Tải bản đầy đủ - 0 (trang)
III. Factors Affecting the Dynamics of Phosphorus in Runoff and Streams

III. Factors Affecting the Dynamics of Phosphorus in Runoff and Streams

Tải bản đầy đủ - 0trang



Some of the theories developed to describe water movcment in soils can

be applied to evaluate the potential loss of P from various soil types as

a result of subsurface runoff. Gardner (1965) developed equations to describe the movement of nitrate in the soil profile due to leaching. The

chemical interactions that occur between dissolved inorganic P and soil

components (discussed later), when water percolates through the soil,

must also be taken into consideration. Inclusion of a term in the equations

developed by Gardner (1965) to describe the relationship between P in

particulate and aqueous phases is therefore necessary. This could take the

form of a linear adsorption isotherm relevant to the concentrations of dissolved inorganic P maintained in the solution of a particular soil. Biggar

and Corey (1969) have also reviewed the literature on infiltration and

percolation of water in agricultural soils as it pertains to nutrient


The movement of solid phase material in contact with natural waters

occurs during surface runoff and in streams. The amounts of solid material

capable of entering surface runoff will depend on the intensity of rainfall,

physical and chemical attachment between various solid components, and

the amounts and energy of runoff waters (Guy, 1970). It is the energy

of surface runoff or stream water, however, that governs the amounts of

a specific size fraction of particulate materials which will remain in suspension during water flow.

The primary source of particulate material to surface runoff and streams

is eroding soil (Guy and Ferguson, 1970), although in urban areas with

little ongoing development, particulates may be dominated by specifically

urban detrital material (e.g., street litter and dust) and organics derived

from urban vegetation.

The total on-site losses of soil due to sheet and rill erosion are not necessarily delivered to streams. The amount of sediment that travels from a

point of erosion to another point in the watershed is termed the sediment

yield (Johnson and Moldenhauer, 1970). Consequently the Universal Soil

Loss Equation used to predict field soil losses on an average annual basis

(Wischmeier and Smith, 1965) must be corrected when used to predict

sediment loads in streams because deposition of particulates may occur

on the land surface as a result of slope variations before surface runoff

reaches a stream. It is for this reason that estimates of soil loss in surface

runoff from sites within a particular watershed cannot be translated into

total P losses through a knowledge of the total P content of the soil, if

the P loss is to be related to P enrichment of surface waters.

An associated complication arises from the fact that soil P is primarily

associated with the solid phase. As soil erosion is a selective process with

respect to particle size, selectivity has been observed for P loss in surface



runoff. The extent of the selectivity depends on the particle sizes with

which most of the soil P is associated. This observation has led to the

concept of enrichment ratios (ER) , which for P are calculated as the ratio

of the concentration of P in the particulate phase of surface runoff to the

concentration of P in the source of the particulate phase. This effect was

first considered by Rogers (1941), who observed ER values of 1.3 for

total P and 3.3 for “0.002 N H,SO, extractable” P for a silt loam situated

on a 20-25% slope. Other values range from 1.5 to 3.1 for total P

(Knoblauch et al., 1942; Neal, 1944; Stoltenberg and White, 1953),

whereas Massey and Jackson (1952) observed values between 1.9 and

2.2 for “water-soluble plus pH 3 extractable” P for silt loams in Wisconsin.

The selective nature of surface runoff with respect to P is due to selective

removal of fine soil particulates as a result of the energy limitations of

runoff and the fact that a large percentage of total soil P is frequently

associated with clay-sized material (Scarseth and Chandler, 1938; Williams

and Saunders, 1956; Syers et al., 1969). Greater selectivity of fines and

consequently particulate P will occur as the energy of surface runoff decreases. Stoltenberg and White (1953) observed that as precipitation disposed of through surface runoff decreased from 70 mm to 0.25 mm per

hour, the clay content of eroded material from a soil with a clay content

of 16-18% increased from 25% to 60%. This has obvious implications

in relation to the nature of the sediment load carried by a stream and the

interactions of P between the solid and aqueous phases, particularly during

periods of surface runoff. It should be pointed out, however, that although

the P content of the sediment load may increase as surface runoff diminishes, as may be predicted from the work of Stoltenberg and White

(1953), the total P load may not change, or may even decrease, owing

to lower sediment loads.

The particulate material carried in streams may be divided into bed load

and wash or suspended load. The bed load, which may also have a contribution from existing stream sediment, is that which moves along or close

to the stream bed, whereas the wash load is maintained in the flow by

turbulence (Johnson and Moldenhauer, 1970). By inference from the selectivity of surface runoff for fine soil particulates, the wash load will be

high during surface runoff events. Furthermore, Johnson and Moldenhauer

(1970) suggested that the wash load travels at about the same velocity

as the water with which it is in contact. Consequently, P associated with

the clay- and silt-sized particulates constituting the wash load will move

between any two points in the stream profile at the same speed as the

ambient dissolved forms of P.

Increased turbulence in streams during high flow, or arising from an

increasing gradient, will tend to maintain in suspension particle sizes more



characteristic of the bed load, and may even resuspend existing stream

bed sediment. In a study of total P loads in the Pigeon River, North Carolina, Keup (1968) noted that an increase in gradient from 2.81 to 4.35

m/km, over which no tributaries entered the main stream, resulted in a

90.8 kg/day increase in the total P load carried.

It appears that in the majority of cases a large proportion of particulate

P in streams arises from soil erosion. Phosphorus may be stored in stream

bed sediments, but unless the stream is actively aggrading, the amount of

P stored will be less than the inflow (Keup, 1968). That which is stored

is liable to resuspension and transport owing to turbulence during periods

of high flow.



1 . Nature of Soil P

Soil P may be divided into two broad categories: inorganic P, namely,

that associated with soil mineral particles; and organic P, which forms an

integral part of the soil organic matter fraction.

a. Inorganic P . O n the basis of solubility product criteria, it has been

postulated that discrete phase crystalline Fe and A1 phosphates exist in

noncalcareous soils (Kittrick and Jackson, 1956; Hemwall, 1957; Chakravart and Talibudeen, 1962). The general occurrence of discrete Fe and

A1 phosphates seems doubtful on the basis of the ion product data presented by Bache (1964) and the experimental observations of Hsu

(1964). It is now generally accepted that secondary inorganic P in many

soils exists primarily in association with oxides and hydrous oxides of Fe

and Al, as surface-bound forms or within the matrices of such components.

However, that discrete Fe and A1 phosphates are formed as temporary

phases in the vicinity of phosphate fertilizer particles due to conditions

of localized high acidity and P concentration is well established (Lindsay

and Stephenson, 1959; Huffman, 1969). Such compounds will not be

stable as the dissolved inorganic P concentration in the soil solution or

aqueous portion of other soil-water ecosystems decreases.

The calcium phosphate mineral, apatite (Shipp and Matelski, 1960) and

calcic fertilizer-soil reaction products (Huffman, 1969) have been identified in soils. The amounts of apatite are appreciable only in weakly

weathered soils (Williams et al., 1969), as predicted by the weathering

indices of Jackson ( 1969). Calcic fertilizer-sail reaction products may be

present in neutral and calcareous surface soil horizons, and their importance in maintaining high concentrations of dissolved inorganic P in

soil-water ecosystems should not be overlooked.



Consequently three basic forms of inorganic P may exist in unfertilized

soils (Syers and Walker, 1969; Williams and Walker, 1969): apatite,

which is a discrete phase P compound; P sorbed on the surfaces of Fe,

Al, and Ca soil components (nonoccluded); and P present within the

matrices of Fe and A1 components (occluded). In fertilized soils, a variety

of P fertilizer-soil reaction products may exist as transient phases. As the

solubility product of pure apatite in water is low (0.03 pg per milliliter

at pH 7, Stumm, 1964) and the P held within the matrices of Fe and

A1 components is virtually chemically immobile, except under reducing

conditions in the case of Fe, major emphasis should be directed toward

the reactions involving P in solution and that sorbed on the surfaces of

Fe, Al, and Ca components as well as the release of P due to dissolution

of fertilizer-soil reaction products.

b. Organic P. Elucidation of the composition of soil organic P is restricted by lack of extractants capable of removing organic P from soils

in a relatively unaltered form and by the inadequacy of current methods

for mildly degrading extracted organic P-organic matter complexes. Existing data indicate that most of the organic P in soils is associated, in an

ill-defined manner, with the humic and fulvic acid complex of soil organic

matter (Anderson, 1967). Of the specific forms of organic P that have

been identified in soils, inositol phosphates are present in largest relative

amounts, comprising up to 60% of the total organic P (Anderson, 1967;

Cosgrove, 1967; McKercher, 1969). Other specific organic P compounds

are present in soil in much lower quantities: nucleic acids account for

5-lo%, and other phosphate esters, such as phospholipids, sugar phosphates, and phosphoproteins, for less than 1-2% (McKercher, 1969).

2. Sorption of Dissolved P by Soils

Whenever water containing a particular concentration of dissolved P

comes into contact with soil material, there is a possibility for sorption,

desorption, or dissolution reactions to take place. The types of reactions

are the same regardless of whether they occur under conditions existing

in the soil profile, surface runoff, or streams. Although in some cases biological assimilation may initially affect the distribution of P between dissolved and particulate phases of soil-water systems, the distribution of P

between these phases will be determined by the nature of the inorganic

particulates and the concentrations of dissolved P in solution (Keup, 1968;

McKee et al., 1970; Ryden et al., 1972b).

a. Inorganic P. It has been demonstrated that the uptake or sorption

of P from solution by soils is significantly related to the presence of shortrange order (amorphous) oxides and hydrous oxides of Fe and A1 (Williams et al., 1958; Gorbunov et al., 1961; Bromfield, 1965; Hsu, 1964;

Saunders, 1965; Syers et al., 1971). Furthermore, “pure” oxides and hy-



drous oxides of Fe and Al, and short-range order aluminosilicates have

also been shown to be particularly effective in the sorption of inorganic

P from solution (Gastuche et al., 1963; Muljadi et al., 1966; Hingston

et al., 1969). The sorption of inorganic P by Fe and A1 oxides and hydrous

oxides is known to be rapid, as is the sorption of P by soils. Furthermore,

short-range order Fe and A1 oxides and hydrous oxides are ubiquitous in

soils (Hsu, 1964), their relative amounts depending on parent material,

climatic and drainage conditions, and occur mainly as coatings on other

soil components. Shen and Rich (1962) and Jackson (1963) have noted

the occurrence of A1 hydroxypolymers and Dion (1944), and Roth et al.

( 1969) have reported the presence of F e oxide and hydrous oxide coatings

on clay mineral surfaces. Such coatings, in conjunction with the greater

surface area of the clay fraction compared to that of the other particle-size

fractions in a soil, explain the observation of Scarseth and Chandler (1938)

that up to 50% of the total P in soils may be associated with the

the clay fraction, as well as the enrichment ratio effect for P as a result

of soil erosion.

Attempts have been made to correlate P sorption with the clay content

of soils (Williams et al., 1958). Correlations between P sorption and clay

content after removal of Fe and A1 oxides and hydrous oxides often have

been poor. Better correlations may be expected if P sorption is related

to the content of water-dispersed clay. The sorption of P by water-dispersed clay and silt of soils has obvious implications to reactions occurring

between dissolved and particulate P in surface runoff and streams.

Sorption of inorganic P by CaC03 has also been demonstrated (Cole

et al., 1953). The nature of the surfaces of calcite in calcareous soils may

be very different from those of pure calcite (Buehrer and Williams, 1936;

Lahav and Bolt, 1963; Syers et al., 1972).

The sorption of dissolved inorganic P by soils may be described by sorption isotherms similar to that shown in Fig. 2. Numerous workers have

also shown that sorption may be described by some of the adsorption isotherms developed to describe gas adsorption by solids (Russell and Prescott, 1916; Olsen and Watanabe, 1957; Rennie and McKercher, 1959;

Syers et al., 1973). Similar observations have been made for the sorption

of inorganic P by soil components such as kaolinite and short-range order

Fe and A1 oxides and hydrous oxides (Gastuche et al., 1963; Muljadi et

al., 1966; Kafkafi et al., 1967). Although these studies have been useful

in describing relationships between various soils and soil components with

respect to their P sorption capacities, they have provided little information

regarding P sorption behavior from solutions containing the low dissolved

inorganic P concentrations characteristic of most soil-water ecosystems,

largely because of the high levels of added P used (Ryden et al., 1972b).

Furthermore, Syers et al. (1973) obtained two linear Langmuir relation-



sorbed ( f )


sol I

released (3

FIG.2. Typical isotherm for the sorption of added inorganic phosphorus by

a soil. E = equilibrium P concentration. (From White and Beckett, 1964.)

ships which intersected at equilibrium P concentrations varying from 1.5

to 3.2 pg P/ml, for three contrasting soils-an observation that probably

invalidates interpretations of P sorption made from many previous studies

where high levels of added P were used.

The study of White and Beckett (1964), conducted at initial dissolved

inorganic P concentrations, comparable to those existing in soil-water ecosystems, provides a useful basis for understanding the interactions between

aqueous and particulate phases of P in runoff and streams. Figure 2 illustrates the principle of the approach used. White and Beckett (1964) defined the intersection of the P sorption isotherm and the abscissa, the

“equilibrium phosphate potential” ( 5 p C a pH,PO,) , abbreviated to

“equilibrium P concentration” by Taylor and Kunishi ( 1971) . The intersection is equivalent to the inorganic P concentration in the ambient aqueous phase when there is no net sorption or release of P, i.e., AP = 0. This

is a point of reference which provides a predictive estimation of sorption

or release of P should the P concentration in solution change. Furthermore,

the average slope of the sorption curve over a given final P concentration

range provides information on the ability of the soil to maintain the P

concentration at the equilibrium P concentration. The steeper the slope,

the closer will the final P concentration be to the equilibrium P concentration. The slope of the curve, although not related to total P sorbed, is

related to the extent to which that soil may sorb P over the concentration

range considered. The potential of this approach in predicting the chemical

mobility of P in soil-water systems is clearly evident and has been used

with regard to streams by Taylor and Kunishi (1971) and Ryden et al.

(1972a,b) for rural and urban soils, respectively.

The desorption of sorbed P from soils is not as simple as may be inferred from the sorption-release relationships obtained by White and




Beckett (1964). In fact very few studies have been reported regarding

the desorption of sorbed P, and those reported by Syers et al. (1970)

and Ryden et al. (1972a), involved desorption following sorption of P

from solutions containing P concentrations in excess of those commonly

found in soil-water ecosystems.

In studies involving the sorption of P by kaolinite from solutions containing realistic inorganic P concentrations, Kafkafi et al. (1967) observed

that initially all the sorbed P was isotopically exchangeable. During a subsequent washing or desorption step, however, a portion of the sorbed P

became nonexchangeable, or “fixed,” this portion being dependent upon

the amount of P sorbed, the number of washings, and the nature of the

previous P sorption cycle. Sorption of P was represented by either onestep sorption from a range of solutions of different initial P concentration

or by successive additions of small amounts of dissolved inorganic P. Both

these types of P sorption, as well as an effect analogous to washing, could

occur in soil-water ecosystems.

6. Organic P . Although the mechanisms involved in the retention of

organic P by soils have not been established fully, there is evidence that

inositol hexaphosphate, and possibly other organic P compounds, are retained by a precipitation rather than a sorption reaction. Nevertheless, removal of dissolved organic P from solution appears to be a rapid process.

Pinck et al. (1941 ) reported that many commonly occurring water-soluble

organic phosphates, e.g., salts of glycerophosphate, hexose diphosphate,

and nucleic acids, become nonextractable with water at almost the same

rate and as completely as dissolved inorganic P.

The retention of water-soluble organic P by sorption reactions may

occur by at least two basically different mechanisms (Sommers et al.,

1972). Goring and Bartholomew (1950) observed that removal of “free

iron oxides” considerably reduced the amount of fructose 1,6-diphosphate

sorbed by subsoil material, suggesting that the sorption of organic P may

occur through orthophosphate groups by a similar mechanism to that for

inorganic P. It is also possible that organic P can be retained by interaction

of the organic moiety of the phosphate ester with inorganic soil components. For example, nucleic acids and nucleotides are protonated at pH

5 (Jordan, 1955) and could consequently be retained on clay surfaces

by displacement of exchangeable cations. Furthermore, physical adsorption, also through the organic portion of the molecule, is possible, particularly if the molecular weight of the compound is high, as suggested by

Greenland (1965). In such cases retention is weak and is accomplished

by van der Waals and ion-dipole forces. Greaves and Wilson (1969) have

implicated physical adsorption in the retention of nucleic acids by montmorillonite. It is also possible that retention occurs indirectly through other



soil organic compounds such as fulvic and humic acids after interaction

of organic phosphates with these species (Martin, 1964).

The desorption of sorbed organic P has not been extensively studied.

The hypothesis that inorganic P added to soils displaces sorbed organic

P to solution (Latterell et al., 1971) was not supported by the data presented by Wier and Black (1968). Although organic P may be leached

from soils, it appears that a large proportion of that removed may not

be in a dissolved form. After incubating sucrose with ammonium nitrate

in the upper portion of a calcareous soil, Hanapel et al. (1964) found

that most of the organic P removed by leaching was present in a particulate

rather than a dissolved form.

3. Chemical Aspects of P in Subsurface and Groundwater Runoig

Losses of P in subsurface and groundwater runoff have been considered

minimal in the past, but, as will be discussed later, such losses can amount

to a significant proportion of losses from agricultural land, and possibly

a major proportion from forest lands. The supposition that P losses in subsurface and groundwater runoff are low probably stems from the concept

of P immobility based on the P sorption properties of soils using added

inorganic P concentrations far in excess of those normally present in the

soil solution.

It is of interest to note that many of the reported mean concentrations

of dissolved inorganic P in subsurface runoff are within the range of values

expected to be maintained in the soil solution. Pierre and Parker (1927)

reported values ranging from 0.020 to 0.350 pg P/ml, with an average

of 0.090 pg/ml, for several surface soils from the southern and midwestern

states of the United States. These workers also noted that dissolved inorganic P concentrations could be maintained at a fairly constant level. Barber et al. (1963) reported similar values for the upper 15 cm of 87 soils

from the midwestern United States, with an average of 0.180 pg of P per

milliliter; the frequency distribution of the values obtained, however, suggested a mode of between 0.040 and 0.060 pg of P per milliliter.

As water percolates through the soil profile, there tends to be a “chemical sieving” of dissolved inorganic P (Black, 1970). This arises as a result

of the sorption of inorganic P by soil components. The low concentrations

of P found in groundwater runoff, which has experienced the maximum

effects of deep percolation with concomitant increase of contact with

P-deficient particulates of the subsoil, are undoubtedly a direct result of

the chemical sieving effect. The principle of this effect is illustrated by

other data presented by Barber et al. (1963). For the same 87 soils mentioned previously, the average dissolved inorganic P concentration at a

depth of 46-61 cm was 0.089 pg/ml, less than half that for the upper



0-15 cm. Another illustration is observed in results presented by Ryden

et al. (1972a) for the P sorption properties of successive soil horizons

of a Miami silt loam profile. The concentrations of dissolved inorganic

P maintained in solution after shaking with solutions of different initial

added inorganic P concentrations at a solution: soil ratio of 40: 1 are given

in Table I.


Dissolved Inorganic Phosphorus (P) Concentrations Maintained

by Soil IIorizoiis of Miami Silt Loam after Equilibration

with Solutions of Different Initial Added Inorganic

P Concentrationsn


Depth (cm)

Initial P conc.


Final P coiic.















Data extrapolated from Ryden et



The concentration of dissolved inorganic P in subsurface and groundwater runoff will depend on the nature and amounts of P-retaining components in the profile, the surface area exposed to percolating waters, and

the ease of percolation which affects the contact time of dissolved inorganic

P with the retaining components. In studies of P leaching through columns

of organic soils in the laboratory, Larsen et al. (1958) observed that P

retention, measured by srP autoradiographs, was closely correlated with

the total hydrous Fe and A1 oxide (“sesquioxide”) content. Similarly,

losses of P due to leaching through a deep siliceous sandy soil were demonstrated in W. Australia by Ozanne (1963). When 225 kg/ha of 32P-labeled

superphosphate was broadcast during winter on a fallow sandy soil, over

50% of the P had penetrated to more than 1 m below the surface within

38 days, during which 230 mm rain had fallen. Ozanne (1963) also demonstrated that the potentially large losses of P to subsurface and groundwater runoff from sandy soils compared to that from loamy soils were

due to quantitative rather than qua1itativ.e differences in P-retaining


Although major emphasis has been placed on P losses in surface runoff,

it appears that losses of P to subsurface and groundwater runoff, although

of little significance from an agricultural standpoint, may under certain

conditions constitute a significant loss of P from agricultural watersheds

in terms of the P enrichment of surface waters, as will be discussed



later. Losses of P to subsurface and groundwater runoff are even more

difficult to evaluate than those in surface runoff and demand further investigative attention.

4 . Chemical Aspects of P in Streams

As discussed previously, surface runoff from agricultural land constitutes

a heterogeneous and relatively short-lived system. Any attempt to consider

the distribution and chemical mobility of P between solid and aqueous

phases before entry into the receiving stream would be pointless as a new

and more homogeneous system is rapidly established. Surface runoff in

urban areas is somewhat different because in most cases it is channelized

shortly after origin by alteration of surface drainage patterns; under such

circumstances it is analogous to a stream in an artificial channel. Consequently, the chemical mobility of P will be discussed from the standpoint

of the stream environment.

The potential of suspended particulates derived from eroding soil to

modify the dissolved inorganic P concentration of streams has been suggested by Taylor ( 1967) and Biggar and Corey (1969). Wang and Brabec

(1969) also implied that inorganic P was sorbed by suspended particulate

material from observations of dissolved inorganic P concentrations in the

Illinois River at Peoria Lake.

An evaluation of the possible effects of eroded soil materials on the dissolved inorganic P concentrations of streams may be obtained from P

sorption studies (Taylor and Kunishi, 1971; Ryden et al., 1972a,b). It

is essential, however, that conditions realistic of those existing in streams

are used if meaningful results are to be obtained (Ryden et al., 1972a).

Widely differing interpretations can be made as solution: soil ratios and

initial dissolved inorganic P concentrations are changed from those conventionally used in P sorption studies to those realistic in terms of the stream

environment (Fig. 3 a-c). The data in Fig. 3a suggest that inorganic P

released from the A1 horizon, which contained a P fertilizer-soil reaction

product, would be largely resorbed by the noncalcareous B1 horizon and

to some extent by the calcareous 3C1 horizon, should the horizons erode

together. Sorption studies employing low initial added inorganic P concentrations and a wide (400: 1 ) so1ution:soil ratio (Fig. 3c) indicate that

the B1 horizon has a much lower ability to remove dissolved inorganic

P from solution than expected, this being equal to or only slightly greater

than that of the 3C1 horizon. In fact for mixtures of varying ratios of

A1 and B l , and A1 and 3C1 horizons, it was found (Ryden et al., 1972b)

that the latter mixtures were able to maintain lower dissolved inorganic

P concentrations than the former. The conditions used by Ryden et al.

(1972a,b) to predict the potential of eroding soils to modify the dissolved












+ Ir

















Final dissolved inorganic P Concentration Wgll)

FIG. 3. Sorption of added inorganic phosphorus by horizons of a Miami silt

loam profile from solutions of varying initial dissolved inorganic P concentrations

and at varying so1ution:soil ratios. ( a ) High added P (0-6 pg/ml) and narrow

so1ution:soil ratio (50: 1 ) . ( b ) Low added P (0-0.2 pg/ml) and narrow solution:soil

ratio ( 4 0 : l ) . (c) Low added P (0-0.2 pg/ml) and wide so1ution:soil ratio ( 4 0 0 : l ) .

[From Ryden et al. (1972a), reproduced with permission of the American Society

of Agronomy.]

inorganic P concentrations of streams, gave results comparable to those

obtained in simulated stream systems using a solution: soil ratio of 1000:1

This is equivalent to a sediment concentration of 1000 mg/liter, which

lies well within the range of values cited by Guy and Ferguson (1970)

and Johnson and Moldenhauer ( 1970).

The P sorption studies reported by Taylor and Kunishi (197 1) and

Ryden et al. (1972a,b) involved closed systems, i.e., soil in contact with

the same aqueous phase. This may be justified on the grounds that the

wash load of a stream travels at the same velocity as the water in which

it is suspended (Johnson and Moldenhauer, 1970), as discussed


Tài liệu bạn tìm kiếm đã sẵn sàng tải về

III. Factors Affecting the Dynamics of Phosphorus in Runoff and Streams

Tải bản đầy đủ ngay(0 tr)