earth worm activity was high. Muddying of forage by trampling, raindrop splash and earthworm casts indirectly contributed to soil-ingestion levels in the New Zealand studies (Mayland et al. 1975). Healy (1968) found that soil ingestion peaked in autumn and winter under New Zealand conditions. Under dry conditions or ’droughty’ conditions, Mayland et al. (1975) found that faecal-soil concentrations generally increased as the amount of available forage decreased. This result was consistent with increased likelihood of soil exposure as pasture availability declined.

Analysis of published data

Given the range of values reported across different seasonal conditions and the ‘noise’ and perceived ‘outliers’ within the data, we conclude that the average soil-intake rate is considered more representative than are maximum values. The average soil consumption ranged from 0.33 to 1.22 kg/head.day for dairy cattle (Healy 1968; Green and Dodd 1988). Beef cattle had a lower average range of 0.2–0.99 kg/head.day (Mayland et al. 1977; Abrahams and Thornton 1994). We note that the maximum and minimum consumption rates were higher for dairy cattle than for beef cattle, as presented in Tables 1 and 2. So, we initially concluded that the average soil consumption should be considered higher for dairy cattle. For dairy cattle (Table 1), there were nine values in the ‘average’ soil-intake column. We calculated the average of the averages to be 0.59 kg/head.day. We performed the same calculation in Table 2 for beef cattle; the average of eight averages was 0.62 kg/head.day. Unexpectantly, these values were quite consistent and nearly equal.

We recognised that outliers could influence these calculations to produce the unexpectantly consistent value of ~0.6 kg/head.day. So, we removed the highest and lowest figures from the averages column and recalculated the average of the average. After removing the two extreme values, we found that beef and dairy cattle ingested 0.63 and 0.55 kg/head.day respectively.

Another approach would be to use the median value. In Table 1, the median of nine average values is 0.47 kg/head.day for dairy cattle. In Table 2, the two median values were 0.62 and 0.73, and the mid-point is 0.67 kg/head.day for beef cattle. So, using the median, we calculated that dairy and beef cattle ingested 0.47 and 0.67 kg/head.day respectively. Again, the calculated value for beef cattle was higher than that for dairy cattle. Before making any further conclusions, we undertook an interrogation of the scientific rigour behind the published data, to understand any additional factors that may be causing the wide range of reported results, and in particular values at the extreme ends of the range.

Published studies report values largely not applicable to Australia

We did not find any Australian studies about bovine soil ingestion; therefore, our deduced average soil ingestion levels are speculative. There are several reasons why our deduced levels may not be applicable to Australia. Australia is the driest continent (El Saliby et al. 2009) and ingestion levels in other countries are likely to have been evaluated on moister soils or in wetter climates. Overseas data may be applicable only in the wettest periods in Australia (Dey et al. 2019), such as for 2010–2012, which was the wettest 2-year period on record in Australia (White and Fox-Hughes 2013).

Intentional soil ingestion may provide essential minerals or elements (Kreulen 1985). However, Australia has a uniquely high proportion of nutrient-poor soils which are unlikely to provide essential minerals or trace elements (Kirby and Stuth 1980; Orians and Milewski 2007). Australian soils are different from those on other continents and are depleted in elements such as phosphorus, iodine, cobalt and selenium (Orians and Milewski 2007). Therefore, we assume that Australian soil consumption is likely to be incidental rather than an intentional part of animal husbandry.

Soil ingestion by other organisms is variable, largely dependent their style of feeding (Beyer et al. 1994). Sandpipers (Calidris spp.) feed on invertebrates in mud or shallow water; sediments composed 7–30% of their diets (Beyer et al. 1994). The nine-banded armadillo (Dasypus novemcinctus, soil is 17% of diet), white-footed mice (Peromyscus leucopus, 15%), American woodcock (Scolopax minor, 10%) and raccoon (Procyon lotor, 9%) had higher rates of soil ingestion, presumably because they feed on soil organisms. Conversely, herbivores consumed less soil, including bison (Bison bison, 7%), black-tailed prairie dog (Cynomys ludovicianus, 8%), wild turkey (Meleagris gallopavo, 9%) and Canada geese (Branta canadensis, 8%) (Beyer et al. 1994). Therefore, studies on ingestion by these organisms cannot be used as a surrogate for soil ingestion by cattle but were indicative that many organisms do ingest soil.

Highest values are not representative

Values at the upper end of reported range represent a point in time, and are not representative of a year-round average. The highest soil ingestion values in cattle were recorded in a dairy herd in New Zealand (Healy 1968) and a beef herd in England (Thornton and Abrahams 1983; Abrahams and Thornton 1994). These values are at the extremes of the range of measured soil ingestion for cattle, representing a kilogram more than other results reported in the same study, and in other studies. They reflect temporary increases in soil ingestion, possibly being linked to season, stocking rate and other factors, and are not an indication of ongoing daily ingestion rate. Healy (1968) surmised that annual intakes will range from about 0.25 kg/day to 1.24 kg/day, with peaks occurring in autumn and winter under New Zealand conditions.

Weak methodologies in published studies

Published studies largely used weak methodologies that relied on crudely estimated input variables. Soil-ingestion estimations in our reviewed studies were based on the ingestion of a non-absorbed indicator, such as acid-insoluble ash or titanium oxide. Calculations relied on assumed values for total dry matter consumed per day and the fraction of dry matter digested (digestibility). Fries (1996) noted that the use of conservative values for these parameters could lead to unrealistically high soil-ingestion values. For example, because a static value for digestibility and DMI in soil ingestion was assumed, calculations did not take into consideration changes in pasture availability throughout the year, such as when forage is sparse or when livestock receive supplementary feed. The approach would overestimate the amount of soil ingested. For this reason, Fries (1996) concluded that many of the higher values for soil ingestion in the literature were incorrect. Consistent with Beyer and Fries (2005), Fries (1996) considered that even the average values reported in several studies on soil ingestion were too high, and may be at least twice the true amount of soil ingested. Also, several authors, including Beyer and Fries (2005), commented that the use of a single value for DMI in soil-ingestion calculations did not consider variations in feed intake within a herd; potentially over-estimating intake for younger animals. Additionally, Jurjanz et al. (2012) noted the limitations of early studies, which considered mixed herds. We discuss the risk of these limitations in this approach below.

International guideline values for risk assessments

Soil ingestion is a potential pathway for exposure of livestock to contamination and entry point of substances of concern into the food chain (American Petroleum Institute (API) 2004). Conservative livestock intake rates were calculated for a range of parameters to inform assessment of risk to the health of animals exposed to pollutants in soil or other environmental media, and to humans from consumption of products from those animals. In Australia, the reference values commonly used to inform risk assessments came from the American Petroleum Institute guidance document from its work on the protection of livestock exposed to petroleum hydrocarbons. A soil-ingestion rate of 2.4 kg/day for dairy cattle and 2.1 kg/day for beef cattle were adopted (American Petroleum Institute (API) 2004). These threshold values, which the authors acknowledged to be conservative, were derived from the maximum percentage of soil found in the diet of cattle, as reported in any of the studies reviewed by American Petroleum Institute (API) (2004). Although a range in values was reported, from 0.2% to 18.8%, the adopted values were derived from the highest end of this range. The values compare with the maximum ingestion rates reported by Healy (1968) and Thornton and Abrahams (1983) and Abrahams and Thornton (1994). There is considerable risk/limitation in adopting this approach as the reported values are likely to be an overestimate of the amount of soil consumed (per Beyer and Fries (2005) and others). Additionally, the high-end values represent a ‘point in time’ and are not representative of average conditions on a day-to-day basis.

The US EPA recommended significantly lower assumed soil-ingestion values of 0.4 kg/day in dairy cattle and 0.5 kg/day in beef cattle for risk assessments (USEPA 2005). These default values were based on a different approach recommended by Fries (pers. comm. to US EPA in 1994) and validated by Fries (1996) and Beyer and Fries (2005), by using the mean as a more representative of average yearly soil ingestion. The 0.4 kg/day intake rate for dairy cattle was equal to approximately 2% DMI if the cow was consuming 20 kg of dry matter per day, while the value accounted for supplementary feeding, a standard dairy farming practice in the United States. Also, Beyer and Fries (2005) noted that the conservative value was reasonable for subsistence farmer scenarios (i.e. where cattle are largely pasture-raised/range-fed/grass-fed). The higher assumed/calculated value of 0.5 kg/day was recommended for beef cattle raised on pasture. Because the value is derived from a DMI typical of mature cattle, it was considered by Beyer and Fries (2005) to overestimate soil intake by younger (i.e. smaller) animals. Younger animals make up the largest portion of the meat supply in both the United States and Australia and DMI approach could mis-inform risk assessments. Also, Jurjanz et al. (2012) noted the limitations of early studies that considered mixed herds.

Average intake more informative than maximum values

When evaluating the significance of soil residues for meat animals, Fries et al. (1982) recommended using the average yearly soil intake as a more representative value of soil ingestion and likely contamination levels in meat than a single highest value. However, because milk was more responsive to changes in chemical levels in diet, worst-case modelling of chemical contamination in dairy cattle should use the average daily soil intake during the month of greatest intake (Fries et al. 1982). Importantly, regardless of meat or milk consideration, Fries et al. (1982) did not recommend the use of maximum soil-ingestion value under either circumstance.

Recommended value most representative of Australian conditions

Earlier in this paper, we calculated average and median values from soil-ingestion rates reported in published literature. Then, we interrogated the methodologies of those studies, and highlighted the limitations in many of the reported values (used in our earlier analysis), and that reported ingestion rates were at the higher end of the range. The widely cited maximum levels reported in New Zealand occurred under specific conditions where dairy herds were wintered on agistment blocks at higher stocking rates. We note that livestock systems, climate, soil type, pasture species and other cumulative factors contributing to soil ingestion rates in the New Zealand study, and most overseas studies reviewed, are unlikely to be comparable to conditions across most of Australia.

In considering the reviewed literature, the study of cattle held under semiarid conditions in USA (Mayland et al. 1975) is considered more comparable to conditions in Australia. This work reported daily soil-ingestion levels of less than 1.33 kg/day and a median value of 0.5 kg/day in beef cattle. Importantly, cattle weights in this study included young steers and cows of the size usually entering the food chain, ranging up to 450 kg in weight, within the liveweight range required by Australian butchers and supermarkets in Australia (Andrews 2015). Additionally, calculations of soil intake consider varying inputs for DMI and digestibility, providing a more robust framework. The median 0.5 kg/day soil ingestion rate (Mayland et al. 1975) aligns with the research of Jurjanz et al. (2012), who also used known, rather than estimated, input values to calculate soil ingestion in dairy cattle, reporting an average of 0.47 kg/day.

We recommend a 0.5 kg/day soil ingestion rate as appropriate for Australian conditions. Additionally, this value aligns with existing international guidelines currently in place in the United States (USEPA 2005).

Application

In agricultural settings, soil ingestion by domestic animals is recognised as an important pathway of animal exposure to toxic pollutants in soil and residues in animal products (e.g. Beyer and Fries 2005). Human health and ecological risk assessment is a process used to understand and manage potential risks to a population from exposure to pollution in the environment. The accuracy of variables used to inform a health risk assessment will determine the quality of the estimates provided and, therefore, appropriate management of those risks. Contamination in farming settings, based on practices not consistent with current expectations and standards, may occur through direct application of or spray drift from pesticides (e.g. orchards, sheep dip sites), application of soil amendments such as biosolids or compost, or indirectly, through chemical leakage from discarded farm materials such as lead batteries, mine contamination, legacy use of chemicals, or where land is subject to flooding and flood waters contain contaminants.

Consistent with Beyer and Fries (2005), we consider that the 2.4 kg/day soil-ingestion rate (American Petroleum Institute (API) 2004), commonly used by Australian risk assessors, is overly protective. Although a precautionary approach is appropriate in assessing risks to human populations, use of such conservative default values could potentially have an adverse effect through over-estimation of risk and raising unnecessary concern among the general population. Our recommendation is for 0.5 kg/day value be used as a default for agricultural settings in Australia (consistent with Beyer and Fries (2005)).