SCOPE 50 - Radioecology after Chernobyl, Chapter4, Terrestrial Pathways

Artificial radionuclides enter terrestrial ecosystems from atmosphere (e.g. as a result of stack discharges or worldwide contamination following weapons testing), via surface waters (e.g. as a result of discharges to inland waters or leaching from shallow land disposal sites), or from ground water (e.g. as a result of releases from waste repositories). The emphasis in this chapter is on behaviour after atmospheric input, though many of the pathways and processes discussed are relevant to other modes of input.

Behaviour of radionuclides in terrestrial ecosystems is determined by the initial chemical form of the radionuclide entering the ecosystem. The global fallout after nuclear weapons testing was largely made up of water-soluble and exchangeable forms of ¹³⁷Cs and ⁹⁰Sr. In contrast, deposition of ¹³⁷Cs and ⁹⁰Sr after the Chernobyl accident was often in non-exchangeable forms (Bobovnicova et al., 1990). Also, close to Chernobyl, a considerable fraction of the deposited radionuclides was present as fuel particles insoluble in water or neutral solutions (Konoplyov et al., 1988).

Deposition processes to ground are discussed in Chapter 3. Transfer of radioactivity from atmosphere to vegetation involves: deposition (usually treated by means of a deposition velocity, V_g); interception (usually treated as an interception fraction, r, defined as the proportion of the total deposit to ground which is initially intercepted by vegetation); and retention (usually treated with a weathering half-life, T_w, defined as the half-life for decline in vegetation concentration on a mass basis, or content on an area basis, following initial deposition). All three parameters are sensitive to the particle size and chemical characteristics of the depositing material, the vegetation or ground surface characteristics, prevailing weather conditions, and state of growth of the ground cover.

There is substantial published information on the interaction of particulates with vegetation surfaces both from experimental (e.g. wind tunnel) and field studies. Much of this work was pioneered by Chamberlain and co-workers who identified the three important processes as: gravitational sedimentation, inertial impaction, and eddy diffusion deposition. Chamberlain (1970, 1975) provided mathematical treatments for these processes on the basis of experimental investigations. Many of the field data are for ⁹⁰Sr and ¹³⁷Cs deposited in weapons-testing fallout, though information has also been obtained for particulates contaminated with heavy metals (Little, 1979; Martin and Coughtrey, 1982).

It is generally very difficult to obtain reliable estimates of V_g from field studies, especially if deposition occurs over prolonged periods and by both wet and dry mechanisms. As a result, there is often a significant discrepancy between values obtained from carefully controlled experiments and those obtained from field measurements. Typically, values of 0.001 to 0.005 m/s have been used for V_g in modelling studies as appropriate to small particles with higher values (0.01 m/s) for reactive gases such as iodine and lower values (0.0001 m/s) for unreactive gases and organic forms of iodine (Coughtrey et al., 1983

1985).

Interception varies according to the leaf area index, botanical composition, morphology and surface roughness characteristics of the underlying vegetation. Compared to studies on deposition, there are relatively few published investigations on interception. A well-cited series of studies is that of Milbourn and Taylor (1965) involving application of radioisotopes in solution as sprays to different types of pasture. Chadwick and Chamberlain (1970) proposed that initial retention by herbage could be modelled according to:

Chadwick and Chamberlain calculated a mean value for µ of 2.3 m² /kg. The data of Milbourn and Taylor (1965) indicate a range of 1.2 to 6.2 m² /kg for soluble deposits applied to a variety of pasture types. Miller (1980) advocated the use of a normalized deposition velocity (V_d) to take account of the effects of herbage density. Relative to pastures, very few data exist for interception by crops. For cereals, Aarkrog (1983) showed a strong effect of stage of development on interception and retention of radionuclides applied in solution. Eriksson (1977) showed large differences in interception between particulate-bound and soluble radionuclides in wet or dry conditions.

For prolonged deposition to vegetation (as occurred after weapons testing), Chamberlain introduced the concept of normalized specific activity (NSA), defined as the ratio of radionuclide concentration in foliage (Bq/kg dry weight) to the rate of deposition to ground (Bq/m²/d). A value of 40 m²d/kg was considered appropriate for herbage in good growing conditions. NSA values are much higher for poor growing conditions and some types of native vegetation (Martin and Coughtrey, 1982).

Radioactive materials can be removed from atmosphere by rain-out (i.e. incorporation in raindrops) and by wash-out (i.e. removal of particulates from below the cloud base by falling raindrops). Generally, both wash-out and rain-out are incorporated into a dimensionless wash-out ratio (WR) (Chapter 3). Very little information exists on the interception of radionuclides after wet deposition, though the data base on this subject has been extended recently by Hoffman et al. (1989) and Caput et al. (1989).

Weathering half-lives in the range of 12 to 17 d are commonly encountered in the literature. Such values are probably appropriate to initial retention of fission products after weapons testing for which losses occur as a result of re-entrainment of carrier particles, sloughing of leaf surface waxes, and leaching by rainfall. Much longer half-lives or more complex patterns in retention can be associated with radionuclides applied in solution, sub-micron particles, and gaseous or volatile radionuclides such as those of sulphur and iodine. Losses of tin and iodine by volatilization have been reviewed by Arnold (1990).

To obtain reliable estimates of T_w from field data, supporting data are required on the pattern of growth of the vegetation and the timing and type of precipitation after initial interception. It is difficult to evaluate models for retention from field data because of the confounding effects of foliar absorption and translocation, or root uptake.

Most published information on behaviour and movement of radionuclides in soils reflects patterns observed after initial deposition to soil surface, either in field conditions (from sites surrounding nuclear installations or affected by fallout from weapons testing), or from laboratory experiments using soluble tracers. Behaviour in soil and subsequent removal reflect the physico-chemical characteristics of the radionuclide, the properties of the soil (including changes with depth from the surface), the type of vegetation, hydrology, and underlying geology. There have been several studies involving sampling of soils at different depths and comparison of the observed distribution and total content as a function of time. Residence times for turnover of radionuclides in a defined soil layer have been derived but generally on the basis of the total contents of radionuclides rather than on the basis of specific fractions extractable with defined extracting agents. As a result, models for radionuclide movement in soils are generally rather simple and usually empirical (Coughtrey, 1988). For some radionuclides e.g. ⁹⁰Sr and ¹³⁷Cs more detailed observations have been obtained and movement simulated on the basis of diffusion theory (Prokhorov, 1973; Frissel and Pennders, 1983; Kirk and Staunton, 1989; Voitsekhovitch et al., 1991).

Figure 4.1 Schematic model for radionuclide transformation processes in soil.

Table 4.1 Rate constants (d^-1) for dissolution of ⁹⁰Sr and ¹³⁷Cs from fuel particles in the 30 km zone surrounding Chernobyl


Site	Soil type		⁹⁰Sr	¹³⁷Cs

Chernobyl	Soddy-podzolic		1.5 x 10^-3	0.87 x 10^-3
		loamy sand
Benyouka	Alluvial soddy acid		1.4 x 10^-3	1.0 x 10^-3
Kopachi	Slightly podzolic		4.0 x 10^-4
		loamy sand
Korogod	Arable soddy-podzolic		3.6 x 10^-3	1.5 x 10^-3
		loamy sand

From Konoplyov and Bulgakov, 1991.

The simplest equation to describe migration is the mass conservation equation, i.e.

		(4.2)
where C is the concentration in solution;
	S is the concentration in solid phase;
	t is time;
	x is the vertical distance;
	V is the interstitial solution velocity;
	D is the apparent diffusion coefficient; and
	S is the source term that describes processes other than adsorption and desorption.

The application of this equation is not limited to soils. It can also be applied to migration in rivers or to migration near a waste repository.

The whole collection of radionuclide transformation processes in soil can be described schematically as shown in Figure 4.1. Radionuclides present as original deposited particles, irreversibly sorbed to soil particles or present in ion-exchange sites on soil particles are subject to processes involved in particle migration. Radionuclides present in soil solution or as organic complexes are subject to processes involved in solution transport.

Processes involved in the movement of radionuclides in soil have been reviewed by Morgan (1990b). Important processes include: sorption, resuspension, mass transport and leaching.

With respect to dissolution of fuel particles deposited after the Chernobyl accident, rate constants derived from the 30 km zone surrounding the source are given in Table 4.1. Similarity between the values obtained for both ⁹⁰Sr and ¹³⁷Cs indicate that the processes involved are not determined to a large extent by the chemical characteristics of the different radionuclides.

The potential for transport of a radionuclide through the soil profile is determined by chemical and biological `availability'. Information on this subject derives from many sources, the most important of which are studies involving the extraction of soil with a sequence of chemicals thought to remove different chemical fractions. Results are then compared with studies on plant uptake or vertical transport to provide chemical explanations for the observed behaviour. In the context of environmental radioactivity this is a relatively new field, however it is well-supported by many years of work in the soil and plant sciences.

In the absence of information on chemical extractability, the most common approach has been to use a distribution coefficient (K_d), defined as the ratio of the concentration of a radionuclide in the solid phase (Bq/g dw) to that in the solution phase (Bq/ml). K_d is sensitive to a number of factors including: chemical characteristics of the radionuclide, chemical and physical properties of the soil, methods used to obtain the K_d, and the solution concentration range.

Application of K_d in terrestrial radioactivity studies is confounded by the fact that the chemically soluble fraction of a radionuclide in soil that is available for transport via mass flow is not necessarily that fraction which is biologically available. Nevertheless, determination of K_d values for a range of radionuclides under comparable conditions does allow general conclusions to be drawn about the likely behaviour of a radionuclide in the environment. Radionuclides with high K_ds under specified conditions (e.g. plutonium, caesium) generally show little movement in soils and have a high potential for resuspension, whereas those with low K_ds (e.g. technetium in lower valency states, strontium) show greater movement and have a low potential for resuspension.

Resuspension of radionuclides attached to soil particles provides a mechanism for loss from the system by water- or wind-mediated erosion, and for contamination of vegetation surfaces. Resuspension has generally been treated by means of an empirical resuspension factor (RF) defined as the ratio of a radionuclide concentration in air (Bq/m³) to the ground surface content (Bq/m²).

Surface soils are subject to considerable turnover as a result of the action of burrowing organisms. Bishop (1989) reviewed information on this subject and concluded that biotic transport of radionuclides by, or as a result of the actions of, burrowing animals, could represent a significant and rapid pathway for radionuclide migration in surface soils. Bishop noted that the maximum amount of soil moved by earthworms in established pastures is in the order of 10 kg/m²/y (equivalent to a soil depth of around 14 mm). For radionuclides which become fixed to soil particles and therefore unavailable for leaching, redistribution within soil profiles will be dominated by mass transport. Mass transport may also occur by physical mechanisms, e.g. the movement of soil particles through macropores.

Net losses from a soil profile via leaching are difficult to quantify, even with complex models. For well-mixed soils the generally accepted approach is to make use of a soil leaching coefficient, S₁ which is defined as follows:

To apply the above, both V_w and K_d must be defined. For V_w, a complete hydrological model is required. Such models are beyond the scope of this work. Since radionuclides often exist in soil as more than one chemical species, a separate K_d is required for each. Also, although ion exchange is often the major absorption mechanism, for some radionuclides and soil types, other mechanisms of sorption or fixation apply.

For unmixed soils which have distinct horizons of contrasting physical, chemical and biological characteristics, more complex models are required. These are generally multi-compartment models with transfer from one compartment to another controlled by first order kinetics (Haywood et al., 1980; Thorne and Coughtrey, 1983).

The two routes for uptake of radionuclides by plants are foliar absorption and root absorption. Scott Russell (1966) used the term `plant base' absorption for certain specific conditions but the physiological meaning of this term is unclear. Translocation has generally been considered only in the context of foliar contamination with the `translocation factor' defined as the proportion of a surface deposit subsequently transferred to edible fractions.

Foliar absorption of soluble radionuclides has been given little attention in the literature; the processes involved and the patterns to be expected for different radionuclides are not clear. Foliar absorption of gaseous materials is better described and quantified, primarily from analogy with non-nuclear contaminants. Thus, for ³⁵S, studies on sulphur dioxide can be used to determine the resistance to passage at various stages of the transfer. For ¹⁴C and ³H a substantial literature exists on photosynthesis, respiration and transpiration, and models of varying complexity exist to simulate absorption and subsequent redistribution. The degree to which surface-deposited radionuclides are absorbed by plant surfaces is very important in interpreting information on deposition, interception and retention.

Watkins (1990) reviewed information on absorption and translocation of radionuclides applied in solution. Foliar absorption may occur via the stomata or across the cuticle. Stomatal penetration by aqueous solutions involves many mechanisms and the main determining factors are: surface tension of the solution; contact angle (wettability); and morphology of the pore. Cuticular penetration is determined by the morphology of the surface and the chemical composition of the cuticle. The longer a contaminant remains on the plant surface the more likely it is to be absorbed. Thus rainfall is an important factor. An additional factor is the role of phyllosphere micro-organisms which may utilize surface contaminants rendering them either more or less available for foliar absorption. Foliar absorption and translocation of ¹³⁷Cs proved to be of some significance after the Chernobyl accident for fruits and pulses. A notable example was the observed transfer to hazelnuts (Monte et al., 1990).

Both soil and plant factors are involved in root uptake. Some information is available on the capacity of roots of different species to absorb different radionuclides from solution. This has generally been obtained from solution culture experiments using carefully controlled conditions. Unfortunately, such information is rarely directly applicable to field conditions due to the lack of knowledge on the partitioning of radionuclides between 'non-available' and 'plant-available' fractions in soils. Additionally, root absorption is not explained in all cases by physical and chemical theory. In many cases absorption relies on biological factors such as the presence or absence of micro-organisms intimately involved with the root (the mycorrhiza), and the release and uptake of carrier molecules at the root surface.

In general, the majority of studies on plant uptake from soil have evaluated the available data in the context of a soil-to-plant transfer factor (SP) defined as the ratio of a radioisotope concentration in plant (Bq/kg dry, wet or ash basis) to that in soil to a defined depth (Bq/kg dry, wet or ash basis). The main soil characteristics affecting SP are: pH; cation exchange capacity; organic matter content and composition; and clay content and composition. Extensive data exist in the literature for a range of plant and soil combinations, climate and growth conditions, different radionuclides and forms of contamination. Unfortunately, many of the available data are difficult to interpret without supporting data for the state of growth and biomass distribution of the plant concerned, relevant physical and chemical characteristics of the soil, and details on depth distribution of the radioisotope.

To overcome these difficulties, a working group of the International Union of Radioecologists (IUR) defined standard conditions for experimental studies and evaluated data from these experiments in relation to measured soil parameters (IUR, 1982). The main difficulty in the derivation, understanding, and application of SP is that it is defined on the basis of total concentration of a radionuclide in soil. For many radionuclides the total concentration in soil bears little or no resemblance to subsequent plant uptake. Additionally, SP (like K_d) is defined for equilibrium conditions.

For radiostrontium and radiocaesium an alternative approach is based on the similarity of the ecological behaviour of Cs and K and of Sr and Ca. The hypothesis is that the mineral composition of a crop is rather constant and that the Cs and Sr content of the crop can be determined from the corresponding Cs/K and Sr/Ca ratios for the soil (Wirth et al., 1985).

A further approach involves analysis of the kinetics of solution to root transport (Nye and Tinker, 1977; Shaw and Bell, 1989). This requires substantial information on the chemical composition of solutions and roots.

The study of soil solution and its interaction with soils and plant roots is only possible if suitable methods are available for the isolation of soil solution. A review of these methods is given by Lembrechts (1991).

Once in the plant, retention of a radionuclide depends on relative distribution and a variety of plant physiological parameters. `Availability' of radionuclides for leaching has been given little attention in the literature, though early studies reported by Tukey et al. (1958) indicated substantial differences between those that are readily leached (isotopes of Na and Mn) and those that are effectively retained (isotopes of Fe, Zn, P and Cl).

Following root or foliar absorption, radionuclides can be translocated either above or below the point of entry. In many published studies, emphasis has been placed on the above-ground fractions of the plant and particularly the edible fractions. This neglects the fact that radionuclides may have been taken up by the plant but not translocated to the organ(s) of interest and, as a result, may have undergone chemical transformations which could have a profound effect on subsequent behaviour. In these circumstances, root uptake and downwards translocation may even affect the observed distribution of the radionuclide in soil as has been suggested for plutonium (Coughtrey et al., 1983

1985).

There is a substantial literature on the uptake, distribution and retention of radionuclides in animals. Much of this literature concerns laboratory or non-domestic mammals and has been obtained and evaluated for the specific purpose of developing models for assessing dose to man following inhalation or ingestion of radioactive materials.

Coughtrey et al. (1983

1985) reviewed data on transfer of radionuclides to animals and specified mathematical models for use in radiological assessments. More recent information was reviewed by Van den Hoek (1989) and by Coughtrey (1990). The important parameters involved in transfer to and metabolism in animals are:

Unlike studies on the behaviour of radionuclides in soils and plants, many published studies on animals were undertaken in the context of mathematical analysis of time-dependent trends. Nevertheless, there are a number of radionuclides and animals for which data are less extensive and for which more empirical approaches were required. In particular, transfer of radionuclides to eggs has been studied in only a few cases, as has transfer to milk of animals other than cows. An area of particular interest is transfer and dynamics of radionuclides in young animals compared to mature animals. In this context, changes in f₁ with increasing age, possibly as a result of changes in gut structure, changes in distribution between different tissues, and changes in retention during development of the animal all need to be considered.

The traditional analysis of transfer from pasture to animal is the use of a plant-to-animal transfer factor (PA) defined as the ratio of the radionuclide concentration in meat or milk (Bq/kg or Bq/l) to the daily intake (concentration in diet (Bq/kg) dietary intake (kg/d)). Values of PA have been derived from field data for a number of animal and radionuclide combinations, but these values are usually subject to uncertainty due to difficulties in obtaining accurate estimates of radionuclide concentration in diet, dietary intake, and other factors. These latter factors include: plant species composition; form of contamination; the extent to which it includes soil ingested directly or indirectly; dry weight and ash weight content (including supplements); and stable element content. Other factors include animal husbandry practices and the metabolic state of the animal during the period of observation. Additionally, the PA obtained from a field study is specific to the chemical form of the radionuclide and other conditions in that study since estimates of f₁ are not included explicitly in the calculation of PA. Finally, PA was defined originally by Ward and co-workers in the 1960s in the context that equilibrium should have been achieved between intake and meat or milk concentration (Ward and Johnson, 1986).

In the absence of equilibrium conditions, estimates of PA can be based on integrated intake over part or the whole feeding period, though even this can be unsatisfactory when there are changes in the level of dietary contamination or dietary intake, or the approach to equilibrium is slow.

Most interest in PA has been in respect of transfers to milk, with transfers to meat receiving less attention. A collation of transfer factor data was provided by Ng (1982).

Interest in the behaviour of radionuclides in natural ecosystems first developed during the 1960s because of the observed transfer of ¹³⁷Cs along the lichen

reindeer

man food chain, though a wider interest also developed at the same time because of the recognition of the importance of artificial radionuclides as a tool for tracing and understanding nutrient turnover in ecosystems. A survey of the literature from the 1960s and 1970s shows a strong interest in the then new subject of `radioecology' with several symposia devoted to its various aspects (e.g. Schultz and Klement, 1963; Cushing, 1976). An introduction and historical perspective of radioecology is provided by Whicker and Schultz (1982) and introductions to the Russian literature by Klechkovskii et al. (1971) and by Kulikov and Molchanova (1982).

Radioecological studies in semi-natural and natural ecosystems have not been limited to radioisotopes associated with weapons testing or to those providing tracers for nutrients. Auerbach (1987), for example, cited information on the longterm contamination of forest ecosystems with plutonium and technetium in the vicinity of Oak Ridge National Laboratory. He also emphasized the importance of such studies in understanding and predicting the behaviour of long-lived radionuclides. Grasslands, forests, and tundra environments have all been considered in the context of actinide distribution and transport (Hanson, 1980) and commonly provide conditions in which enhanced transfers from soil to plant or plant to animal can occur.

For radiocaesium, extensive studies in the 1960s during the period of weapons testing indicated that arctic and alpine ecosystems were particularly important in relation to impact on man. Subsequent studies (e.g. Hutchinson-Benson et al., 1985) demonstrated the persistence of ¹³⁷Cs in arctic environments many years after the peak of fallout, and a similar situation was recognized by Bunzl and Kracke (1984) for heathlands in Germany.

Decomposition provides a mechanism for transfer of radionuclides from dead vegetation to the soil system. Many semi-natural ecosystems have an accumulated organic layer at the soil surface which may provide a source of available radionuclides for subsequent transfer to other ecosystems or man.

The significance of semi-natural systems was further emphasized in many studies following the Chernobyl accident as summarized in the work by Desmet et al. (1990).

The consequences of radionuclide contamination of terrestrial ecosystems are strongly dependent upon a number of seasonal parameters (e.g. state of growth of the vegetation, agricultural practice, weather conditions) which control many of the processes involved in radionuclide transport summarized above. Experience after the Chernobyl accident highlights the need to take account of seasonal factors when describing the behaviour of radionuclides in terrestrial ecosystems after acute or chronic contamination. Seasonality is also highly relevant when making decisions on the application or relative effectiveness of countermeasures (NEA, 1991).

The following sections of this chapter provide state-of-the-art overviews of the behaviour of radionuclides in different types of ecosystem. Conclusions and recommendations are presented in Section 4.5.

4.2 AGRICULTURAL ECOSYSTEMS

The general theories for deposition, interception and retention (Section 4.1.1) also apply to agricultural products. In fact most models and parameters were derived for agricultural products from temperate zones. A distinction is made between dry deposition and wet deposition. In both cases only a part of the radionuclides is intercepted and/or retained by the vegetation. Interception factors are given in Table 4.2. Calculations of deposition are usually rather simple (IAEA, 1991 a) and three types of air contamination are distinguished. First noble gases; these are very inert and not deposited. Secondly aerosols of chemically non-reactive compounds, e.g. aerosols of Sr, Cs and transuranic elements. Rates of deposition and wash-out are dependent on particle size and not on the type of radionuclide (Table 4.3). Therefore general aerosol theories apply. Thirdly chemically reactive gases or aerosols, e.g. iodine.

Table 4.3 Dry deposition velocities (cm/s) for aerosols of chemically nonreactive compounds (IAEA, 1991a)

Radionuclides react with vegetation or soil and scavenging is very intense, consequently deposition is relatively high. An important factor is the type of the surface and/or vegetation. The rougher the surface, the higher the scavenging and thus the deposition. Deposition increases in the order: smooth surface, rough surface, small crops, large crops, forests. The difference between the lowest and highest deposition rates is a factor of 100 (Frissel et al., 1989). For larger crops sometimes the state of growth of the crop or leaf area index (LAI) is proposed as a measure for interception (IAEA, 1991b).

For accidental releases, rainfall can contribute substantially to total deposition. This is an important lesson which radioecologists have learnt from the accident at Chernobyl. With the exception of the immediate vicinity of Chernobyl, all severely contaminated areas were contaminated via wash-out.

The solubility of most radionuclides which are formed in nuclear reactors or other nuclear devices is low, consequently the major part of radionuclides released into the environment will finally accumulate in either the upper layer of soils in terrestrial systems or of sediments in aquatic systems. In general, Pu and Am hardly migrate in soil. Neptunium and Sr migrate over short distances, i.e. only 1 cm in a few years. The migration of Cs depends very much on soil characteristics. Usually it is very strongly adsorbed and migrates less than Sr. Exceptions to these rules are discussed later.

Migration can be described by a process which in chemistry is known as chromatography. The soil is considered to be divided into a number of layers. Radionuclides are assumed to be present in the top layer and partly in solution. Rain or irrigation water causes the solution to flow from the upper layer to the second layer. Part of the radionuclides adsorb in the second layer, depending on K_d value (Section 4.1.2). The studies of Frissel and Pennders (e.g. Frissel and Pennders, 1983) using the mass conservation equation cover a period of 20 years and indicate vertical migration velocities for Sr, Cs, Pu and Am of 0.3 to 0.5 cm per year.

In illitic soils, K, NH₄ and Cs ions are adsorbed within the clay mineral platelets of the illite. This adsorption is so strong that release occurs very slowly, only large amounts of K or NH₄ fertilizers promote growth of crops. However, desorption may counteract any reduction. This problem is associated with the composition of the soil solution and is discussed later. A very detailed analysis of Cs fixation is given by Cremers et al. (1988). Despite limitations, K_d is one of the best variables to indicate differences in leaching behaviour in soils of the temperate region. Table 4.4 provides default values from some radionuclides.


Radionuclide	Sand	Loam	Clay	Organic soil

Sr	1.4x10¹	2.0x10¹	1.1x10²	1.5x10²
Cs	2.7x10²	4.5x10³	1.8x10³	2.7x10²
Np	4.1x10⁰	2.5x10¹	5.5x10¹	1.2x10³
Pu	5.5x10²	1.2x10³	4.9x10³	1.8x10³
Am	2.0X10³	9.9x10³	8.1x10³	1.1x10⁵

The chromatography concept is rather complicated, and sometimes simpler concepts are preferred. A very simple approach is the apparent half-life concept. The assumption is that in a certain period half of the radionuclides disappear from a certain layer; the processes are not considered. A refinement of this concept is that of the residence time. The soil is assumed to be distributed into different layers. The radionuclide moves through subsequent layers, staying in each layer with a certain residence time. The advantage of the approach is that the processes which cause migration do not have to be considered; the disadvantage is that diffusion, dispersion and vertical transport cannot be distinguished.

Migration theories have two shortcomings. The first is that mechanical mixing is not considered. The effect of ploughing is obvious, but also the mixing effect of earthworms and trampling by cattle can be important in the upper soil layers. The second is incorporation of radionuclides in clay or other soil minerals which causes migration to decrease with time. Radionuclides produced by a natural chain reaction in Oklo (Gabon), some 1.7 billions of years ago, and running for about 100 000 years, migrated over much shorter distances than calculated with commonly used K_d values. The accident at Chernobyl occurred only 6 years ago, a rather short period to measure migration in soil. In almost all reports it is concluded that radionuclides are adsorbed in the upper few centimetres of the soil and migrate very slowly (Bokori et al., 1989). In areas with high rainfall a small part of the radioactivity may immediately have penetrated deeper, thereafter it did not migrate further (CEC, 1991). Because of increasing fixation with time; the velocity decreases with time, an example is described by Bunzl et al. (1988). A remarkable, so far unexplained, exception was reported by Pavlotskaya et al. (1991), who noted shortly after the accident at Chernobyl that in hydromorphic soils the major part of Pu was already present at a depth of 10 cm. The accident at Kyshtym in 1957 caused severe contamination of soils with ⁹⁰Sr. Most soils have been ploughed and no migration pattern was determined. Recent investigations on some non-ploughed areas showed that ⁹⁰Sr had reached a depth of about 15 cm, indicating a migration rate of 0.5 cm per year. However, the ⁹⁰Sr content in the upper layers of soil is lower than expected. It is possible that part of the ⁹⁰Sr leached to deeper soil layers. Detailed information is lacking and cannot be traced back (Molchanova, J. V., personal communication, 1990).

A special problem is that caused by hot particles which may constitute a reason for an increase of the available fraction of radionuclides with time (Alexakhin, in press).

Application of transfer factors is complicated by the thickness of the layer in which radionuclides are assumed to be present. A good choice should be the thickness of the rooting layer, but that layer is variable. The IUR recommends as standards for TF calculations average concentrations over 0

10 cm for `grass' and 0

20 cm for all other crops. If the contamination is very superficial, and for instance in a litter layer or sward, the IUR convention is not practical. Instead Bq per kg crop/Bq per m² has been used, particularly for Chernobyl data.

Table 4.5 shows some representative TF values. They are suitable for screening purposes, but the number of conditions which influence uptake is so large that for realistic assessment calculations, site and condition specific values have to be used. For Cs, time lag, i.e. the time elapsed since the contamination, is most important. A decrease of availability of Cs by 50 per cent in the first year after contamination is the most frequently reported value. Values range, however, between 30 and 90 per cent. Another problem is the large variation, partly caused by factors which vary from year to year depending on weather conditions, and local conditions. After three to five years the uptake drops a further 50 per cent. After 10 years usually only 10 per cent of the original availability remains; but this value certainly does not apply to all soils. An example of the reduction of Cs content of crops in the Gomel area of Belarus is shown in Table 4.6 (Grebenshchikova et al., 1991).


Nuclide	Crop	Crop part	TF value

Sr	Cereals	Grain	0.13
Sr	Fodder		0.95
Sr	Grass		1.3
Sr	Bean	Pod	1.2
Sr	Carrot	Root	0.46
Sr	Potato	Tuber, flesh	0.17
Sr	Green vegetables		2.3
Cs	Barley	Grain	0.03
Cs	Wheat	Grain	0.018
Cs	Alfalfa	Fodder	0.26
Cs	Grass		0.21
Cs	Bean	Pod	0.023
Cs	Potato	Tuber	0.1
Cs	Spinach		0.24
Np	Wheat	Grain	0.0023
Np	Potato	Tuber	0.0067
Np	Green vegetables		0.037
Pu	Wheat	Grain	0.00001
Pu	Potato	Tuber	0.00018
Pu	Green vegetables		0.00007
Am	Wheat	Grain	0.00002
Am	Potato	Tuber, flesh	0.00039
Am	Green vegetables		0.00066

Table 4.6 The reduction of the Cs content of crops with time observed near Gomel, Belarus (Grebenschikova et al., 1991)

The bond between Cs and illitic clays is very strong and can only be broken by an excess of K or NH₄ ions. This explains why an application of K-fertilizers can have different effects on contamination of crops. An excess of K in soil dilutes Cs ions which decreases uptake, the excess may cause desorption of fixed Cs which increases the uptake. Application of ammonium fertilizers causes the same effect. A complication is that ammonium ions may be formed by microbiological decomposition of organic soil material. Most arable soils are well drained, in which case ammonium is oxidized to nitrate which hardly influences Cs uptake. In wet soils, particularly if they are rich in organic material, ammonium is not oxidized and consequently, Cs remains more available to the solution phase. This situation applies to all soils with a sward or litter layer, and is important for forests and pastures of upland and alpine soils.

The availability of Sr also decreases with time, but the decrease is small and limited to a few per cent per year. In short-term experiments it may be overshadowed completely by annual variations. The reason for the decrease is that Sr ions are incorporated into the crystal lattice of soil minerals.

Another important factor which influences uptake is the pH. For important fission products, uptake increases with decreasing pH values. A factor of 10 can be obtained easily (Frissel and Pennders, 1991). A complication is that soil type, pH and organic matter content are related. The uptake of Cs from a sandy soil is about five times higher than from a clay soil; for a loam soil this factor is about 2. For well-cared agricultural soils, soil type and organic matter content describe the uptake sufficiently and the pH does not have to be considered separately (CEC, 1991). For less well-cared soils and semi-natural soils the pH will be the first parameter which has to be assessed to estimate uptake. After the accident at Chernobyl Soviet scientists applied three pH classes for their rehabilitation programme: 4.5

5.5, 5.5

6.5 and 6.5

7; the ratios of the Cs uptake of the classes are about 4:2:1 (Table 4.7; Prister, 1991). The interaction between soil type and pH is also important for the uptake of Sr. Jouve (1990) found that the uptake from sand soils is 5 times as high as from clay soils. In his prognosis model he included pH, K_d and exchangeable Ca content of a soil.

Depth of contamination, rooting depth and soil moisture regime form another set of related factors. It is obvious that uptake will increase if root activity increases in the contaminated layer. The presence of soil moisture activates root activity. In particular, in superficially contaminated dry soils this behaviour offers a possibility to influence the uptake of radionuclides. Subsoil irrigation will activate the root below the contaminated layer and decrease uptake. On the contrary, surface irrigation promotes uptake. N. Zezina (personal communication, 1990) succeeded in increasing Cs uptake by maize by a factor of 100 after applying sprinkler irrigation to a coarse sandy soil in an experimental field 10 km west of Chernobyl.

Table 4.7 Some soil-to-plant transfer values determined for¹³⁷Cs in the Kiev region of the Ukraine (Prister, 1991)


	Bq/kg per kBq/m²
Crop	pH range		pH range	pH range
	4.55.5		5.56.5	6.57.5

Wheat	0.5		0.2	0.5
Barley	0.3		0.1	<0.01
Lucerne	0.8		0.4	0.2
Clover	0.8	2.3	0.3	0.3
Potato	0.3		0.1	0.04
Onions	0.6		0.2	0.11
Cabbage	0.3		0.1	0.04

Flooding is responsible for changes in physical, chemical and biological properties of soils. It also results in modification of growth and development, particularly of the root system. An important example of the effects of flooding on radionuclide uptake is rice growing (Bourdeau et al., 1965; Myttenaere, 1972; Myttenaere et al., 1975), where besides direct contamination from deposition of fallout, the plants can absorb radionuclides from the soil or from the water in which they stand. Investigations have been carried out in the field using fallout radioisotopes as tracers, and under controlled conditions in lysimeters, glasshouses and growth chambers. In the field, covering had little or no effect on the levels of ¹⁴⁴Ce, ¹⁰⁶Ru, ¹³⁷Cs, ⁵⁴Mn or ⁹⁰Sr in the roots of flooded rice plants. In contrast, covering reduced ¹⁴⁴Ce in shoots to 28 per cent of the control, and ¹⁰⁶Ru, ¹³⁷Cs and ⁹⁰Sr to 75

77 per cent. ⁵⁴Mn was affected only slightly (down 4 per cent). The ¹³⁷Cs/⁹⁰Sr ratio in the grain hull was 1.6 whereas that for hulled rice was 10. For covered plants, the ¹³⁷Cs/⁹⁰Sr ratio in the hull was 3.1 whereas that in the hulled grain was 8.2, indicating a lower mobility of ⁹⁰Sr compared to ¹³⁷Cs within the plant. The major fraction of ¹³⁷Cs and ⁹⁰Sr found in the shoots and grain of flooded rice was due to indirect contamination, the two radioisotopes being equally available via this route.

The relative importance of soil and water in the indirect contamination of flooded rice with radiocaesium has also been investigated in model rice fields using a dual tracing technique. Concentration factors between plant and water and plant and soil (Table 4.8) varied with the plant part and the origin of the contamination, but were much higher for ¹³⁷Cs coming from water than for ¹³⁴Cs coming from soil.

Table 4.8 Concentration factors between plants and water (¹³⁷Cs) and between plants and soil (¹³⁴Cs)


	CF₁	CF₂	CF₁CF₂	CF
Plant part	Plant/water^a	Plant/soil^b		in nutrient solution

Roots	63	0.572	110	225
Leafy shoots	111	0.052	2135	55
Panicle minus grains	70	0.019	3684
Hull	45	0.010	4500	94
Caryopsis	20	0.005	4000	32

^aActivity per g dry matter
Activity per cc irrigation water
^b Activity per g dry matter
Activity per g dry soil

Another important factor is the adhesion of soil to crops. Even the presence of 10 g soil per kg dry crop equivalent to 1 g soil per kg fresh weight (for many green crops) causes a TF of 0.01 Bq/kg crop per Bq/kg soil. Strontium TF values range from 0.1 to 3, thus the increase of the TF due to soil adhesion is unimportant. Apparent TF values for Cs range from 0.01 to 0.3; soil adhesion therefore contributes significantly. Soil adhesion is mainly caused by rainsplash and in that case is limited to heights of about 40 cm. Rainsplash, together with wind erosion may, however, also be the cause of resuspension. It is difficult to give a default value for soil adhesion, as the value depends so much on local conditions. The IUR (1990) reported typical values of 10 and 4 g soil per kg dry crop for crops lower than and above 40 cm respectively. An extreme value is perhaps 250 g per kg soil.

The determination of soil adhesion values is rather difficult. One way is to determine the load of an element which is not or hardly taken up by crops (e.g. Ti, Sc). Plutonium concentration is sometimes used for this purpose. Alternatively experiments can be done with and without conditions which prevent resuspension. Resuspension must be prevented if reliable TF values are required for Pu, Am and Np, otherwise soil adhesion dominates the root uptake contamination. The TF values for transuranics shown in Table 4.5 and reviewed by Frissel (1990) are based on experiments which avoided soil adhesion. The same review shows that for field experiments higher TF values for Pu and Am were observed, suggesting that soil adhesion played a dominant role.

Only generic values for soil to plant transfer are needed to answer assessment questions in which only simple steady-state screening models are applied. In this case all processes leading to plant contamination can be included in one parameter which describes the maximum transfer from soil. Greater accuracy is required for models intended to assess dose or environmental concentrations. Such models should consider more transfer processes and should use best estimate values instead of maximum values. In particular, for transuranics, one should consider the direct and indirect uptake separately. The biological availability of both contaminants is expected to be quite different.

Once an area has been contaminated two major remedies exist, i.e. removal of radionuclides by removing contaminated soil or reduction of uptake. Removal of bare soil is not discussed here. An alternative method starts with temporary fixation of Cs and Sr by applying polyacrylate or carboxy methyl cellulose derivatives. Sowing of clover or grass produces a sod which contains the major fraction of the radioactivity and which can easily be removed by conventional equipment. Another effective way to reduce uptake is deep ploughing to bury radioactivity below the rooting zone. In districts of the Southern Urals which were contaminated with ⁹⁰Sr after an accident at Kyshtym, good results were obtained with ploughing to a depth of 50 cm.

Alternative methods consist of the application of fertilizers, lime, peat, zeolites and other compounds which may influence the uptake of radionuclides. The reduction factors which can be obtained are usually not greater than a factor of four. On poor soils, fertilization will boost crop production so much that contamination expressed per unit weight crop decreases considerably. However, total uptake of radiocaesium may be higher. For ⁹⁰Sr countermeasures which increase pH are useful. Table 4.9 shows the impact of applications of various products; lime is the most successful immediate measure. Ploughing and waiting appear to be the best measure.

Table 4.9 Impact of agricultural countermeasures on the uptake of ⁹⁰Sr on meadows observed near Gomel, Belarus (Grebenschchikova et al., 1991)


		kBq/kg grass fresh weight
		1st cut 1988	1st cut 1989	1st cut 1990

Control		9.25	3.33	2.33
Ploughed, N,P, K applied		2.41	1.66	0.33
Ploughed, N,P,K + zeolite		1.37	0.70	0.33
	(10 000 kg/ha applied)
Ploughed, N,P, K + phosphorgypsum		2.55	1.48	1.04
	(5000 kg/ha) applied
Ploughed, N,P,K + lime
	(5000 kg/ha) applied	1.30	0.67	0.56

The importance of animals, and more particularly of domestic mammals, with regard to the radioactive contamination of the food chain resides in the fact that edible products such as milk, meat and eggs are important items of the human diet. The general principles of metabolic behaviour of radionuclides in domestic animals and in free wild animals are similar and they need not be discussed separately.

In this section, the factors influencing the transfer of the most important radionuclides to the major food-producing animals are discussed as well as the metabolism and distribution in tissues, and secretion into milk and other products. As a result of the importance of milk in the human diet, particularly for young children, the transfer of radionuclides to milk of lactating cows has occupied a central place in many studies.

The important parameters for the transfer of radionuclides to and their metabolism in animals are listed in Section 4.1.4. In order to be metabolized, a radionuclide will have to enter the animal organism. Three possibilities need to be considered: absorption through the skin, inhalation, and oral ingestion. The superficial layer of the skin in domestic animals is impermeable to many substances even when they are applied in aqueous solution. An exception must be made for tritium which may penetrate through the skin when it occurs as tritiated water.

A damaged skin may lose its protective function to a certain extent. Appreciable absorption of radionuclides through the intact skin of domestic animals under field conditions is very unlikely to occur. Inhalation is the process of air movement in the respiratory system which consists of air passages leading to the lungs where the actual gas exchange of oxygen and carbon dioxide takes place. Radionuclides may enter the respiratory system either as a gas or adhered to particles of variable size. Particulate material will only reach the lungs if the particles are small enough, since larger particles will be trapped by the cilia of the epithelial cells lining the mucous membrane of the air passages, e.g. trachea, bronchi, bronchioli. Such trapped particles will be removed quite efficiently by the upward movement of the cilia. Particles larger than 5

10 µm in diameter are generally considered to adhere to the mucous membrane of the air passages and do not reach the lungs. When the particle size decreases, more particles will reach the lower respiratory tract and the lungs where retention may occur. Translocation of radionuclides from retained material or from inhaled gases to other tissues depends predominantly on their solubility. Rapid clearance with retention half-times of a few days or less have been observed for certain chemical forms of iodine and strontium. Much longer retention half-times have been found for plutonium oxide (Bair, 1960).

Absorption of radionuclides by animals via inhalation is of limited importance. Hvinden et al. (1964) showed clearly that the intake of radionuclides through inhalation was at least three orders of magnitude less than that from pasture. Since the respiratory tract is not usually an edible product for man, the significance of the inhalation route for food chain contamination is further reduced.

Oral ingestion of radionuclides with contaminated feed is the predominant route for uptake by animals. However, contamination of animal produce will occur only if the radionuclide under consideration can be absorbed from the gastro intestinal tract. Whether or not it will then be retained in edible tissues or secretions of an animal, depends on its physiological properties.

Differences in length, size and number of compartments of the digestive tract exist between carnivorous, omnivorous and herbivorous animals (Swenson, 1975). The digestive tract is much longer in herbivores which are very important species from the point of view of transferring radionuclides to man. Ruminants have a complex stomach system in which the very large first stomach, the rumen, plays a significant role in the digestion of fibrous cellulose-containing plant material. This function is performed in the large intestine, particularly the colon, of non-ruminant herbivorous animals such as the horse and rabbit. Although some absorption of ingested material may take place in the rumen, e.g. part of the ingested ¹³¹I,and also in the large intestine, it is generally agreed that the bulk of the digested food is absorbed from the small intestine. Absorption rates are different for different radionuclides. This is because the epithelial lining of the mucous membrane in principle is impermeable to the large majority of electrolytes and organic compounds in the lumen of the gut. Absorption may be a passive process such as simple diffusion, or a result of special mechanisms requiring energy which ensure rapid and efficient transport across the epithelial cells from the lumen of the gut into the blood capillaries and small lymph vessels in the wall of the alimentary tract. These different transport mechanisms may lead to different absorption rates for certain elements relative to their physiological significance. For example, calcium is absorbed by both active transport and simple diffusion, and strontium by simple diffusion only. This partly explains the much slower absorption of strontium compared with that of calcium. For many elements the epithelial cells of the gut wall are completely impermeable such as is the case for the rare earths. Near impermeability is found for barium and the actinides resulting in their very poor absorption.

Another point of interest is the movement of substances in the opposite direction, that is excretion from the blood into the lumen of the gut. This so-called `endogenous excretion' has been observed for many radionuclides and it may be quite significant when the actual absorption is high, e.g. for caesium.

Following the entry of absorbed radionuclides into the systemic circulation, their physiological characteristics will eventually determine their fate. They may be concentrated in a particular organ, e.g. in the thyroid for iodine, the bone for strontium and the bone or soft tissues for caesium. Radionuclides may also be secreted into milk of a lactating animal, or excreted into urine or faeces. The selective capacity of the epithelial cells of the mammary gland for secretion into milk is responsible for the different amounts of various radionuclides which can be recovered from milk.

Transfer of a radionuclide from an animal's diet into milk or meat is usually described by a transfer coefficient (Section 4.1.4). For milk, the symbol is F_m, and for meat it is F_f. The use of a transfer coefficient for predictive and other purposes has obvious advantages but caution is necessary because of its shortcomings. These are summarized in Section 4.1.4, and are discussed in detail by Ward and Johnson (1989). Table 4.10 (CEC, 1987) presents F_m and F_f values in milk and meat for radionuclides of greatest radiological importance for man under conditions of continuous intake.

Table 4.10 Transfer fators (d/kg or d/l) for caesium, strontium and the actinides in various animal products


		Nuclide
Animal	Product	^134,137Cs	^80,90Sr	Actinides (²⁴¹Am)

Cow	Milk	0.01	0.002	4 x 10^-7
	Meat	0.06	3 x 10^-4	2 x 10^-5
Calf	Meat	0.6	0.003	2 x 10^-4
Sheep	Milk	0.06	0.02
	Meat	0.3	0.002	2 x 10^-4
Lamb	Meat	0.8	0.02	0.002
Hen	Meat	4.0	0.02	0.005
	Eggs	3.0	0.3	0.008

After accidental releases, such as was the case at Windscale and Chernobyl, ¹³¹I (halflife about 8 days) is the radioisotope of immediate concern. In soluble form, it is rapidly and almost completely absorbed from the gastrointestinal tract. The absorption of iodine may begin in the first stomach of ruminants, and this explains the rapid appearance of ¹³¹I in milk when animals are out on pasture at the time of release. The metabolism of iodine is characterized by the ability of certain tissues to concentrate it. This is particularly true for the thyroid gland which incorporates iodine in its hormones. The mammary gland also has this capacity to concentrate iodine from plasma but differences in efficiency between species do occur. The mammary gland of the lactating cow is rather inefficient as compared to that in goat and sheep. A release of radioactivity into the environment usually is a single contamination event, at least initially. This means that animals will ingest decreasing amounts of radioisotopes each day. Maximum concentrations of ¹³¹I in milk will be found after 2 to 4 days and these will generally represent an F_m of 0.002

0.004 d/1 (Bustad et al., 1964; Handl and Pfau, 1987). The decrease in amounts of ¹³¹I ingested result not only from the physical decay of the radioisotope but also from its removal from pasture due to meteorological and other conditions. Experimental data on the amounts of iodine transferred to milk from goats after a single contamination event have not been recorded. When extrapolating data for continuous daily dosing experiments, 5 to 10 times as much ¹³¹I may be expected in goat's and ewe's milk than in the milk of dairy cows.

Strontium belongs to the alkaline earth metal series, of which calcium is the most important member from a physiological point of view. It is convenient to consider strontium metabolism in relation to that of calcium because dietary calcium levels have been observed to influence decisively the absorption of strontium from the digestive tract (Comar et al., 1961). Another reason is the homeostatic control of calcium which results in a high degree of constancy of calcium levels such as in blood plasma and milk. The transfer of calcium across biological membranes is usually more efficient than that of strontium which leads to discrimination against strontium. The value of this discrimination factor at the level of the digestive tract in the ruminant and also in other animal species is about 0.2, and that at the level of the mammary gland is 0.5. This means that strontium is transferred to milk about 10 times less efficiently than is calcium. In practice, most values of the observed ratio between strontium and calcium levels in milk and diet vary between 0.09 and 0.16, although quite important fluctuations have been observed, e.g. from 0.04 to 0.20 (Van den Hoek, 1989). Satisfactory explanations to account for these differences could not be given in all instances.

After absorption from the digestive tract, strontium, like calcium, may be incorporated into bone from which it is released very slowly. There is a wealth of information on strontium metabolism in bone because of the radiation hazard to bone marrow in the human being. The significance of the contribution from deposits in bone to contamination of other products has not yet been quantified.

In an experiment simulating an accidental contamination event while cows were out on pasture, peak levels of strontium in milk were found 3

4 days after contamination, representing an F_m of 0.0005

0.0007 d/l (Van den Hoek et al., 1969). After reaching their maximum these levels decreased with half-lives of about 10 days. It may be expected that administration of excess calcium will lead to a decrease in strontium absorption. This was reported to be the case in several experiments.

Caesium, like potassium and rubidium, belongs to the alkali^. metals. Up until the accident at Chernobyl, the contamination of animals with caesium was considered generally of lesser importance for contamination of the agricultural food chain than that of iodine and strontium, and caesium metabolism received proportionally less attention. The high caesium levels in sheep in some upland areas in the United Kingdom and in Ireland, and also those in free-ranging wild animals such as reindeer, moose and deer in Scandinavia and in some alpine regions as well as the persistence of these high caesium levels have prompted a great deal of experimental work. Our knowledge of caesium metabolism in animals has extended considerably since 1986. Caesium is taken up and metabolized very easily by mammals. Like potassium it is distributed in the soft tissues of the body (muscles and organs) and it occurs intracellularly. However, there is not the same degree of interdependence between caesium and potassium, as there is between calcium and strontium.

Absorption of caesium from the gastrointestinal tract varies quite significantly between animal species, and is influenced by a number of other parameters, the effects of which are not always understood. Apparent absorption of caesium in monogastric animals, such as the pig, may be virtually complete whereas in polygastric animals usually it is not much higher than 60 per cent. Absorption will be greater when caesium is ingested in soluble form, e.g. as CsCl in tracer experiments or in milk. Other factors such as the crude fibre content of the feed and the amount of clay particles ingested have been reported to reduce the absorption of caesium. After absorption, caesium is distributed readily and fairly uniformly into the soft tissues. At equilibrium it occurs in approximately similar concentrations in muscular tissue and in several organs, notably the kidney. Its levels in liver and spleen are lower. Caesium concentrations in muscles and organs of young animals are generally higher than in those of adult animals for the same intake. The higher availability of caesium in the milk that young animals drink, a greater capacity of their digestive tract for the absorption of caesium (and of some other radionuclides as well), a higher tissue retention of caesium and perhaps other factors may be responsible for the higher caesium levels observed in many young animals. This is illustrated by the different F_f values for lamb and ewe muscle reported for different experimental conditions of caesium intake (Howard, 1989).

Caesium is also readily secreted into milk. Important variations in milk levels for the same intake of caesium were reported after Chernobyl (i.e. F_m = 0.002

0.02 d/1). Typical values in cow's milk lie around 0.008 d/l for continuous intake of caesium. These values are 0.1 d/l for the goat and 0.03 d/l for sheep. After accidental contamination, the milk levels will be lower because animals are ingesting decreasing amounts of caesium each day. A value of 0.0026 d/l was reported in cow's milk in May 1986 (Handl and Pfau, 1987). This compares quite well with values of 0.0031

0.0045 d/l which had been obtained in a simulation experiment under natural conditions (Van den Hoek et al., 1969).

Although a number of radionuclides may be released under accidental conditions, transfer to man via animals is usually of limited significance either because of short physical half-life, e.g. ⁸⁶Rb or as a result of very low absorption from the digestive tract, e.g. ⁹⁵Zr, isotopes of ruthenium, cerium and other elements of the lanthanide series.

After the Chernobyl accident, ^110mAg was reported to be present in the livers of cows and sheep. Noble metals are absorbed very poorly from the digestive tract but their absorption may be enhanced by the formation of complexes, which is more likely to occur in polygastric animals. Silver-110m is not metabolized actively after absorption. Most of the silver in the plasma is filtered out by the liver. This explains the relatively high levels of ¹¹⁰mAg in liver tissue and its virtual absence in muscular tissue and milk (Beresford, 1989).

The transuranic elements (plutonium, americium and neptunium) occupy a special place because of their great radiological significance for man. This is particularly true for plutonium which has been studied in greater detail than the other elements. Absorption of transuranics from the digestive tract, although generally very poor, may vary according to the chemical state of the element, the age of the animal and other factors. In adult animals, typical absorption values lie between 0.01 and 0.001 per cent; these values may be higher in young animals by one to two orders of magnitude. After absorption, actinides are preferentially deposited in bone and liver in variable fractions although the smaller fraction is usually found in liver. Their transfer into milk is very poor and variable, with F_m in the range 1 x 10^-8 to 1 x 10^-5 d/l.

Evidence was presented in several studies carried out during the 1960s that farm management practices might influence quantitatively the transfer of deposited radionuclides into animal produce (Hansen et al., 1964). As a result of better pasture management and adequate use of fertilizers, a higher grass yield could be obtained. This would lead to a greater dilution of the radionuclides deposited per unit area, although the greater interception of radionuclides by the pasture would presumably counteract partly the diluting effect. The net effect, however, was a reduced intake of radionuclides by grazing animals.

The contamination of large areas of land with caesium isotopes after the Chernobyl accident has drawn attention to the importance of plant selection by free-ranging animals (Howard, 1989). Seasonal changes in dietary composition led to unexpectedly high caesium levels in sheep and also in wild animals. For example, the increased intake of fungi by roe deer and other species in autumn, led to peak values of caesium in these animals during the hunting season (Johanson et al., 1990). Other dietary changes such as higher intake of lichen, heather and other plants have had similar effects.

The ingestion of soil by animals may be a significant source of radionuclides in certain situations. Whether or not the ingested radionuclides will also be available for absorption by the animal will depend primarily on how firmly they are bound to soil particles. Specific soil characteristics such as clay content are probably decisive for the availability of caesium for animals.

Countermeasures are used to prevent animal produce from becoming unfit for human consumption. Milk, contaminated with ¹³¹I, may be converted into milk powder, and the milk powder can be stored until the radionuclide has decayed to acceptable concentrations. For radionuclides other than ¹³¹I attempts are made to prevent absorption from the digestive tract. The secretion of strontium into milk may be reduced by increasing the calcium intake of the animal's diet or by the application of fertilizers containing calcium to pasture (Van den Hoek and Binnerts, 1967). The absorption of caesium isotopes can be efficiently reduced by including ammonium ferric cyanoferrate (a Prussian Blue derivative) in the diet (Unsworth et al., 1989). Inclusion of bentonite and other clay minerals may also reduce significantly the absorption of caesium from the alimentary canal.

Only very scattered information is available on the behaviour of radionuclides within tropical agricultural ecosystems, presumably reflecting the more limited and relatively recent development of a nuclear industry in such areas, compared with temperate regions. In view of the paucity of data, pathways of radionuclide transfer in tropical agricultural systems are considered here in generic terms, taking into account soil characteristics and relating these to knowledge of transfer processes in the temperate world.

Ecosystems in the tropics have an extremely wide range of soils and climatic conditions. Thus, depending on altitude and location, ecosystems may range from desert to rain forest to snowline communities. However, this section considers only those agricultural ecosystems which are restricted to tropical zones. The classical notion of a tropical soil involves the concept of a highly leached, well-weathered mineral substratum which has evolved over a period of some millions of years. While there are some specific tropical soil weathering processes, such as laterization, differences from temperate soils in many cases lie in the degree and intensity of these. However, the main difference from temperate zones is the major role of termites in soil development in many areas, which may be considered as comparable to the bioturbation induced by earthworms (Russell, 1973). Lack of disturbance during the soil development period may result in soil profiles being extremely deep. These profiles are commonly characterized by the development of deep porous oxic horizons, in which chemical alteration is so complete that the original parent material cannot be identified (Birkeland, 1974). Long and intense periods of weathering result in the almost complete depletion of cations and silica, commonly rendering the principal fabric of the soil as quartz grains, strongly cemented by oxides and hydrous oxides of iron and aluminium. The silt to clay ratio is commonly below 0.15, although the clay fraction is dominated by 1:1 kaolinitic minerals (Russell, 1973). The dominance of these minerals renders tropical soils relatively unfertile (Pitty, 1979). However, the cation exchange capacities of the soils are typically over 10 meq/100 g of clay and it is postulated from this that other 2:1 clay minerals (including illite) must usually be present.

Biological productivity in tropical soils is high, with a rapid turnover of organic matter, resulting in very rapid nutrient cycling. Infiltration of water is often at a high rate in tropical soils, resulting in the intense leaching which is associated with the development of the oxic horizons noted above. Loss of nutrient anions, such as nitrate, can therefore be considerable, despite the domination by hydrous oxides which result in a much higher anion retention capacity than in temperate soils, whereas for cations the converse occurs. Tropical soils tend to be acid, with penetration of low pH to deeper horizons than for soils in temperate regions, due to high cation uptake from depth by roots of the natural vegetation.

Soil loss due to rapid runoff is a problem in many tropical areas, but this can be mitigated by modifying agricultural practices, to avoid the development of rills and gulleys. A large proportion of agricultural land in the tropics is derived from rain forest, which is often cleared with minimal attention given to subsequent adverse environmental consequences in the form of erosion by wind and rain. Thus development of agriculture in tropical areas could potentially give rise to large-scale resuspension and mobility of radionuclides if contaminated forest lands were cleared. On the other hand, removal of timber could result in transport of large amounts of radioactivity out of a contaminated ecosystem; a large proportion of nutrient elements is locked up in standing biomass, because of highly efficient nutrient cycling.

While many radionuclides may be released into tropical ecosystems under different circumstances, a limited number of these are of particular interest in view of their radioecological significance over human life-span and their importance with respect to discharges by the nuclear industry. Radionuclides of particular interest include isotopes of caesium, strontium, technetium, actinides, and miscellaneous fission and activation products of the transition metal and lanthanide series, such as ruthenium, cerium and cobalt.

It is well known that caesium can become strongly fixed into specific clay minerals, the degree of fixation depending on the nature of the clay mineral concerned. Wahlberg and Fishman (1962) reported that caesium adsorption onto clay minerals could be ranked in the order montmorillonite > illite > halloysite > kaolinite. As kaolinitic clays dominate in tropical soils, it can be assumed that fixation of caesium will be generally lower than occurs in most temperate soils, in which illitic minerals are more prevalent. Caesium may remain more available for cycling through soil

plant systems in the tropics than in other areas, as unfixed radionuclides are readily absorbed by roots. In view of the heavily leached nature of many tropical soils, reduced concentration of potassium may lead to increased caesium availability (Shaw and Bell, 1991). While deep penetration of caesium may occur within the profiles of lateritic soils, there are reports of fallout ¹³⁷Csbeing restricted mainly to the top 15 cm of soils in different parts of India, over the 1966

70 period (Mishra and Sadavisan, 1972). Similarly Lowe (1978), reported strong ¹³⁷Csfixation in the top 10 cm of `claylike' soils in the Malayan Peninsula.

Strong eluviation of clay minerals has been reported from surface horizons of monsoon-affected soils on the Ganges Plain (Tomar, 1987). This depletion of the surface horizons may lead to Cs remaining in a relatively unfixed state and hence more available for biological absorption. There is evidence of the monsoon in itself influencing ¹³⁷Csuptake, with D'Souza and Mistry (1980) demonstrating increased concentrations in foliage of grasses growing on a high cation exchange capacity soil in Bombay in the monsoon season. D'Souza and Mistry stated that the observed greater uptake may, in all probability, be due to the greater availability of radionuclides from soil under conditions of high moisture status.

Strontium, a divalent element which can undergo simple cation exchange, shows less fixation within soils than does caesium. As a result, it has a marked degree of environmental mobility, exacerbated by the depletion of other basic cations in the surface layers of tropical soils. Evidence of this is provided by Mishra and Sadavisan (1972) who showed fallout ⁹⁰Sr penetrating to greater depths than ¹³⁷Csin a range of soils in different regions of India. Technetium as pertechnetate may be more readily retained in leached tropical soils as a result of their relatively high anion exchange capacity.

Soils are known to form the major environmental reservoir for Pu, whose major chemical sinks are organic and oxide fractions (Livens et al., 1986). Further, the chemistry of Pu and other actinides is strongly controlled by redox conditions in the soil. The preponderance of metal hydrous oxides in highly leached tropical soils may be expected to provide a major chemical sink for Pu, although in soils affected by monsoon rainfall the chemistry may be made significantly more complex in a manner not clearly understood. The degree of overall biological incorporation of the actinides is generally much less than for Cs, Sr and Tc, but any accumulating layers of dead organic material may again provide a significant environmental sink. The influence of rapid organic matter turnover on biological availability of organically bound actinides remains to be elucidated.

Very few studies have been performed on soil to plant transfer with crops which are only grown in tropical regions. One such study was reported by Delmas et al. (1987), in which uptake of ⁹⁰Sr and ¹³⁷Csinto peanut, banana and pineapple from three African soils was examined in an experiment in Senegal. The substrata concerned were a hydromorphic, a tropical ferruginous (lateritic) and a volcanic soil, respectively. However, Delmas et al. (1987) only gave results for peanut growing on the lateritic soil.

Scarcely anything is understood about the behaviour of radionuclides deposited directly onto foliage of tropical plants, and it is more difficult to make generalizations here than for contamination of soils. D'Souza and Mistry (1980) reported the pattern of field loss of ¹⁰⁶Ru, ¹²⁵Sb, ¹⁴⁴Ce and ¹³⁴Cs applied to hybrid napier grass (Pennisetum purpureum) which was then exposed in the field in Bombay. The rates of loss of the four radioisotopes were fairly similar, with a retention half-life of 12

16 days, on a concentration basis which is comparable to that reported in temperate regions. However, as is pointed out by the authors, there may be a marked seasonal influence on field loss in tropical regions, depending on when contamination occurs in relation to the monsoon and other rainy seasons. Rainfall of monsoon intensity is rarely experienced in temperate zones and its influence in contaminating crops via wet deposition, in removing foliar contamination, and in soil splash remain areas of ignorance.

The leaves of trees in moist tropical regions often bear large numbers of epiphyllae, such as algae, lichens and bacteria, on their surfaces. These epiphyllae are adapted to receive their nutrition from the atmosphere and consequently have been shown to accumulate large quantities of fallout radionuclides. Thus, Lowe (1978) reported that canopy leaves on Malaysian trees contained four times more ¹³⁷Cs when they were covered in epiphyllae. Similar evidence showing a major role of epiphyllae in increasing foliar levels of fallout caesium has been produced by Kline et al. (1973) for natural vegetation in Puerto Rico; this phenomenon was considered to explain the increase in ¹³⁷Cs levels observed at higher altitudes on the island. In the case of a nuclear accident, the presence of epiphyllae on the foliage of agricultural arboreal crops could have major radioecological significance.

In conclusion, there is an urgent requirement for both field and controlled experimental studies on the behaviour of radionuclides in tropical agricultural ecosystems, considering the various combinations of soils, crops and climatic conditions which pertain in these ecosystems.

4.3 SEMI-NATURAL ECOSYSTEMS

The Chernobyl accident in April 1986 exemplified the importance of forest ecosystems as a link in the transfer of radioactivity to man (Desmet and Myttenaere, 1988). Before the accident, several studies had considered the transfer of radioisotopes (particularly ¹³⁷Cs) in forest ecosystems after introduction as tracers (Auerbach et al., 1964) or after contamination of the soil (Auerbach, 1987). Nevertheless, the transfer atmosphere-canopy, the first important step after an atmospheric release (Ronneau, 1990) had not been studied.

Several mathematical models had been developed to predict the cycling of various radionuclides (Groom and Rasdale, 1980; Garten, 1980; Jordan et al., 1973; Olson, 1965; Prokhorov and Ginzburg, 1971). In 1986, Auerbach (Auerbach, 1987) had summarized results on the behaviour of radionuclides in three forest ecosystems and emphasized the soil compartment as the major accumulator of caesium. Most of the radioactivity transported to the canopy leaves was shown to be leached, but in the case of radiocaesium the analysis of first fallen leaves and of the throughfall water revealed that part of the radioactivity was returned to the tree before leaf fall (Auerbach et al., 1964).

These results underlined the need to study the biogeochemical cycle of deposited radionuclides within forest ecosystems. Also, since forests appeared to be very effective accumulators of atmospheric radioactivity (Sokolov et al., 1990; Sombre et al., 1990; Tikhomirov, 1990; Van Voris and Dahlman, 1976) and since they occupy slightly more than one third of the land surface, studies of the behaviour of the most important radionuclides released in the event of an accident are required. Moreover, understorey plants and wild animals which inhabit forests may be highly contaminated after the redistribution of radioactivity trapped by the canopy. The consequence of secondary transfers is a possible contamination of food and exposure of man and populations to ionizing radiation.

The use of forests for recreation may also expose man to direct gamma shine through contact with the ground, trees and other vegetation. The industrial uses of timber may also be a source of irradiation. The analysis of risk associated with deposition of radioactivity into forest ecosystems requires a forest dose assessment model to evaluate the importance of the forest ecosystem as a source or sink of radioactivity (Berg, 1990).

4.3.1.2 Radioactivity transfer and cycling in forest ecosystemsmain pathways

The above processes provide a basis for developing simulation models for forest ecosystems. The biogeochemical cycles of some radionuclide analogues are well established and may assist in designing scenarios for radionuclide transfer. A review of potassium and calcium biochemical cycles in forest ecosystems throws some light on the possible behaviour of ¹³⁷Cs and ⁹⁰Sr in these environments (Van der Stegen de Schrieck and Myttenaere, 1992).

The majority of radionuclides are transported in the atmosphere as aerosols with various physico-chemical properties and this gives rise to different deposition mechanisms (Section 4.1.3).

The final step in the deposition process occurs after particles or gases have been brought close to the canopy surface by mechanisms such as diffusion, gravitational settling and impaction.

The efficiency of capture depends upon particle diameter and wind speed in a somewhat complex manner, with a minimum deposition velocity of the order of 0.1 cm/s for particles of about 0.2

0.5 µm (Ronneau et al., 1987). Moreover it has been shown that deposition velocity is higher for Scots pine than for grass. Recent results (Tveten, 1990a) show that the deposition velocities for trees are about one order of magnitude greater than onto smooth surfaces. For single trees it was also shown that the local deposition velocity is proportional to bulk mass in much the same manner as for grass.

Forest canopies may also be contaminated by radioactivity in rain, and part of the radioactivity trapped during the previous dry periods is leached mostly by throughfall. Rain acts as a powerful cleaning agent of both the atmosphere and canopies at the expense of soils which are the ultimate sink of air pollutants. Normally, during a full meteorological cycle, dry deposition accounts only for a small fraction (few per cent) of the wet deposition.

The wet deposition of radioactivity from Chernobyl was studied in a spruce forest in Belgium using gauges in clearings and under spruce canopies (Ronneau, 1990; Ronneau et al., 1987). Table 4.11 gives accumulation factors (AF) observed after rain scavenged the atmosphere on 4 May 1986. The AFs were close to unity except for radiocaesium (0.2); such a low value was never observed in a several-year study on non-radioactive pollutants. This is attributed to the solubility of caesium which allows needles to absorb it very efficiently.


	Bulk deposition (Bq/l)
Radionuclides	CL	CAN		AF(CAN/CL)^a

¹⁰³Ru	226	180		0.80
¹³¹I	584	620		1.06
¹³²I	763	965		1.26
¹³²Te	805	1075		1.34
¹³⁴Cs	97	19		0.19
¹³⁶Cs	23	5	.4	0.23
¹³⁷Cs	178	38		0.21
¹⁴⁰La	29	21	.6	0.75

^aAccumulation factor (AF) =	radioactivity collected under trees
	radioactivity collected in clearings

Deciduous trees were less at risk from the Chernobyl episode because, given the time of year, they were less able than the evergreen species to intercept water.

Observations two years after the accident gave different results (Table 4.12) because at that time leaching of the incorporated radiocaesium was responsible for the high values obtained under canopies. Moreover, determinations made on rain collected under spruces and oaks showed that there was a good correlation between the deposition fluxes of both radiocaesium and stable potassium, suggesting that potassium may be used as an indicator of the behaviour of ¹³⁷Cs (Olson, 1965).

Experiments conducted in controlled conditions with thermogenerated aerosols have shown that the interception was of the order of 80 per cent. Other experiments conducted recently have given higher figures (~99 per cent). At Chernobyl the crown of trees trapped 60 to 90 per cent of the radionuclides deposited (Sokolov et al., 1990).

Retention properties are a function of the species (biotic layer) and of the inactive particles (natural or anthropogenic) previously deposited and trapped onto the surface of the leaves (abiotic layer). The surface of leaves is covered with particles of various sizes and the possibility of some interaction between radiocaesium and the soil-derived fraction of the abiotic layers has been suggested (Sombre et al., 1990) to justify its preferential retention.

The weathering half-lives of intercepted radiocaesium have been calculated for oak and pine trees contaminated with a fallout simulant (¹³⁴Cs adsorbed on 88

175 µm quartz particles; Witherspoon and Taylor, 1969) and for deposited thermogenerated radiocaesium (Al Rayes et al., 1988; Sombre et al., 1990). The first experiment showed the influence of species (oak leaves lost caesium quicker). In both cases the retention curves showed that at least a two-compartment model has to be used. Spruces contaminated by thermogenerated aerosols and then exposed to natural rain showed two weathering rates (Tb₁ 5 d ~ 50 per cent of the activity; Tb₂ 50 d) (Sombre et al., 1990). These results were confirmed by determination of caesium retained by leaves and by results obtained near Chernobyl four years after the accident (Sokolov et al., 1990; Tikhomirov, 1990). The non-removable caesium may penetrate leaves and become available for transfer to other parts of the tree.

The translocation in woody plants has been examined by inoculating vegetation with radioisotopes. These experiments, as well as data obtained for stable elements (nutrients or not), have shown that potassium as well as caesium is very well transported in the plant and may be readily transferred to wood before winter. On the contrary, calcium as well as strontium is stored in older tissues and no recycling within the tree occurs (Van der Stegen de Schrieck and Myttenaere, 1992).

These observations may explain why, several years after deposition, radiocaesium levels in leaves do not decrease or might even increase (Ronneau, 1990). The last observation might also be due to an absorption by roots of activity deposited or redeposited on soil.

In the long term, deposited radionuclides are removed to the forest floor where chemical adsorption, ion exchange and complexation occur on contact with the organic rich humic and moss layer (Witherspoon and Taylor, 1969).

The litter of temperate woodlands on acid soils tends to accumulate in thick layers in which three horizons may be distinguished:

In temperate woodlands the annual litter fall may take 3

5 years to decompose (Schell and Myttenaere, 1989). On the contrary, in tropical forests the leaves disappear completely within six months, whereas in coniferous forests the needles may take more than a decade in cool temperate or boreal sites (Packham and Harding, 1982). In extreme cases litter may be broken down and incorporated in the mineral soil within a year (mull) or it may accumulate for many years on top of the mineral substrate (mor) thereby returning nutrients (Witkamp and Van der Drift, 1961). The fate of radionuclides in forest soils may thus vary with climate conditions.

Post Chernobyl observations (Thiry et al., 1990) have confirmed that forest soils accumulate and strongly retain radionuclides, particularly caesium, in the upper organic zones. The thickness and nature of the organic matter present at the surface of the floor seems to be the main factor which influences the vertical distribution of the deposits.

Such results may explain maximum concentrations of radiocaesium found in some compartments of the understory. Only a few data obtained in situ give information on bioavailability of radionuclides deposited on multilayered soils. Recent studies have underlined the necessity to review the methodology (Andolina and Guillitte, 1990) as well as the concepts and expression modes (Thiry and Myttenaere, 1991) for radionuclide behaviour in multilayered soils. Thiry and Myttenaere (1991) highlighted the potential for misinterpretation of activity measurements for forest soils if based on expression by weight (the different layers having a different density). On the other hand, the calculation of a K_d based on the soil solution activity and the volumetric soil activity gives a better indication of bioavailability and mobility of radionuclides in multilayered soils. This confirms the fact that estimates of soil-to-plant transfer require a better understanding of chemistry of soil solution (Nisbet and Lembrechts, 1990).

Radiocaesium half-lives in the top 5 cm of different forest soils were given by Belli et al. (1990). For some sampling sites the data showed a great spatial variability which may be accounted for by different physical and chemical characteristics of the soils.

Tables 4.13 and 4.14 give the ¹³⁷Cs activity of the components of a pine forest situated 15 km from the Chernobyl nuclear power plant (Thiry et al., 1990). The understorey is much more contaminated than the overstorey. A calculation based on the activity of the different compartments of the same forest (soil included) showed that four years after the accident most of the ¹³⁷Cs radioactivity was found in the organic layer of the soil.

Table 4.13 Distribution of ¹³⁷Cs activity (percentage) in a contaminated pine forest at Bourakovka, Chernobyl in 1990


	¹³⁷Cs (MBq/ha)	Percentage

Needles	47
Twigs	189
Bark	201	1.45
Wood	108
Roots	?
Litter	1612	4.29
Soil
0 (02 cm)	30 350	80.72
Ah (25 cm)	4048	10.76
B (510 cm)	1045	2.78

Table 4.14 Distribution Of ¹³⁷Cs in a contaminated pine forest at Bourakovka, Chernobyl in 1990. Data expressed in Bq/g dry matter

Contamination levels of mushrooms vary with the soil layer explored by the mycelium. Moreover the migration of caesium in the soil profile with time may determine the value of any individual species as a bioindicator. Lichens, mosses and fungi which grow on forest soils and which explore different layers of the soil surface may be differently contaminated; this means that a species may be a good bioindicator at only a certain time after an accident.

Such restrictions require from the scientist a very good knowledge of the dynamic aspects of the movement of radionuclides in the soil and of the growing habits of the living forms considered.

After the Kyshtym accident a reduction was observed in the contaminated territory in the populations of invertebrates with a long life cycle and an extended phase of development in the forest litter (Spirin et al., 1990). These effects were most marked for earthworms, myriapoda and mites and at contamination levels above 3.7 MBq/m. Less marked effects were observed for flying insects and for invertebrates with an exoskeleton even though they spend most of their lives on the surface of the forest litter. Birds and mammals received lethal doses in autumn and winter (1957

58) if they lived permanently in areas where contamination exceeded 37 MBq/m. For mammals, the most notable effects were observed in mouse-type rodents (increase in mortality and decrease in life expectancy for 37 MBq/m). After 15 years populations were comparable to those of other animals. No similar effects were observed in other animal populations.

A substantial proportion of the areas affected by Kyshtym and Chernobyl were covered by forest (respectively 20 per centand 50

60 per cent). One of the main differences why the two accidents had a different impact on the environment was the time of occurrence (Kyshtym 29 September 1957, autumn

Chernobyl 26 April 1986, spring). Radioactive contamination occurred in Kyshtym when plants and many animal species were in their physiologically dormant state during which the process of radiation damage and regeneration were slowed down. Deciduous trees were less at risk after the Chernobyl episode because, at that time of year, they were less able to intercept radioactivity than were evergreen species (Arkhipov et al., 1991).

The development of a general model of transfer for a defined radionuclide through a generic forest may be extremely difficult. A descriptive model of radionuclide pathways through a forest ecosystem has been developed recently (Van Voris et al., 1990). However, many of the processes affecting radionuclide behaviour in forest ecosystems are not well understood and the models developed previously reflect this lack of knowledge.

The goal of future studies should be to determine and model the relevant processes involved in the transfer of radionuclides within and out of the forest. One of the main questions that such as a study has to answer is: How long will the forest maintain a level of radionuclides harmful to man?

The fate of man-made radionuclides in arid and semi-arid environments (defined collectively herein as arid sites), takes on a special significance when considering that over 30 per cent of the Earth's land surface, 22 per cent of the world's arable land, and 16 per cent of the world's population is tied to low precipitation areas. While there is an enormous amount of literature on radionuclide behaviour in the environment this discussion is based on research sponsored by the US Department of Energy (DOE) on plutonium in arid areas of the US (US DOE, 1980, 1985).

More than a third of the world's land surface is classified as arid (receives less than 250 mm of annual precipitation) or semi-arid (250 mm to 500 mm of annual precipitation). The continental interior deserts can be hot or hot and cold depending upon elevation and latitude. They are subject to both summer thunderstorms and winter frontal precipitation patterns where snow can be an important part of the annual water balance.

Of interest in radionuclide transport processes in arid areas are the relative amounts of wind and/or water erosion. Wind erosion is prevalent in areas where vegetative and rock cover is sparse or absent and where winds blow over dry surfaces. Under these conditions, much soil can be moved and, if contaminated, result in significant transport of radionuclides.

In many of the continental deserts, and especially in their associated semi-arid areas, high-intensity thunderstorms can result in flash-flooding and the resultant movement of large amounts of water and sediment in affected areas. This is a very efficient means of transporting sediment-associated contaminants. In areas subject to heavier winter rains and snow, runoff and erosion can occur over extensive areas. While the severity of these winter events may be less than their summer counterparts, their areal extent may be much larger and they may result in significant movement of sediment and contaminants.

Snow and snowmelt in the colder arid and semi-arid areas may result in more soil water recharge and relatively more movement of sediment-associated contaminants into the soil profile than in areas dominated by rainfall. The significance of these different processes in transporting radionuclides will depend upon specific site and radionuclide source conditions.

Most sources of radionuclides presently in arid ecosystems in the US were initially deposited on the ground surface decades ago as a result of single contamination events or as chronic events over long periods of time.

Results of field studies in arid and semi-arid areas in the western US show quite convincingly that soil will eventually be the primary repository of plutonium and many other radionuclides with long half-lives. For example, the percentage of the total plutonium inventory associated with the soil component in four different arid site ecosystems was, in all cases, greater than 99 per cent, despite the fact that the source of the plutonium at each site was different.

The relative amount of plutonium in the surface 2.5 cm of soil from four study areas in New Mexico, ranged from 4 to 41 per cent of the total measured in the sampled profiles (Table 4.15). Mortandad Canyon, at the time the data in Table 4.15 were collected, was actively receiving treated radioactive liquid effluent from Los Alamos research and development activities. Note that the plutonium inventory in sediment profiles was about equal in the various depth increments. In contrast, over two thirds of the inventory of plutonium in sediments from Acid Pueblo Canyon, which had not recieved liquid effluents for about 20 years when the data in Table 4.15 were collected, was in the 12.5

30 cm depth profile. Depletion of plutonium in the near surface profiles in Acid Pueblo Canyon reflects the combined influence of runoff, erosion and downward movement into the alluvial soil. At the Trinity Site over a 23-year period, the content in the 0.2

5 cm depth profile decreased by a factor of about 40 (Table 4.16).

The highest concentrations of many radionuclides in soil are usually associated with the smaller particle size fractions. Even though concentrations of radionuclides may be much higher in a particular soil size fraction, the total inventory of radioactivity in that fraction will depend on its mass relative to the whole soil. For example, despite the higher concentrations of plutonium in the silt-clay fraction of Los Alamos alluvial soils, less than 15 per cent of the total plutonium inventory in whole soil was in this fraction. In contrast, the silt-clay fraction at a Trinity Site location not only contained the highest plutonium concentrations, but it also contained over 70 per cent of the total soil plutonium inventory.

Table 4.15 Mean percentage plutonium inventory in soil profiles from Los Alamos and Trinity Site study locations in New Mexico (adapted from Hakonson and Nyhan, 1980)


Trinity Site ^a			Los Alamos ^a

Depth (cm)	Area GZ	Area 21	Depth (cm)	Mortandad	Acid Pueblo

02.5	29 (0.78)^b	41 (0.46)	02.5	20 (0.44)	4 (0.76)
2.55.0	18 (0.72)	19 (0.63)	2.57.5	36 (0.23)	10 (0.48)
510	21 (0.81)	6 (0.88)	7.512.5	22 (0.55)	20 (1.3)
1015	15 (0.67)	8 (0.92)	12.530	22 (0.79)	66 (0.18)
1520	10 (1.1)	9 (0.95)
2025	7 (l.3)	7 (l.0)
2533	ND^c	10(l.2)

^an = 8 for Trinity Site data; n = 10 for Los Alamos data.
^bValue in parentheses is the coefficient of variation, CV (standard deviation/mean).
^cNot detectable.

Table 4.16 Comparison of plutonium concentrations (kBq/m²) in surface (0

2.5 cm) soils from Chupadera Mesa as a function of time after the atomic bomb test at Trinity Site in 1945 (from Hakonson and Nyhan, 1980)

Soil will serve as the primary reservoir of many long-lived radionuclides in arid environments and complex spatial, chemical, and physical relationships will determine the mechanisms of transport and their relative importance. Wind and water sort soil particles during the detachment, transport, and deposition phases of erosion. Because radionuclide concentrations can be strongly related to soil particle size, there is a potential to enrich radionuclide concentrations in eroding soil. This potential for enrichment can affect the long-term redistribution of radionuclides and their transport to biological components of ecosystems.

Despite the host of chemical, biological and physical factors which can modify the physiological (chemical) availability of radionuclides and subsequent transport to internal plant tissues, field studies suggest that contamination of foliage surfaces with soil particles containing the radionuclides is a major transport mechanism under many arid site and radionuclide source conditions (Arthur and Alldredge, 1982). For example, comparative studies of plant uptake of plutonium under both field and laboratory conditions generally yield results of the type shown in Table 4.17. Laboratory studies focused on root uptake of plutonium from soils yield concentration ratios which are at least 1 order of magnitude (and often 2

3 orders of magnitude) lower than ratios observed under comparable conditions at field sites. The differences in concentration ratios between laboratory and field studies imply that mechanisms exist in arid environments for delivering at least 10 times more plutonium to vegetation than does direct transport via roots.

Table 4.17 Comparison of plutonium concentration ratios for field and glasshouse conditions (Romney and Wallace, 1976)

Studies at Los Alamos, New Mexico, demonstrated that rainsplash of soil particles with subsequent deposition on foliage surfaces can easily contribute most of the plutonium measured in field-site vegetation (Dreicer et al., 1984). More importantly, those studies, which employed a labelled-soil particle technique and the scanning electron microscope, have shown that relationships that govern translational movement of plutonium by soil erosion processes also govern transport of plutonium to foliage surfaces. For example, the energy of impacting raindrops caused an enrichment of the smaller soil particles on foliage surfaces. In general, only the highly transportable silt-clay particles, which often contain higher concentrations of plutonium at Los Alamos, were retained by plant surfaces. While absorption of plutonium through leaf surfaces has been demonstrated (Cataldo and Vaughan, 1980) it is considered to be of limited importance in the field, particularly for annual or deciduous vegetation.

Studies on the uptake of plutonium by vegetable crops grown in field sites at Los Alamos show that as much as 50 per cent of the plutonium in crop samples was surficial contamination that could be removed by standard food preparation procedures (White et al., 1981). Plutonium that cannot be removed from vegetable crop surfaces in arid environments does not necessarily reflect plutonium incorporated into plant tissues. Cataldo and Vaughan (1980) showed that submicron particles on foliage surfaces are difficult to remove by either simulated wind or rain.

The concentrations of plutonium and caesium in animals collected from field sites in arid environments indicate that gut availability of these elements in the environment is relatively low, as shown by the low concentrations in internal organs and tissues across a variety of sites and source conditions. In addition, highest concentrations of plutonium and caesium are often measured in tissues exposed to contamination with soil particles. For example, plutonium in the pelt, gastrointestinal tract and lungs accounts for nearly all of the animal's body burden (Hakonson and Nyhan, 1980; Moore et al., 1977).

The high mobility of large herbivores coupled with natural elimination processes could provide a mechanism for radionuclide transport across the landscape. Although the amount of plutonium transported across the landscape by this mechanism is considered to be small in areas where the extent of contamination is large (i.e. fallout areas) relative to the home range of the animal, there are circumstances where this transport mechanism becomes important. For example, in a nuclear waste burial site at Hanford Washington, jack rabbits (Lepus californicus) gained access to buried waste, ingested radioactive strontium, and subsequently excreted it on the ground surface of the site and surrounding area (O'Farrell and Gilbert, 1975).

Studies on pocket gophers (Thomomys bottae) inhabiting a low-level waste site at Los Alamos, New Mexico (Hakonson et al., 1982), showed that the burrowing activities of this animal can greatly perturb cover profiles placed over low-level radioactive waste disposal trenches. Over a one year period, gophers excavated about 11 metric tonnes of soil per hectare from within the trench cover and created about 3000 m of tunnel system in the cover profile. In addition to altering the hydrological properties of the soil, animal burrowing activities can also alter radionuclide distributions within the soil profile, as has been shown for pocket gophers in plutonium contaminated sites at Rocky Flats, Colorado (Winsor and Whicker, 1980) and for other small mammals at the Radioactive Waste Management Complex at Idaho National Engineering Laboratory (Arthur and Markham, 1983).

Present data demonstrate that soils and sediments serve as the major repository of plutonium, caesium, and strontium in arid ecosystems and that processes which redistribute soils and sediments can also cause major changes in the environmental distribution of these elements. It is clear that there is a need to determine the relative importance of radionuclides associated with soil and sediment as a source of contamination to biota. Available data on long-lived radionuclides in arid ecosystems point to the potential importance of the physical movement of soil and sediment through food webs.

Any phenomenon which retains radionuclides in contact with the biosphere for extended times has the potential to increase risks to biota due to exposure to these elements. The interception properties of vegetation, mechanical disturbance by organisms living in the soil, and processes which resuspend soil all contribute to maintaining radionuclides in contact with arid site biota. The ultimate fate of long-lived radionuclides in arid ecosystems, including transport to biota, will ultimately depend on the changes that occur in their distribution in soils and sediments and, barring significant changes in the environmental `solubility' of elements such as plutonium, the fate of the soil and sediments themselves.

The term `upland' can be used in a purely geographical sense to define land over a certain altitude or to describe a range of habitats and their associated ecosystems. With regard to radioactive contamination it is the latter definition which seems most often implied in the literature. `Upland' ecosystems are determined by climate, there is a low temperature regime and often a high rainfall. Thus under suitable conditions `uplands' may occur at sea-level; in many respects they are determined by latitude rather than by altitude. The vegetation is poor in agricultural terms and management is often minimal or non-existent.

The use of upland ecosystems varies from that where man harvests the native plants and animals to systems where grazing pressures are manipulated and some fertilizer is applied. The higher the latitude the more man is dependent upon products from the grazing animal until, as described in Section 4.3.4, there are populations totally dependent upon animals such as reindeer and caribou. Research on upland systems has been dominated by the effects of the Chernobyl fallout. The research emphasis varies from country to country depending on whether, as in the UK, agricultural grazings were affected (Beresford et al., 1987; Coughtrey et al., 1989), or, as in Scandinavia, where natural ecosystems contribute largely to the pathways to man (Hove et al., 1990). Whereas work has concentrated on radiocaesium, other radionuclides such as plutonium isotopes from bomb fallout are present. There is still little information on the factors controlling the behaviour of other radionuclides in upland ecosystems.

The dominant source of radionuclides to upland areas is wash-out of aerosol material by rainfall. After the Chernobyl accident some dry deposition of ¹³¹I took place but compared with wash-out this was small. Radiocaesium deposition in the UK has been shown to be closely correlated with rainfall (Clark and Smith, 1988) and direct deposition from clouds (occult precipitation) at altitude is another source which has not been given as much attention. However, the chemical form of the aerosols does not appear to have been well determined although it is thought that they range from soluble forms such as hydroxides and carbonates to refractory particles.

Geology affects soil type and vegetation mainly according to the hardness and chemical composition of the rock. In high rainfall areas on acidic rock types highly organic soils are formed, and as the rainfall increases this may occur even over limestone. These organic soils, in combination with waterlogging, lead to the formation of a range of wetland habitats ranging from nearly open water fens to Sphagnum dominated raised bog.

Table 4.18 Caesium isotope distributions in soil profiles. All data decay-corrected to 1 August 1989


Sample site	Depth (cm)	¹³⁴Cs (Bq/m²)	¹³⁷Cs (Bq /m²)

Italy
Tarvisio Podzol	03	1610	9700
	38	135	1240
	813	<26	310
	1318	<18	145
	1823	<20	41
	23+	<24	26
Pozzis Brown Earth	03	4860	26 100
	38	1900	12 400
	813	160	1210
	1318	89	570
Sella Nevea Podzol	18+	41	200
	07	10 300	56 800
	715	690	4140
	1522	350	2500
	22+	200	1200
Stolvizza Calcareous	05	8020	47 000
Brown Earth	511	560	5880
	11+	110	990
Norway
Geiteseter Ranker	05	7410	40 500
	510	395	2030
	1015	170	830
	1520	100	530
	2025	46	210
	2530	<28	120
Stuttgonglia Brown Earth	05	7350	39 900
	510	570	3040
	1015	590	3300
	1520	88	490
	2025	35	310
	2530	<30	180
Grinningsdalen Fen/Mire	05	420	2290
	510	560	3060
	1015	<25	65
	1520	21	130
Scotland
Sherrabeg Brown Earth	05	1100	7630
	510	144	1540
	10-15	80	680
	1520	35	350
Glen Shirra Podzol	05	1530	9320
	510	97	650
	1015	90	1290
	15-20	<27	280
	2025	<22	86
	2530	<30	<35
	3035	<21	46
	3540	<17	56
Creag Meagaidh	05	400	2330
Ranker	510	140	1520

Samples collected August 1989. Data from Livens et al. (1991).

Though radiocaesium is leached readily from organic soils (Munson, 1985; Coughtrey et al., 1983

85) it is also apparent that the presence of organic matter enhances uptake by pasture plants according to clay mineral content and composition (D'Souza et al., 1972, 1980). The dominance of organic soil types in Western Europe and the mobility of radiocaesium in these soils has led to high uptake by vegetation and a consequent contamination of animal tissues. A knowledge of the physico-chemical properties of soils allows the mobility of radiocaesium to be determined in a relative manner. The `immobilization capacity' of a soil (Livens and Loveland, 1988) is a reflection of clay mineral content and type, organic content, pH, ammonium content and potassium status. Although this is not quantifiable it does permit ranking of soils, and has been used with some success in the field. Soils classified as most vulnerable in Cumbria, UK, correspond closely with areas where sheep movement is prohibited. A closer approach to, quantification has been made by Cremers et al., (1990). The potential of a soil to absorb radiocaesium is defined as the product of the number of absorption sites and the selectivity for caesium with respect to potassium. Knowing these values and the ammonium-potassium status of the soils good agreement with the field situation may be obtained.

The distribution of radionuclides in upland soils will obviously depend upon the soil type. This can vary from a pure waterlogged peat through podsols to a variety of brown earths. It has been found (Cawse and Horrill, 1986; Sandalls et al., 1988) that radiocaesium originating from both weapons testing and Chernobyl fallout are moving relatively slowly down the soil profile. Chernobyl radiocaesium dominates in the upper parts of the profile (Table 4.18) and over the first few years after deposition resided nearly all within the rooting zone of the vegetation. Shortly after the Chernobyl accident radiocaesium from this source was more mobile than weapons testing fallout but a gradual decline in mobility has been recorded with time.

In contrast to the approach whereby the physico-chemical characteristics of the soils are considered the biological activity within a soil cannot be ignored. In many upland areas the soil microflora is important in determining the breakdown processes. In some types of highly organic upland soils Dighton et al. (1991) have suggested that the bulk of the radiocaesium could be held in the fungal mycelia. Clint et al. (1991) have shown that soil fungi can hold an appreciable fraction of the soil radiocaesium.

It would appear therefore that in upland ecosystems there is still considerable doubt as to the reasons why the Chernobyl radiocaesium is recycling around the system so readily, and it is possible that an integration of both physical and biological factors is needed to explain the problem.

The range of habitats found in `upland' areas is wide and includes grasslands, heaths, rock communities, and wet and dry peaty habitats. Woodlands cannot be excluded as the tree line is determined by exposure and the potential limits can extend from sea-level to thousands of metres. As a consequence of this wide spread of habitats the range of species is wide. Most frequent, however, are those species of the wet acidic habitats including many lower plants represented mainly by bryophytes, lichens and fungi.

Radiocaesium uptake varies widely depending on plant species, habitat and season. However, there are some generalities which may be made. It was first recorded by Bunzl and Kracke (1986) that accumulation within the family Ericaceaea was marked. This finding has since been confirmed by other workers. Even within the Ericaceaea consistent differences have been found between species. Horrill et al. (1990) found similar rankings for ericoid species collected from a series of dry heathland sites (Table 4.19).

Bryophytes and lichens obtain the bulk of their nutrients by trapping aerosols or as they rarely have conducting tissues, the capillary movement of groundwater up the plant body. It is frequently found that these lower plants have higher concentrations of radiocaesium than other plants. This is illustrated by Livens et al. (1991) where data for species from 10 sites from Italy, Norway and Scotland show the bryophytes often have radiocaesium concentrations an order of magnitude greater than higher plants from the same habitat. Relatively high concentrations of radionuclides were reported in lichens during the period of maximum fallout from weapons testing. Cigna Rossi (1971), for example, reported data for ⁵⁴Mn, ⁹⁰Sr,⁹⁵Zr, ¹⁰⁶Ru, ¹²⁵Sb, ¹³⁴Cs, ¹³⁷Cs and ¹⁴⁴Cein samples collected in Italy and concluded that a selective absorption mechanism existed only for ⁹⁰Sr.

Radionuclide concentrations in different species from the same locality have been found to vary widely even when species are only separated by a few centimetres (Table 4.20). Many possible explanations for this have been put forward, including physiological and morphological differences, but the true reasons are still obscure. Kirton et al. (1990) demonstrated that the temporal pattern of radiocaesium concentrations varies according to the growth patterns of individual species, and it has been demonstrated that regrowth after grazing can result in enhanced concentrations of radionuclides.

Fungal fruiting bodies have been found to play an important part in determining the contamination levels of animals in upland regions. Selective uptake of radiocaesium can be shown by using the ⁹⁰Sr/¹³⁷Cs ratio which was about 0.01 in the fallout and 2

3 orders of magnitude lower in mushroom species (Mascanzoni, 1990). Radionuclides can accumulate to high levels in certain species, up to 445 kBq/kg for ¹³⁷Cs (Hove et al., 1990) and selective grazing by animals such as goats takes place. Goat milk in late August in Norway has been recorded as averaging 800 Bq ¹³⁷Cs per litre. In Scandinavia the unfortunate coincidence of the fungal fruiting and hunting seasons means that harvesting of wild animals takes place at the worst possible time with regard to contamination of meat.

Table 4.19 Comparative performance of ericoid species^a on dry heathland sites (Bq/kg dry weight)


Site	Species	¹³⁷Cs	¹³⁷Cs in soil

Grasmoor	Calluna vulgaris	469	669
NGR NY 167198	Vaccinium myrtillus	220
360 m	Erica cinerea	93

Hindscarth	Calluna vulgaris	366	730
NGR NY218177	Empetrum nigrum	107
490 m	Vaccinium myrtillus	78
	Erica cinerea	33

Yewbarrow	Calluna vulgaris	227	76
NGR NY 16709	Erica cinerea	40
85m

Robinson	Calluna vulgaris	126	500
NGR NY 193176	Vaccinium myrtillus	96
335 m	Erica cinerea	65

Gawthwaite	Calluna vulgaris	39	98
NGR SD263848	Vaccinium myrtillus	170	713
200 m	Erica cinerea	28	98

^aAll above-ground parts of the plant sampled. Collected in August/September 1988.
Data from Horrill et al. (1990).


		¹³⁷Cs	¹³⁴Cs

Higher plants
	Calluna vulgaris	6590	2075
	Vaccinium myrtillus	3964	1194
	Agrostis canina	1247	379
	Juncus squarrosus	1132	303
	Juncus effusus	936	295
	Eriophorum angustifolium	595	121
	Nardus stricta	514	153

Bryophytes
	Sphagnum spp.	7472	2813
	Polytrichum commune	4026	1445
	Polytrichum alpestre	3543	1187

^aAll above-ground plants sampled. Collected in spring 1987.

Animal populations of the uplands of north-west Europe are those for which most data are available and the radionuclides best documented are ⁹⁰Sr and ¹³⁷Cs. Both have originated from atomic weapons testing, nuclear reprocessing and accidents such as those at Kyshtym and Chernobyl. The range of animals affected is wide and includes wild species and domestic stock. The type of land used by domestic animals varies widely from natural vegetation to enclosed managed pastures.

Radionuclide uptake by wild species is highly dependent on many factors even within the same area. Lowe and Horrill (1991) found significant differences in concentration factors between food and muscle tissue depending on species, food, sex, breeding condition and age (Table 4.21). Here animals from the same general area show great differences in tissue concentration. Rabbits (Oryctolagus cuniculus) feeding on grasses on mineral soils, where radiocaesium mobility is low, always had the lowest tissue concentrations whereas the red grouse (Logopus logopus) feeding exclusively on heather had the highest.

Table 4.21 Analysis of samples of flesh and vegetation from Loyal Estate in Sutherland (Bq/kg fresh weight)


		Flesh		Vegetation
		¹³⁴C_S	¹³⁷Cs	¹³⁴C_S	¹³⁷Cs

Red grouse	(cock)	325.0	962.5	147.1	408.3
	(hen)	602.1	1684.3	267.9	692.4	^a
Red deer	(hind)	111.9	310.7	78.3	202.7
	(calf)	186.3	535.4	70.6	189.7
Rabbit	(buck)	8.1	17.8		8.2
	(buck)	5.1	13.1		11.9
Fox	(vixen)	175.6	460.6

^aHeather shoots collected by hand, all other vegetation samples obtained from the gut of
the animal, Samples collected in autumn 1988. Data from Lowe and Harrill (1991).

The high concentrations found in many of the plant species of the uplands, particularly the lower plants, combined with seasonal scarcity of fodder, often means that dietary intake at certain times of the year can be high in radionuclides. The high accumulation by food species such as lichen means that the intake by reindeer can be 10 times the activity of radiostrontium and radiocaesium of grasses growing in the same area (Nevstrueva et al., 1966).

The Chernobyl accident put a single pulse of radionuclides into the ecosystem. It became possible, by use of the ¹³⁴Cs isotope, to compare not only the movement of radiocaesium from two sources, weapons testing fallout and Chernobyl, but also to study the ecological half-life in some systems. Pre-Chernobyl ¹³⁷Cs concentrations reached over 600 Bq/kg fresh weight for red grouse and roe deer in the UK. The ecological half-life of ¹³⁴Cs in native roe deer has been calculated as 28 days (Lowe and Horrill, 1988).

Due to their economic importance domestic stock have been extensively studied with regard to radiocaesium, as has the passage to man in meat and milk products. Control measures after Chernobyl have received much attention particularly from an economic point of view.

Radiocaesium uptake by animals in the uplands is rapid. Within a few days of the Chernobyl material being deposited, radiocaesium was found in milk and shortly afterwards in meat. Elimination of radiocaesium is equally rapid, the biological half-life being 11

15 days for sheep when removed from contaminated areas to lowland pasture (Howard et al., 1987).

A wide range of radiocaesium concentrations have been recorded for domestic stock with values ranging from hundredsto thousands of Bq/kg fresh weight depending on location. Uptake of radionuclides is complex depending, as for wild animals, on species, food, age and breeding condition. For instance the transfer to goat's milk appears to be higher than other animals. The state of the radionuclide, whether incorporated in vegetation or bound to attached soil particles, is of great importance as the latter is much less available for uptake.

Ameliorative measures to reduce radionuclide contamination in upland systems have to be cheap and easy to use. They fall into two categories

management and diet additives. Management techniques rely on removing the animals from contaminated pastures before sale or use and allowing the short biological half-life of the caesium isotopes to operate. It is too expensive to consider treatment of large areas of land but diet additives can be economic. The use in Norway of salt licks or administering boli containing the binding agent ammonium-ferric-cyano-ferrate (AFCC) has proved effective in the case of reindeer and sheep (Hove and Ekern, 1988).

Considerable interest in artificial radionuclides in arctic and sub-arctic ecosystems, particularly those in circumpolar regions, began with simultaneous and independent reports of appreciable concentrations of worldwide fallout in successive links of the atmosphere

lichen

caribou/reindeer

man food chains of northern Alaska and Scandinavia during the 1960s (Liden, 1961; Palmer et al., 1963). Several circumpolar research programmes were initiated to promptly evaluate the radiation exposures of Inuit (Eskimos), Indians, Sami (Lapps), and Soviet reindeer herders because of their dependence upon caribou and reindeer for food and livelihood.

This section is based on radioecological research on worldwide fallout from nuclear weapons tests in arctic ecosystems during 1959

1980, with ancillary consideration of radioactive debris from satellite re-entry episodes. Emphasis is placed upon ¹³⁷Cs because of its greater accuracy of measurement, consistent ratios to other fallout radionuclides, and radiological implications in arctic regions (Hanson, 1966).

Radiation ecology studies were conducted over the general area of northern Alaska between 66° North latitude and the Arctic Ocean, constituting about 310 000 km². Within this area are five major physiogeographic provinces, the north-western terminus of the North American continental divide, four major caribou herds which constitute the largest aggregation of big game animals in North America, and several Inuiat Eskimo and Athabascan (Kutchin) Indian villages representing four major ethnic groups. A spectrum of ecosystems of varying complexity are utilized by the migratory caribou herds and by the native groups inhabiting the region, and were the subjects of the studies. The approach proceeded from the general description of radionuclides in the northern biosphere to specific investigation of the most promising components of critical food webs, for purposes of defining routes, rates and concentrations of worldwide fallout in arctic food chains. The Arctic Coastal Plain, a northern extension of the Great Plains of interior North America, consists of a 100 km wide smooth plain that rises imperceptibly from the Arctic Ocean to a maximum elevation of 180 m at its southern margin, where it abuts the Arctic Foothills. The Plain is poorly drained, marshy in summer, and covered by elongated thaw lakes. There are no glaciers, although the entire land area is underlain by more than 300 m of permafrost. A network of ice-wedge polygons covers the Plain; an active layer (depth of summer thaw) varies from 10

100 cm and allows limited percolation of moisture. Annual precipitation is 10

15 cm, including 50

90 cm of snow.

The Arctic Foothills consist of rolling plateaus and low linear mountains extending 150 km southward from the Coastal Plain. The northern section rises from 180 m where it joins the Plain to 350 m on the south, with broad east-trending ridges dominated by mesa-like mountains; the southern section is characterized by irregular buttes, knobs, and other landforms of about 350

450 m elevation, with local relief up to 750 m. In common with the Coastal Plain, there are no glaciers and the entire province is underlain by permafrost. The substrate contains ice wedges, stone stripes, polygonal ground, and other features of a frost climate. Annual precipitation ranges from 15 to 25 cm, with about 1

1.2 m snow.

The Arctic Mountains Province consists of a 125 km wide by 1000 km long area of features carved chiefly from folded and over-thrust Paleozoic and Mesozoic sedimentary rocks during Pleistocene glaciation. An irregular and indistinct northern boundary is shared with the Foothills Province, in contrast with the abrupt southern front. The central and eastern Brooks Range is an area of rugged glaciated east-trending ridges of 2100

2400 m elevation in the northern portion and 1200-1800 m in the southern portion. Small cirque glaciers are common in the higher parts of the Range, such as the 2700 m Schwatka and Romanzof Mountains in the north-east section.

In general, the Arctic Coastal Plain and Foothills Provinces are mostly covered by Intrazonal Tundra and Bog soils, usually strongly acid and saturated by water. These soils are poorly drained silt loam or silty clay loam and of least importance in the mountains, where they are confined to the valleys. Arctic Brown is the zonal soil in which the dominating influences of climate and vegetation are expressed; however, it is present on a small portion (1 per cent) of the area, being most common in dry, well-drained environments of the Foothills Province and on kame terraces on valley floors within the Arctic Mountains. With increasing ground moisture, Arctic Brown soil is successively replaced by Upland Tundra, Meadow Tundra, Half Bog and Bog soils. Upland Tundra soils appear most often on steep slopes which are better drained than the gentle slopes of the valley floors, where Meadow Tundra, Half Bog and Bog soils occur.

Vegetation of the Coastal Plain Wet Tundra is dominated by graminoid species, particularly Carex aquatilis; cottongrass tussocks (Eriophorum vaginatum) characterize the Moist Tundra environment of the Foothills; and low, mat-forming heath communities occupy extensive fellfields of Alpine Tundra in mountainous areas. Lichens and mosses are common components of all three tundra vegetation types, with lichen carpets most extensive on stream terraces in mountain valleys. Stream drainages are dominated by high brush, such as willows (Salix spp.), alder (Alnus crispa), with isolated stands of cottonwood (Populus balsamifera). Alpine Tundra and Moist Tundra share much of the southern exposures of the Brooks Range mountains, intergrading with closed spruce

hardwood forests composed of white and black spruce (Picea glauca and P. mariana), paper birch (Betula papyrifera), aspen (Populus tremuloides), and balsam poplar (Populus balsamifera) with a rich understorey of shrubs and extensive lichen carpets that are the major winter forage of caribou.

Inventories of ¹³⁷Cs in surface (top 5 cm) soils in northern Alaska during 1975

1979 declined at an effective half-life of 9 y, including minor (6 per cent) increases due to fallout deposition during snow-free summer months. Much of the radioisotope loss from surface soil is a combination of surface erosion during snowmelt and percolation to depths >5 cm. Sampling below lichen carpets showed that 5 per cent of the ¹³⁷Cs and 15 per cent of the ⁹⁰Sr inventories was in A₀ horizons (humus layer) and 3

4 per cent for both radioisotopes in A₂ horizons (organic mineral soil).

Plutonium in surface soils at northern Alaska and northern Greenland study sites decreased with a 0.4

0.5 y half-life and was measured with considerable difficulty, due to the large samples required to provide statistically valid measurements.

Lichens are particularly efficient accumulators of fallout radionuclides, of which ⁹⁰Sr and ¹³⁷Csare the most important biologically. Concentrations of ¹³⁷Cs in lichens tend to be higher than in nearby unvegetated soils. This is due to a number of factors such as the ability of lichens to absorb mineral metabolites from rainwater and snowmelt, radionuclide loss from soil by surface erosion, and upward cycling of ¹³⁷Cs from underlying soil horizons. Thus, lichens provide an enriched, long-term, source of important radionuclides which tend to be concentrated in upper portions of the preferred winter forage of caribou and reindeer.

The possibility of geographical and ecological differences in ¹³⁷Cs and ⁹⁰Sr concentrations in lichen communities over the very large study area of northern Alaska was investigated during 1967 and 1972 by intensive sampling at 20 locations from north-western Canada to the Chukchi Sea, including three locations across the wintering range of the caribou that provided critical food to the native villages. Three-way analysis of variance of results for ¹³⁷CS showed no significant difference between three ubiquitous species, three physiogeographic provinces from which they were collected, or sampling years (Hanson, 1973). Student's t test of combined means of all species tested between years indicated that there was usually a highly significant (P < 0.01) difference between ¹³⁷Cs concentrations of the nine species from the three provinces, with samples from the Arctic Coastal Plain and Arctic Foothills more similar than those of either province compared to the higher values in the Arctic Mountains (Brooks Range).

The upper 12 cm (active growing portion) of Cladonia stellaris lichen communities generally contain 70

90 per cent of the ⁹⁰Sr and > 90 per cent of the ¹³⁷Cs inventory in the lichen carpet, with some indication of fractionation between upper and lower 6 cm increments. ¹³⁷Cs is more readily translocated than ⁹⁰Sr, resulting in higher concentrations of ¹³⁷Cs in upper layers of lichen mats. Experiments with ⁹⁰Sr and ¹³⁷Cs in lichens indicate that translocation along lichen thalli is primarily diffusive in character but complicated by cation exchange, especially for strontium. Fixation of strontium to the thalli explains the greater mobility of caesium between compartments of the lichen carpet and underlying strata. Several circumpolar studies after weapons-testing fallout indicate effective half-lives of about 5 ± 2 y for ¹³⁷Cs and 1.0

1.6 y for ⁹⁰Sr in arctic and sub-arctic lichen communities.

Comparison of simulated and observed data in a deterministic model of ¹³⁷Cs concentrations in lichens collected at Anaktuvuk Pass during 1963

1973 (Thomas et al., 1982) demonstrated the deficiencies in W.C. Hanson's model despite intensive sampling and analysis. A major need was accurate measurements of fallout deposition rates at the lichen sampling sites, which would have greatly improved model performance in the reality of variable inputs of fallout due to discontinuous nuclear weapons test series and weather systems.

Plutonium isotopes in lichens from northern Alaska during 1967

1979 occurred in pronounced peaks during 1968, 1972, 1974, and 1976 that correlate well with periods of high-yield nuclear tests. Both ²³⁸Pu and ^239,240Pu isotopes showed the same pattern of concentrations, with ²³⁸Puconsistently 0.1 times the ^239,240Pu values rather than the 0.022 ratio reported in fallout. This factor of 5 disparity was maintained throughout the series, indicating appreciable retention in the lichens. Results of studies in northern Alaska and central Sweden suggest that ²³⁸Puis more tightly bound in the upper 6 cm stratum of lichens than is ^239,240Pu. Both plutonium isotopes demonstrate an effective half-life in the lichen carpet of about 6 y.

Plutonium-239 was reported to have a mean residence time of 4.3 ± 0.5 y in the top 3 cm of the Cladonia alpestris carpet in Sweden, compared to 6.1 ± 0.5 y in the entire 12 cm carpet (Holm and Persson, 1975), with the difference ascribed to ingrowth of lichen biomass and low solubility of the fallout plutonium oxide.

Concentrations Of ¹³⁷Cs in caribou and reindeer flesh follow an annual cycle with low values during autumn months and maximum values in spring months (Figure 4.2). The low values occur in animals returning from their summer range on which sedges and other fresh forage provide a diet containing relatively low amounts of fallout ¹³⁷Cs and substantial amounts of potassium, which combines with increased summer body water turnover to produce the abrupt decline in net accumulation of ¹³⁷Cs in soft tissues. An effective half-life of 28 days was usually observed for ¹³⁷Cs in caribou muscle between late May and late August, when migration to winter ranges usually began. it is generally accepted that the rate of caesium loss is dependent upon the potassium intake rate; experiments with reindeer (Holleman et al., 1971) showed that, for the slow ¹³⁷Cs component, the biological half-time was about 17 days with a dietary potassium concentration of 1 mg/g dry weight and about 6.7 days with a dietary potassium concentration of 5 mg/g dry weight. Extrapolation of a log

log plot gives a corresponding half-time of about 30 days for a potassium concentration of 0.37 mg/g, which was the mean value measured in 17 lichen samples collected from caribou winter range near Anaktuvuk Pass during 1964

1967.

During autumn, caribou and reindeer gradually shift to their winter diet composed mainly of lichens, and ¹³⁷Cs concentrations in soft tissues begin a steady increase through winter months; in Alaska the ¹³⁷Cslevels usually plateaued during January

April of each year. An appropriate ¹³⁷Cs tracer kinetics model applied during periods of simultaneous sampling of caribou and lichens at the same location on winter range indicated that 4.5

5.0 kg dry weight lichens were ingested per day. ¹³⁷Cs concentrations in caribou flesh at the end of the winter were 4.0±0.9 (SD, range 3.2

5.5) times the ¹³⁷CS concentration of lichens.

Figure 4.2 Fallout ¹³⁷Cs in biota of northern Alaska ecosystems during 1962

1979

Comparison of ¹³⁷Cs levels in other important herbivores such as moose (Alces alces) and Dall sheep (Ovis dalli) on nearby ranges to those utilized by caribou showed no significant seasonal pattern and were significantly lower than caribou values. Sheep muscle contained four times the ¹³⁷Cs concentration in moose muscle but only one-fourth to one-tenth of ¹³⁷Cs concentrations in caribou muscle. These differences were due to food habits; the sheep fed mostly on sedges, forbs, and modest amounts of lichens, while moose were observed to feed mostly on willows and other shrubs, with some minor amounts of grasses.

After prolonged exposure, ⁹⁰Sr concentrations in bone are about 1000 times those in muscle of caribou and reindeer. In northern Alaska, ⁹⁰Sr levels in caribou bone were relatively stable near 740 Bq/kg until 1962 and then began a sharp increase to 2600 Bq/kg in 1966

1969, after the major nuclear weapons test series of 1961

1962. The ratio of concentrations in caribou bone/lichen was about 7.6 during the 1966

1969 plateau period, when a condition of 99 per cent of equilibrium existed. During the period May 1964-November 1974 the values declined at a half-life of 56 months to levels near those measured in 1962. Strontium-90 in flesh followed a more gradual curve of increase, peaking in years of nuclear weapons testing, similar to the ¹³⁷Cs pattern (Hanson and Thomas, 1982).

Concentrations of plutonium in caribou bone samples during the 1971

1975 period of maximum values in lichens were barely detectable. Concentration ratios relative to lichens were usually in the range of 0.02 for ²³⁸Puand 0.001 for ^239,240Pu compared to values in lichens.

Concentrations of worldwide fallout ¹³⁷Csin circumpolar food chains follow an annual cycle dependent upon the subsistence patterns of human populations and the food habits of carnivores, particularly those associated with caribou and reindeer.

Extensive sampling of wolves, foxes, and wolverines in northern Alaska showed a repeated annual ¹³⁷Cs pattern with a rapid increase through autumn and winter that paralleled the increase in ¹³⁷Cs concentrations in caribou. Wolves contained twice the concentrations observed in foxes and wolverines, both of which, in Alaska, are primarily scavengers on wolf-killed caribou and consumers of small mammals that contain much lower ¹³⁷Csconcentrations. The ratio of concentration for ¹³⁷Cs in wolf flesh/caribou flesh averaged 2.7 at a time at which 80 per cent equilibrium had been achieved.

A similar pattern occurred in ⁹⁰Sr concentrations in bone of the wild carnivores and the same relationship was observed; maxima occurred in late winter following utilization of caribou that were increasing their concentrations and wolves contained twice the ⁹⁰Sr concentrations of foxes and wolverines. Wolves consume more caribou bone than do foxes and wolverines and therefore ingest a richer source of ⁹⁰Sr than do scavengers. The concentration ratio for ⁹⁰Sr in wolf bone/caribou flesh at about 50 per cent of equilibrium at time of sampling was 0.4. An effective halftime of about 7 years for ⁹⁰Sr in bone was observed in all three species of carnivores.

Using the lichen forage ingestion model it was estimated that caribou on winter range ingested about 3 Bq ²³⁸Puand about 30 Bq ^239,240Pu per day during the mid-1960s to mid-1970s, yet bone samples contained barely detectable amounts. Assuming averages of 0.7 Bq/kg for ²³⁸Puand 7 Bq/kg for ^239,240Pu in lichens during the above period, concentration ratios of 0.02 and 0.001 are estimated. This indicates that ²³⁸Puwas more readily transferred through the lichen

caribou food chain than was ^239,240Pu; however, the tenuous nature of the values makes such a conclusion speculative.

The island environments of the atolls in the Marshall Islands represent a unique ecosystem where radionuclides that were introduced between 1946 and 1958 have had nearly 40 y to equilibrate. Bikini and Enewetak Atolls were the sites of 66 atmospheric nuclear weapons tests. These atolls are composed of coral limestone, which has accumulated on old igneous seamounts that arise in the Pacific Basin. The depth of coralline deposits beneath these atolls is about 1.2 km at which point there is contact with primordial igneous rocks. An atoll usually comprises a discontinuous series of islands on a ring of coral reef that surrounds a relatively shallow lagoon. An actively growing reef of coral organisms is present on the seaward side of the atoll and coral beaches occur on the lagoon side.

The species diversity of the atolls is quite restricted because of the remoteness from continental masses that typically supply organisms for island colonization. Most of the vegetation on the atolls has arrived there by either water or animal transport. Most of the plant species that comprise the stable, climax vegetation of the atolls in the Marshall Islands are pantropical in distribution. The coconut palm, Cocos nucifera, and the pisona tree, Pisona grandis, are good examples of the pantropical floristic element, and these species occur throughout the Pacific Basin on both islands and atolls.

Areas that were disturbed in the 1950s have not returned to the stable tropical forest type and are present as open woodlands of Messerchimidia argentea and Scaevola frutescens, with vines, sedges and grasses occurring between the clumps of low trees and shrubs.

The amount and frequency of rainfall at the atolls are of paramount importance in the cycling of a mobile ion such as ¹³⁷Cs. Most of the annual rainfall occurs from May through November; the months from December through April tend to be very dry. About 50 per cent of the annual rainfall occurs from August through November.

Rainfall is variable and ranges between 114 and 140 cm/y in the northern Marshall Islands to as much as 380 cm/y in the southern Marshall Islands. The lack of rainfall during some years seriously affects the growth of vegetation, and in drought years some species die back. Most of the research discussed here has taken place in the northern Marshall Islands where native agriculture often does not succeed with rainfall as the only source of water.

The unique properties of the soil at the atolls dictate to a large extent the distribution and cycling of radionuclides in the ecosystem. The composition of coral soil at Bikini and Eneu Islands is listed in Table 4.22. The soil is composed primarily of calcium carbonate (CaCO₃) with some magnesium carbonate (MgCO₃) and no silica-clay component. The pH is high, ranging from 7.7 to 9.0. The surface horizons are high in organic content (as much as 14 per cent), although the organic content of the soil drops markedly with depth in the soil column. The soil is low in exchangeable potassium and marginal in phosphorus and trace-mineral content. Some of the native plant species and most introduced species show definite signs of potassium deficiency. In fact, with most introduced food crops and ornamental plants, growth is very limited without addition of potassium as well as nitrogen, phosphorus and trace minerals.


Island	Total^a							Extractable
location							Organic
and depth		Sr	Ca	Mg	P^c	N	Matter^d	K^e
(cm)	pH^b	(%)	(%)	(%)	(%)	(%)	(%)	(ppm)

Bikini No. 1
0-5	7.7	0.38	30.4	0.95	1.35	0.64	14.4	79
5-10	7.8	0.39	30.8	0.89	1.28	0.62	13.2	26
10-15	7.9	0.39	30.9	0.89	1.29	0.63	12.3	20
15-25	7.9	0.40	31.9	0.86	1.17	0.50	10.6	23
25-40	8.3	0.39	34.3	1.28	0.67	0.19	4.5	4
40-60	8.4	0.31	34.5	2.05	0.16	0.11	1.6	3

Bikini No. 2
0-5	7.8	0.40	31.0	1.02	0.82	0.49	10.7	50
5-10	8.0	0.40	32.4	1.09	0.71	0.46	8.5	24
10-15	7.9	0.38	33.1	1.18	0.56	0.35	7.4	24
15-40	8.2	0.38	34.7	1.79	0.32	0.11	1.6	6

Eneu No. 1
0-5	7.7	0.32	32.0	1.74	0.085	0.30	5.1	41
5-10	8.0	0.34	32.6	1.76	0.055	0.35	5.6	20
10-15	8.0	0.31	34.3	2.08	0.037	0.17	2.6	9
15-25	8.4	0.28	34.0	2.40	0.016	0.06	0.9	1
25-40	8.7	0.28	34.4	2.48	0.014	0.05	0.8	1
40-60	8.9	0.30	33.3	2.37	0.015	0.03	0.6	<1

^aStable caesium was below detection limit (1.3 ppm).
^bpH in water.
^cHigh phosphorus values indicate ancient guano deposition.
^dOrganic matter by wet oxidation.
^eExtractable in N ammonium acetate.

Figure 4.3 The median concentration of ¹³⁷Cs, ⁹⁰Sr, ^239+240Pu and ²⁴¹Am as a function of depth in the soil column on Bikini Island.

The distribution of ⁹⁰Sr and ¹³⁷Cs with depth in the soil column is essentially exponential, with the highest concentrations in the top few centimetres of the soil column (Figure 4.3). The concentrations of ¹³⁷Cs, and ⁹⁰Sr are generally very similar at the atolls. However, ¹³⁷Cs, is the most significant radionuclide contributing to the dose to people inhabiting the atolls. The concentrations of ^239+240Pu and ²⁴¹Am are much less than those of ¹³⁷Cs, and ⁹⁰Sr. Other radionuclides present a few years ago, such as ⁶⁰Co, have essentially disappeared due to radioactive decay. Most of the activity is within the top 25

40 cm of the soil column, which is where most of the organic material is located. Below the organic layer, the radioactivity concentration drops rapidly and is generally very low. The small amount of radionuclide activity present all the way to the groundwater represents the fraction of the soil inventory of ¹³⁷Cs, ⁹⁰Sr,^239+240Pu and ²⁴¹Am that is solubilized and carried to the groundwater when rainfall is adequate to cause through-flow of rain water and recharge of the underground lens system.

The fundamental composition of the atoll soil produces dramatic differences in the uptake of ¹³⁷Cs and ⁹⁰Sr at the Marshall Islands compared to the uptake rates in the published literature, which are based primarily on silica-clay type soils. For example, the concentration ratio (CR), defined as the activity concentration of the radionuclide per gram of wet plant weight divided by the activity concentration per gram of dry soil, is about 0.1 for ¹³⁷Cs, and about 1.0 for ⁹⁰Sr in silicate soils (Ng et al., 1982). However, in the coral soils in the Marshall Islands the CR for ¹³⁷Cs (based on 0

40 cm) is about 5 while that for ⁹⁰Sr is about 0.0001 (Robison et al., 1988; Koranda et al., 1978). It is generally high for ¹³⁷Cs, and low for ⁹⁰Sr in all the plant species that have been analysed (Table 4.23). Strontium-90 replaces Ca in the CaCO₃ matrix of the coral soil, and is relatively unavailable for uptake; ¹³⁷Cs, is bound to the organic fraction of the coral soil and is relatively more available than in silica-clay soils where it is bound by mineral clays.

In many biological systems K is preferred over Cs and plants selectively take up K and discriminate against Cs (Handley and Overstreet, 1961; Middleton et al., 1960; Nishita et al., 1962; Wallace et al., 1983).

Early studies and measurements were concerned with the distribution of radionuclides in the atoll ecosystem. Because of the importance of¹³⁷Cs in the atoll environment, a landscape inventory was made at Enewetak Atoll on Engebi Island to determine the distribution of ¹³⁷Cs in the soil and above-ground biomass. The data indicated that approximately 3

5 per cent of the ¹³⁷Cs inventory in the soil was present in the vegetation at any time. This suggested that the soil-complexed pool of ¹³⁷Cs was tightly held and not available to the plant.

A laboratory study concerned with the availability of soil-bound ¹³⁷Cs was made to verify the low apparent availability in the ecological inventory. Five columns each with approximately 3 kg of soil were set up and leached with distilled water.

Table 4.23 Concentration ratio (CR) for ¹³⁷Cs, ⁹⁰Sr, ^239,240Pu and ²⁴¹Am at Bikini Island for several species of vegetation. Mean (median) CR (Bq per g in fruit, wet weight/Bq/g in 0

40 cm soil, dry weight)

The pore water was allowed to remain in the soil for varying lengths of time and then leached with 10 times the pore volume. The availability of the soil-bound ¹³⁷Cs was observed to be between 3 and 4 per cent in this laboratory experiment, which was in agreement with the ecological measurements made previously.

A field experiment involved a hectare of atoll surface in which the ground was irrigated with overhead sprinklers using seawater, and the presence of ¹³⁷Cs monitored in the groundwater lens beneath the soil. After delivering 20 m of seawater to the plot, it was apparent that a pulse of ¹³⁷Cs appeared in the groundwater compartment. The amount of ¹³⁷Cs appearing in the groundwater lens was on the order of 3

5 per cent of the soil inventory.

These experiments indicated that ¹³⁷Cs was tightly held in the organic stratum of coral soil on the atolls and, in general, did not respond to the leaching effect of distilled water or seawater. The small fraction of ¹³⁷Cs circulating in the biomass pool was found to be transferred to man via terrestrial foods.

The atoll ecosystem is ideal for evaluating the root uptake of the transuranic radionuclides ^239+240Pu and ^{24 1}Am. Determining what fractions of the total plutonium and americium observed in plants is due to root uptake versus deposition on the plants via soil resuspension has been a difficult problem in most environments with the types of food crops generally used. However, in the atoll system most of the food crops (coconut, breadfruit, pandanus, etc.) grow to such heights that the fruit and canopy are well above resuspended soil aerosol and deposition is unlikely to be significant. In addition, the edible fruits have dense, thick layers that protect the edible portion of the fruit; these protective layers must be husked, peeled or removed in some manner in order to eat the fruit. Consequently, the edible portion is in a sense `sealed' from any deposition type of contamination. The plutonium and americium observed in such foods can only be derived via root uptake.

The CRs for plutonium and americium (Table 4.23) generally agree with pot culture studies in glasshouses where experimental design eliminated any resuspension. The general magnitude of uptake of plutonium and americium seems to be about the same over a wide range of soil types, with the coral soils being at one extreme with high pH and nearly pure CaCO₃ plus organic material. This is in stark contrast to the very different CRs observed for ¹³⁷Cs and ⁹⁰Sr in different soil systems.

The CaCO₃ nature of the soil and the low exchangeable potassium characteristic of coral islands provide the conditions for the observed high uptake of ¹³⁷Cs On the other hand, these conditions would seem to provide a situation in which potassium added to the system might alter the uptake of ¹³⁷Cs. Consequently, experimental plantings of coconut palms, pandanus, breadfruit, and other food plants have been established at Enewetak and Bikini Atolls to study radionuclide uptake under semi-controlled conditions and to provide guidance on the resettlement of the atolls by the native population.

Several experiments using large areas of the coconut grove on Bikini Island have been implemented. The experiments include surface application of potassium ranging from 666 kg/ha to 5550 kg/ha, in the form of N,P,K fertilizer or KCl, and the number of trees per experiment ranging from 10 to 120. The total potassium has been applied over a 3-year period for some experiments and in one application for other experiments. Similar experiments have also been conducted on breadfruit and pandanus trees.

The results of all the experiments show a decrease in the concentration of ¹³⁷Cs by a factor of about 20 compared to the pre-treatment concentration in the fruits. The initial results from a recent experiment using 666 kg/ha indicated a reduced uptake of ¹³⁷Cs similar to that observed for the higher rates of added potassium. Examples of the results are shown for two of the experiments in Figures 4.4 and 4.5. In addition to blocking the uptake of ¹³⁷Cs the treatment with KCl (or N,P,K) increases the growth rate and productivity of the plants as has been observed in other tropical locations (Vernon et al., 1976).

The alternative remedial measure for rehabilitation of the islands is to excavate and dispose of surface soil to eliminate the radionuclides, particularly ¹³⁷Cs. Although this method is very effective in eliminating the radionuclides from the island, it has significant environmental consequences. All of the standing vegetation, which includes mature coconut groves that have taken 20 y to develop, breadfruit, and pandanus trees, must be removed. Then the soil can be removed. This contains most of the radionuclides, but also contains most of the organic material that has taken centuries to develop on the coral islands; below the organic-material-containing layer is essentially beach sand. The soil-organic layer provides nearly all the nutrients needed for plant growth and contributes a significant fraction of the water retention capacity of the coral soil. The consequences for excavation of soil will have long-term ecological and agricultural effects; a very long-term commitment, probably the order of decades, would be required to rebuild the soil and to revegetate the islands.

Figure 4.4 The effect of 1110 and 2220 kg K/ha, 500 kg NP/ha, 1110 kg K/ha + 500 kg NP/ha and 2220 kg K/ha on the uptake of ¹³⁷Cs in drinking coconuts at Bikini Island. (Reproduced from the journal Health Physics with permission from the Health Physics Society.)

Figure 4.5 The effect a single 5550 g/ha K application on the uptake ¹³⁷Cs in drinking coconuts at Bikini Island

Coastal ecosystems occur at the interface of land and sea. For the purposes of this section only systems which contain vascular plants and that are above the low tide mark are considered, thereby excluding the algal communities of rocky shores. Similar systems may also exist on the shores of large freshwater lakes. The wide variety of habitat types associated with coasts are described by Ratcliffe (1977) and include the following: coastal lagoons and swamps, salt marshes, beach systems, sand dunes, coastal grasslands and heath, and cliff vegetation.

Coastal ecosystems can be exposed to radioactive deposition resulting from man's activities (see Chapter 5). Resuspension from the water surface, either in the form of droplets or as suspended aerosols, is an important pathway for transport to coastal environments as is direct submersion by the tides. Mechanical movement, either by wind or living organisms, is also a potential contamination pathway. Coastal ecosystems far removed from the source may be affected by liquid discharges.

Coastal lagoons and swamps often provide environments where, due to sheltered conditions, and trapping by vegetation, deposition of fine sediments is possible. In some instances such as the Great South Bay on Long Island, USA, small amounts of low-level radioactive waste can find their way into these habitats. Extensive use of these areas often occurs for fishing and recreation, and pathways to man include both the food chain and external exposure.

Mangrove swamps can constitute an important ecosystem in tropical areas but very few data exist on radionuclide behaviour in these systems.

Salt marshes are typically found in river estuaries. However, even on exposed coasts salt marshes can occur, such as on the Norfolk coast of Great Britain. Plant species characteristic of salt marshes are specially adapted to tolerate inundation by salt water as well as freshwater at certain times of the year. Whilst some salt marshes still contain natural vegetation, many have been modified by grazing pressures. Salt marshes also serve as important feeding areas for large populations of wildfowl which may pick up radionuclides during winter feeding (Table 4.24).

Table 4.24 Mean, confidence limits and range of radionuclide concentrations in pectoral muscles and livers of a Greylag goose and wigeon from Ravenglass and Morecambe Bay (Bq/kg fresh weight) (From Lowe and Horrill, 1986)

In some cases, nuclear installations are situated on the coast in close proximity to areas of salt marsh. The accumulation of radionuclides in salt marshes is largely associated with sedimentation processes, naturally occurring in these habitats. Radionuclides are mainly sorbed onto sediments attached to the surface of the plants, rather than being incorporated in the tissues. In salt marshes of the Ravenglass Estuary in Cumbria, UK (near the Sellafield nuclear reprocessing facility) a wide range of gamma-emitting fission products can be detected in the silts, e.g. ⁶⁰Co, ⁹⁵Zr, ⁹⁵Nb, ¹⁰³Ru and ¹⁰⁶Ru, ¹⁴¹Ce and ¹⁴⁴Ce, ¹⁵⁴Eu and ¹⁵⁵Eu, ¹³⁴Cs and ¹³⁷Cs. Also present are the actinide isotopes ²⁴¹Am, ²³⁸Pu, ^239,240Pu and ^24lPu (Table 4.25).

All these radionuclides originate from the low-level liquid waste released from the Sellafield reprocessing plant. The deposition is related to both the time of tidal inundation and the structure of the vegetation (Horrill, 1983). Animals grazing these contaminated pastures ingest radionuclides attached to silt and quantifiable amounts have been found in their tissues (Howard and Lindley, 1985) (Figure 4.6). Studies of the transfer of radionuclides in sheep grazing the contaminated pastures show that lambs generally have higher concentrations of radiocaesium in their tissues compared with those of the dam. However, radionuclides with a long half-life, e.g. plutonium isotopes, are present in higher concentrations in the dam's liver than in the lamb. The availability of radiocaesium to animals grazing salt marsh is much reduced compared with that taken in the ionic form (Howard, 1989).

Figure 4.6 ^239,240Pu activity concentrations in tissues from sheep which had been grazing tide-washed pastures in estuaries bordering the Irish Sea. Note that samples have not been taken in every year since 1980 (from: Howard and Livens, 1991).

Table 4.25 Radionuclide activity concentrations found in sediments/soils of tide-washed pastures bordering the Irish Sea (from Howard and Livens, 1991)

Beach systems generally either consist of bare sand or shingle. As such they are almost totally devoid of vegetation and provide little scope for the biological accumulation of radionuclides. The larger systems become vegetated as a result of the accumulation of organic material along the strand line which serves as a foothold for vegetation. If sufficient shingle accumulation occurs then the effect of the sea is diminished and a vegetation type containing heathers can develop. Monitoring at the classic shingle beach at Dungeness, UK, has demonstrated that only small amounts of radionuclides, relative to discharge, are accumulated in this type of habitat.

Sand dunes are sub-maritime coastal habitats. They are not inundated by the sea, nor are they strongly saline. The systems result from the prevailing onshore winds and trapping of windblown sand by the vegetation. The type of sand determines the flora of the dune system and can range from base-deficient silaceous to strongly calcareous material (Salisbury, 1952).

Dune systems are well drained with the exception of the hollows, or slacks, between the dunes. Studies of coastal sediments have shown that coarser sands, such as those associated with dunes, have fewer absorption sites than do clay minerals, silts and muds. However, sands with a high proportion of carbonaceous particles accumulate more radionuclides than do other coarse-grained sediments (Jones et al., 1984).

The areas most likely to accumulate radionuclides in active dune systems would be the dune slacks where organic matter accumulates and binding sites for radionuclides will be more frequent. As the dune system ages it becomes more stable and the area behind often develops into a heathland. The soil obviously accumulates more organic matter and the retention of spray and aerosols becomes more effective. Dune systems, however, have not been found to be important habitats where large amounts of radionuclides accumulate (Nellis, 1990).

The source of building material for dunes, the intertidal sands, has been studied on the Cumbrian coast by Eakins et al., (1990). These authors found that the intertidal sands contained about 1 per cent of the alpha-emitting activity released from Sellafield up to 1982 but that only in a few cases did the concentrations of man-made alpha-emitters exceed those of natural alpha-emitters.

As dune systems age and new dunes are thrown up nearer the sea the old dunes are stabilized by the accumulation of organic matter and development of vegetation. With the consequent grazing pressure, either by wild or domestic animals, grassland, and eventually by acidification, heathlands often develop. These systems will be subject to spray and aerosol deposition from the sea and the better-developed soil structure will enable greater retention of deposited radionuclides than the sands of the dunes. The greater accumulation of radiocaesium in heathland habitats as well as the ability of heath species to accumulate radionuclides has been well documented (Bunzl and Kracke, 1984; Horrill et al., 1990).

Grasslands also develop behind the many areas of salt marsh on the coast or in river estuaries. These will be subject to aerial deposition and rare inundations by high tides. Whilst the short-lived radionuclides will soon decay, there could be an accumulation of long-lived radionuclides such as ¹³⁷Cs, ^239,240Pu and ^{24
1}Am as a result of periodic flooding.

Vegetation associated with cliffs varies according to exposure to wind and driven spray, the bedrock type and angle. Vegetation is often severely affected by the presence of bird colonies, but on the other hand cliffs can be refuges for plants sensitive to competition or grazing pressure.

The main route for the input of radionuclides is by atmospheric deposition and by spume from breaking waves (Malloch, 1972). Deposition on cliff vegetation will be very variable and in the majority of cases investigated concentrations are only slightly above background levels (Toole et al., 1990).

4.4 MODELS FOR RADIONUCLIDE TRANSPORT IN TERRESTRIAL ECOSYSTEMS

There are numerous models that can be used to simulate the transport of radionuclides in parts of or whole terrestrial ecosystems. Over the last 10 years much attention has been focused on the development of models that can be used to assess the radiological consequences of releases to atmosphere, land or water. This section summarizes some of the recent developments in modelling terrestrial ecosystems with particular attention to radiological assessment models and their underlying databases.

Though considerable use is still made of the concepts of deposition velocity, washout ratio (or rate) and interception fraction (Section 4.1), it became clear after the Chernobyl accident that such models have considerable uncertainties, not least due to the effects of varying physical and chemical form of the depositing radionuclides. There is clear evidence post-Chernobyl for variable concentration of radiocaesium and radioiodine in pasture herbage according to height within the sward (Coughtrey et al., 1990a). Also, the multilayered nature of the canopy within forests leads to differences in interception and retention. The reasons for this relative partitioning are both physical and biological, and the implications for modelling radionuclide transport are very great. Various authors have previously proposed the use of multilayered interception models (e.g. Lassey, 1983); however it is only recently that such models have been tested to any extent (e.g. Pinder et al., 1989) and there do not, as yet, appear to have been attempts to apply such models to post-Chernobyl data.

Much attention has been focused on the deposition and interception of particulates in dry conditions. Events after the Chernobyl accident identified a requirement for a more detailed examination of wet deposition and interception. This requirement also applies in the context of waste disposal assessment where contamination of herbage and other vegetation may occur via irrigation with ground waters (Watkins, 1990). Post-Chernobyl data have been analysed in a number of ways to determine empirical relationships between vegetation content and wet deposition as is apparent in the model summaries included in BIOMOVS documentation (e.g. BIOMOVS, 1991). Muller and Prohl (cited by Watkins, 1990) modelled interception for wet deposition (f_w) as follows:

4.4.2 ABSORPTION, TRANSLOCATION AND LOSS FROM VEGETATION FOLLOWING SURFACE CONTAMINATION

Methods for modelling absorption, translocation and loss from vegetation are highly dependent upon the model structures adopted. These range from a single compartment to represent above-ground vegetation to multi-compartment models that attempt to separate `surface' contamination from `internal' contamination (BIOMOVS, 1991). Nevertheless, there have been a few recent developments in modelling the processes summarized in Section 4.1. Particularly in conditions of wet deposition or when rainfall follows dry deposition, retention on vegetation (whether measured on a content or concentration basis) shows multi-exponential trends. In the case of data based on concentrations, one component of the observed decline curve will reflect dilution by new growth.

Retention half-lives recorded for radiocaesium in the first few months after the Chernobyl accident ranged from 1 to 30 d with some apparent distinction between values recorded for areas primarily affected by wet or dry deposition (Coughtrey et al., 1990a). BIOMOVS (1991) separated data for four sites into two phases of decline with retention half-lives of 8.3 to 12 d for the first 30 d and of 24

173 d thereafter.

No process-oriented models appear to exist for retention, absorption and translocation. All existing models are based on empirical factors derived from experimental or field observations.

The last decade has seen a rapid development of models describing radionuclide transport in soils. Dynamic, process-oriented models have replaced simple empirical or `black box' models (Coughtrey, 1988). Prior to the Chernobyl accident, the rapid developments in modelling had not been accompanied by a similar expansion in data. Chernobyl provided an opportunity to test and further develop many of the models but the timescale of some soil processes implies that it could be several years before adequate data can be acquired for validation purposes. Nevertheless, Chernobyl highlighted two aspects of radionuclide behaviour in soil which require careful consideration. These are:

Though no specific examples appear to exist in the literature, the former can probably be dealt with on the basis of the physical characteristics of the soil concerned (particularly infiltration capacity, see Section 4.1.2). The latter is more difficult to treat without detailed knowledge of the initial chemical characteristics of the deposited material and treatment of subsequent chemical behaviour within soil (Section 4.2.1). Nevertheless, considerable work has been undertaken, particularly in the former USSR. Konoplyov and colleagues (e.g. Konoplyov and Bulgakov, 1991) have used an approach of the type outlined in Figure 4.1 in which the interchanges between `fuel particle', `cationic', 'ion-exchangeable sorbed', `soluble complexes' and `irreversibly sorbed' fractions in soil are modelled as a function of time. Considerable data have been collected for both ⁹⁰Sr and ¹³⁷Cs to develop and test the model. In general, agreement between predictions and observations has been very good.

Models of the type described above are essential to provide long-term predictions for exposure of man via both ingestion and external pathways. For Chernobyl, simple exponential models or those based on diffusion alone have not provided an adequate description of the behaviour of radiocaesium in soils other than in the first few months following the accident. Dynamic models for soil which take account of different chemical fractions are also valuable in predicting the rate of migration from soils to groundwater, the availability for plant uptake, and the potential for resuspension or erosion.

Data concerning resuspension of radiocaesium after the Chernobyl accident have been considered by Garland and Pattenden (1990) who demonstrated a negative correlation between the calculated resuspension factor for radiocaesium and the total level of deposit. In part, this relation could be interpreted as the effect of long-range transport of resuspended material. An alternative explanation is the effect of rainfall in moving the deposited radiocaesium to lower soil layers and thereby rendering it unavailable for resuspension. Observed declines in resuspension factor with time were compared with a range of resuspension models. The models showed a wide divergence over the time interval occupied by the data reflecting, to some extent, the different applications of the models.

Although resuspension factor or rate methods are commonly applied in radiological assessments, other methods have been applied, for example the massloading approach. In areas close to Chernobyl, opportunities exist for further measurements of resuspension and it is understood that these opportunities are being exploited by some groups from the former USSR to develop revised models.

Though the traditional methods of assessing soil-to-plant transfer are extremely useful for long-term assessments assuming equilibrium between plants and soil (Coughtrey et al., 1990b), they are not always useful when dealing with short times after input to soil or with conditions in which biological availability within soils is changing. Over the last few years there have been several developments to provide realistic alternatives to the traditional soil

plant transfer factor. In general, these approaches require coupled soil and plant models in which the soil model allows for processes leading to biological availability. This generally requires complex modelling of soil processes and the incorporation of simulation methods dealing with root uptake from solution. However, few data exist as yet for the radionuclide, soil and plant combinations of interest. A complicating factor in this type of approach is the determination of the extent to which root-absorbed radionuclides are subsequently transferred to those parts of the plant consumed by man or animals.

For radiocaesium and radiostrontium there has been further interest in the development and application of models based on potassium and calcium uptake respectively. In many plant species, potassium and caesium interrelationships are not well understood and whereas models can be developed for overall transfer to vegetation, these are not necessarily adequate for individual component species (Coughtrey et al., 1990c).

Considerable work has been undertaken in recent years on the models which deal with absorption and turnover of radionuclides in animals. Multi-compartment models are often applied in radiological assessments to provide time-varying concentrations in different products (Coughtrey, 1990). In several cases it is possible to extrapolate data for one species to another taking into account differences in body weight, metabolic rate and physiology. One of the critical factors in models of this type is the method for treating initial absorption from the gastrointestinal tract, particularly in cases where significant endogenous secretion occurs (Section 4.2.1.2). Current research programmes sponsored by CEC are directed towards categorizing the processes involved in gut absorption and subsequent turnover in animal tissues. A major uncertainty in many of the existing radiological assessment models is the rate of ingestion of contaminated produce and the nature of its supply (e.g. comparisons between strip-grazing and free-ranging cattle). More attention is therefore being directed towards integrated soil-plant animal models that can allow for a variety of agricultural scenarios.

A further area of development concerns radionuclide turnover in growing animals such as pigs, calves, lambs and chickens. Models similar to those applied to adult animals have been developed in which transfers between compartments are modified according to animal age and weight and in which gastrointestinal absorption is allowed to vary according to age and/or weaning (Cuff, 1989). Such models can provide predictions which appear to be consistent with published data but much more work needs to be done to extend the database for model development and testing.

The various developments summarized above relate to individual components of dose assessment models which deal with all pathways from source to man. In both waste disposal assessment and accident assessment, much more use has been made in recent years of probabilistic consequence models. For terrestrial pathways, some problems do exist in these approaches mainly because many of the parameters used in the models are correlated. This makes adequate parameter specification (range and distribution) a difficult process. Also, in the context of waste disposal assessment, major uncertainties still remain in the methods that can be used to couple biosphere and geosphere models and in the ways in which changes should be incorporated into biosphere models to simulate, for example, the effects of climate change.

Since models used for radiological assessment often incorporate a large number of potential transfer pathways, comparison between model predictions and measurements requires caution. Without a comparison of predictions and measurements at all intermediate steps in the pathways to man, the potential exists for the model to give an apparently correct estimate of the endpoint but for entirely the wrong reason. Self-compensating errors can apply either within an individual pathway to man or between several different pathways. Errors of this type are apparent in the results of the BIOMOVS intercomparison based on post-Chernobyl data (BIOMOVS, 1991). It is anticipated that exercises such as VAMP will help to elucidate the causes of self-compensating errors.

In general, natural and semi-natural ecosystems only provide a direct route for transfer of radionuclides to man via critical pathways. However, natural ecosystems also provide a long-term source of radionuclides to other systems. Particularly significant in this context are forests and alpine or upland grasslands, both of which can receive a higher-than-average deposition relative to surrounding land, and both of which have individual characterstics which tend to retain radionuclides within biological cycles. At the present time there are significant developments in the modelling of alpine or upland grasslands. Examples are work being undertaken in the UK (e.g. Mitchell and Davis, 1990; Crout et al., 1990) in which the results of recent surveys of radiocaesium behaviour are being used to construct both retrospective and future scenarios for radiocaesium concentrations in vegetation and animals. Where forests are concerned, model structures have been proposed (Section 4.3.1) but the major impetus at present is in collation and interpretation of available data, much of which is being co-ordinated by VAMP.

4.5 CONCLUSIONS AND RECOMMENDATIONS

Radionuclides of varying half-life and ecological characteristics enter terrestrial ecosystems by many routes and become widely dispersed within the component parts of the ecosystem.

Terrestrial ecosystems are important to man since they provide the main source of his diet. The majority of the existing information on radionuclide behaviour in terrestrial ecosystems is for temperate agricultural ecosystems. In contrast, much less information exists for coastal systems (other than salt marshes), forests, upland and wetlands, and tropical systems (other than atolls). The effects of chemical form of input remain an overriding area of uncertainty in the behaviour of radionuclides in terrestrial ecosystems.

Dry deposition of particles and gases to vegetation or soil surfaces is well documented and relatively well understood. A possible exception is the role of the abiotic layer (mostly soil particles) on vegetation surfaces in capturing and retaining deposited materials. In comparison with pasture grasses, relatively few data exist for other types of vegetation, particularly agricultural crops and forests. In contrast to dry deposition, practically no information exists on the processes involved in wet deposition to either soils or vegetation, or on the subsequent redistribution of wet-deposited radionuclides between vegetation and soils. This is equally true for all radionuclides other than radiocaesium and is important when considering inputs via either rainfall or irrigation waters. Additionally, deposition during cloud cover (occult deposition) or via snowfall can be important and is not well documented.

Until recenty, relatively few data existed for resuspension other than in arid environments. Chernobyl is extending the database for temperate conditions but more information is required on the relative importance of processes such as soil erosion by wind or water, rainsplash, volatilization, and the effects of burning. The latter could be of particular importance for contamination of forests or natural grasslands where fire is a natural ecological process.

The behaviour of radionuclides in soils determines radiological impact via both terrestrial and aquatic pathways. However, some aspects of radionuclide behaviour in soils are still poorly understood, especially for sandy or organic soils. It is currently possible to simulate the redistribution of radionuclides in soil profiles but the degree of availabilty is largely unknown. Additionally, Chernobyl has highlighted the need for a better understanding of particle migration in soils.

Present studies on transfer between soil and plants take more account than previously of chemical and biological availability in soil, particularly for those radionuclides which show irreversible binding to soil particles and hence long-term trends in plant uptake.

The role of micro-organisms in both soil and foliar uptake has not been studied sufficiently in the context of radionuclide behaviour.

Very few studies have distinguished between the individual processes of interception, retention, absorption and translocation in terrestrial plants and few data exist on the dynamics of partitioning within plants following absorption. The extent to which soil contamination contributes to radionuclide concentrations of plants has been determined in only a few studies.

Seasonality is an important factor, both in terms of the time of year at which contamination occurs and in terms of plant and soil concentrations following the contamination event. Present data are only for accidents which occurred either in early spring or autumn

significant deviations can be expected for an event which occurs during summer or winter.

Present studies on plant-to-animal transfer have identified a requirement for a better understanding of processes involved in absorption and excretion within the gastrointestinal tract and the need to separate true absorption from apparent absorption. This issue can be resolved by further kinetic studies following input of radionuclides to blood and by the re-evaluation of existing data from less controlled studies. Though some recent studies have been undertaken with young animals (e.g. lambs, chicks), much more work is required before the behaviour of radionuclides in these animals can be predicted accurately. As with plants, few studies have considered the effects of intakes of radionuclides attached to soil. Also, detailed information is still required on the seasonal variation in dietary components of free-ranging animals to interpret seasonal changes in body tissues and burdens.

Considering specific radionuclides, most information exists for biologically important fission products (¹³¹I, ⁹⁰Sr and ¹³⁷Cs).Less information exists for other fission products (e.g. isotopes of Ag, Ru, Ce, Zr) and, other than for ⁶⁰Co, for the majority of activation products (e.g. isotopes of Ni, Mn, Zn and Cd). In respect of releases to atmosphere, the behaviour of ¹⁴C and ³H is relatively well documented. However, especially for ¹⁴C and for other radionuclides of potential significance in waste disposal (e.g. ¹²⁹I, ⁹⁹Tc, U-Th series nuclides) much more information is required, especially if these radionuclides enter soil via groundwater pathways.

In the past decade there have been substantial developments in the modelling of radionuclide transport in terrestrial ecosystems, and a hierarchy of models exists for different parts of the ecosystem and for a number of the processes of importance. Nevertheless many of the models rely on empirical evaluation of existing data without due account of the underlying mechanisms and processes. Such models can rarely be applied to circumstances other than those that applied to the original source data. Models can be developed rapidly relative to the timescale for acquisition of experimental data, hence there are few published examples of rigorous testing of those models that have been developed. In this respect, the model intercomparisons that have been undertaken contain little discussion of the effect of model structure on the end-result, despite the fact that agreement between model prediction and measurement can often be good but for entirely the wrong reasons. More emphasis should be placed on the development of agreed structures which incorporate processes known to be important and, in compartment models, in ensuring that the individual components of net fluxes between compartments have been specified correctly.

Future research should be directed at filling these gaps in knowledge and should also include the following topics.

4	Terrestrial Pathways
	Co-ordinator: R. Kirchman
	Contributors: J. N. B. Bell, P. J. Coughtrey, M. Frissel, T. E. Hakonson, W. C. Hanson, D. Horrill, B. J. Howard,
		L. J. Lane, C. Myttenaere, W. L. Robison, C. Ronneau, G. Shaw, W. R. Schell, J. Van Den Hoek,
		A. Konoplyov, and N. Zezina

4.1 Introduction
	4.1.1 Deposition, Interception and Retention
		4.1.1.1 Deposition
		4.1.1.2 Interception
		4.1.1.3 Retention
	4.1.2 Movement in Soils
		4.1.2.1 Sorption
		4.1.2.2 Resuspension
		4.1.2.3 Mass transport and bioturbation
		4.1.2.4 Leaching
	4.1.3 Plant Uptake and Distribution
		4.1.3.1 Foliar absorption
		4.1.3.2 Root uptake
		4.1.3.3 Translocation
	4.1.4 Transfer to, and Metabolism in, Animals
	4.1.5 Semi-natural and Natural Ecosystems
	4.1.6 Seasonal Effects
4.2 Agricultural Ecosystems
	4.2.1 Temperate Zones
		4.2.1.1 Soils and plants
		4.2.1.2 Transfer to and metabolism in animals
	4.2.2 Tropical Zones
4.3 Semi-natural Ecosystems
	4.3.1 Forests
		4.3.1.1 Introduction
		4.3.1.2 Radioactivity transfer and cycling in forest ecosystemsmain pathways
		4.3.1.3 Deposition and interception by the canopy
		4.3.1.4 Retention and leaching of intercepted radioactivity
		4.3.1.5 Behaviour of radionuclides in forest soils
		4.3.1.6 Contamination of the understorey
		4.3.1.7 Effects on forest plants and animals
		4.3.1.8 Conclusions
	4.3.2 Arid Environments
		4.3.2.1 Radionuclide distribution in arid ecosystems
		4.3.2.2 Biotic processes
		4.3.2.3 Summary and conclusions
	4.3.3 Upland Ecosystems
		4.3.3.1 Introduction
		4.3.3.2 Source term
		4.3.3.3 Soils
		4.3.3.4 Vegetation
		4.3.3.5 Animal populations
	4.3.4 Arctic Ecosystems
		4.3.4.1 Study areas
		4.3.4.2 Soils
		4.3.4.3 Vegetation
		4.3.4.4 Herbivores
		4.3.4.5 Carnivores
	4.3.5 Tropical Islands
		4.3.5.1 Radionuclide distribution in the soil
		4.3.5.2 Caesium-137 and strontium-90
		4.3.5.3 Plutonium and americium
		4.3.5.4 Remedial measures for ¹³⁷Cs uptake
	4.3.6 Coastal Ecosystems
		4.3.6.1 Coastal lagoons and swamps
		4.3.6.2 Salt marshes
		4.3.6.3 Beach systems
		4.3.6.4 Sand dunes
		4.3.6.5 Coastal grasslands and heath
		4.3.6.6 Cliff vegetation
4.4 Models for Radionuclide Transport in Terrestrial Ecosystems
	4.4.1 Deposition to Ground and Interception by Vegetation
	4.4.2 Absorption, Translocation and Loss from Vegetation Following Surface Contamination
	4.4.3 Fixation and Transport in Soils
	4.4.4 Resuspension
	4.4.5 Soil-to-plant Transfer
	4.4.6 Plant-to-animal Transfer
	4.4.7 Dose Assessment
	4.4.8 Natural and Semi-natural Ecosystems
4.5 Conclusions and Recommendations

		(4.3)
where: V_w= velocity of water percolation;
	d_s = depth of soil;
	= soil bulk density; and
	=soil water content


Type of surface	Chemically non-reactive compounds			Reactive gases
	0.01-0.1µm	0.11µm	15µm

Smooth surface	0.02	0.01	0.02	0.2
Grassland canopy	0.2	0.1	0.2	1
Woodland canopy	2	1	2


		CAN
Sampling dates	CL	Spruce	Oak

9 December 1988	0.02	0.49	0.096
2 January 1989	0.01	0.31	0.108
27 January 1989	0.01	0.42	0.103
14 March 1989	0.01	0.27	0.067


Crop mass	Amount of rainfall (mm)
(kg/m² fresh weight)	0.1	1.0	5.0	10.0

0.1	1.5	1.2	0.6	0.3
0.5	1.2	0.72	0.24	0.12
1.0	0.97	0.57	0.18	0.09
2.5	0.4	0.33	0.10	0.05


Crop	Part	Bq/kg per kBq/m²
		1987	1988	1989

Winter rye	Grain	0.24	0.12	0.05
Barley	Grain	0.19	0.06	0.04
Oats	Grain	0.65	0.30	0.22
Potato	Tuber	0.16	0.11	0.04


Overstorey
Needles/leaves	10	.1
Twigs (wood)	7	.6
Bark	25	.1
Trunk (wood)	1	.6

Understorey
Mosses	270	515
Lichens	275	510
Heath	500
Fungi	380


1950	1951	1973

27.6 (0.01)^a	12.6 (0.03)^a	0.67 (0.18)^a
n=6	n=3	n=8

^aValue in parentheses is coefficient of variation, CV (standard deviation/mean).


Soil source	Field	Glasshouse

NTS^a Area 11 B	0.13 to 0.16	0.00015
NTS Area 11C	0.045 to 0.34	0.00018
NTS Area 13	0.078 to 0.44	0.00011

^aNTS (Nevada Test Site).


	¹³⁷Cs	⁹⁰Sr	^239,240Pu	²⁴¹Am

Drinking	5.5 (3.4)	6.1 x 10^-3(4.5 x 10^-3)	2.8 x10^-5(7.1 x10^-6)	5.3 x10^-5(1.4 x10^-5)
coconut meat
Drinking	2.1 (1.3)
coconut fluid
Copra	10 (4.2)	4.6 x 10-³ (3.1 x 10^-3)	1.9 x10^-5(6.0 x10^-6)	3.1 x10^-5(8.2 x10^-6)
meat
Breadfruit	0.8 (0.6)	0.057 (0.055)	1.3 x10^-5(4.2 x10^-6)	2.4 x10^-5(6.5 x10^-6)
Pandanus	15 (11)	0.12 (0.041)	2.3 x10^-5 (1.3 x10^-5)	6.3 x10^-5(3.0 x10^-5)


			Years	Sample
Sample	Species	n	collected	wt (g)	⁴⁰K	¹³⁴Cs	¹³⁷Cs

Ravenglass
Pectoral muscles	Greylag goose	1	1981	342.9	184.6	3.3	57.7
	Wigeon	4	1981	151.3±33.2^a	126.9±48.8	8.5±8.3	158.0±148.9
				123.8-172.6	85.8153.5	3.7-15.5	80.7289.3
Livers	Greylag goose	1	1981	52.9	124.0	-	27.8
	Wigeon	4	1981	18.8±3.3	28.1±60.0	-	99.5±52.7
				15.921.1	3.779.6	-	64.4149.9
Morecambe Bay
Pectoral muscles	Wigeon	2	1984-5	106.2	158.7	-	27.0
				69.7142.7	146.5170.6	-	18.9-35.2
Livers	Wigeon	10	1979-85	17.2±3.0	74.4±79.7	-	60.3±26.2
				12.725.7	3.7348.9	-	12.2141.0

^a±= 95 per cent confidence limits about the mean.


Radionuclide	Activity concentration	Locality	Year
	(Bq/kg dw)

⁶⁰Co	97	R. Esk	1981^b
	112	R.Irt	1982^c
⁹⁰Sr	11 000	Eskmeals marsh	1974^a
¹⁰⁶R_u	130 000	Eskholme marsh	1973^a
	4350	R. Esk	1981^b
	18 700	R. Irt	1982^c
¹³⁷Cs	18 000	Eskholme marsh	1978^a
	up to 30 000	R. Irt	1979
	10 260	R. Esk	1981^b
	16 810	R. Irt	1982^c
	10 000	R. Esk	1982
	5003620	R.Duddon	1985
¹⁴⁴Ce	69 000	Eskholme marsh	1973^a
	690	R. Esk	1981^b
¹⁵⁴Eu	4920	R.Irt	1982^c
	300	R. Esk	1981^b
	441	R.Irt	1982^c
¹⁵⁵Eu	260	R. Esk	1981^b
	620	R.Irt	1982^c
²³⁷Np	4	R. Esk	1986
^239/240Pu	3560	R. Esk	1981^b
	10 750	R. Irt	1982^c
	8000	R.Esk	19821988
	2301570	R. Duddon	1985
²⁴¹Pu	79 000	R. Esk	1989
²⁴¹Am	3650	R. Esk	1981^b
	5650	R.Irt	1982^c
	9000	R.Esk	1982-1988
	2902400	R. Duddon	1985

^aHighest value obtained.
^bMean value for an ungrazed salt marsh.
^cMean value for a grazed salt marsh.

f_w = k₁L[1 - exp(-k₂R)] / R	(4.4)
where: L is the leaf area index;
R is the amount of precipitation (mm);
k₁ and k₂are element-dependent constants.