endosulfan transport: ii. modeling airborne dispersal …dust.ess.uci.edu/ppr/ppr_rba01.pdf ·...

Endosulfan Transport: II. Modeling Airborne Dispersaland Deposition by Spray and Vapor

M. R. Raupach,* P. R. Briggs, N. Ahmad, and V. E. Edge

ABSTRACT elucidating management implications. The first paper(Raupach et al., 2001; henceforth Paper I) provides anEndosulfan (C9H6O3Cl6S; 6,7,8,9,10,10-hexachloro-1,5,5a,6,9,9a-overview, in which process models and data are com-hexahydro-6,9-methano-2,4,3-benzodioxathiepin 3-oxide) and other

agricultural chemicals can be transported from farms to rivers by bined to assess the contributions of each major transportseveral airborne pathways including spray drift and vapor transport. pathway to riverine endosulfan concentrations. This pa-This paper describes a modeling framework for quantifying both of per provides modeling details for the airborne pathways.these airborne pathways, consisting of components describing the Parts of the work have appeared in several technicalsource, dispersion, and deposition phases of each pathway. Through- reports (Raupach et al., 1996; Raupach and Briggs, 1996,out, the framework uses economical descriptions consistent with the 1998; Briggs et al., 1998).need to capture the major physical processes. The dispersion of spray

The present specific treatment of the major airborneand vapor is described by similarity and mass-conservation principlespathways (spray and vapor) has several motivations.approximated by Gaussian solutions. Deposition of particles to vege-First, there is a need for a simple and robust frameworktation is described by a single-layer model incorporating contributionsfor describing airborne pathways within the broaderfrom settling, impaction, and Brownian diffusion. Vapor deposition

to water surfaces is described by a simple kinetic formulation depen- context of a complete integrative assessment of all trans-dent on an exchange velocity. All model components are tested against port pathways, undertaken in Paper I. Second, no ade-available field and laboratory data. The models, and the measurements quate theoretical framework has hitherto existed for theused for comparisons, both demonstrate that spray drift and vapor vapor transport pathway. Third, many of the governingtransport are significant pathways. The broader context, described in processes and conditions are common to both pathways,another paper, is an integrative assessment of all transport pathways including turbulent dispersion and some aspects of(both airborne and waterborne) contributing to endosulfan transport

source and deposition processes, so economy of descrip-from farms to rivers.tion and proper comparisons between pathways are fa-cilitated by using a single framework for these com-mon processes.In the irrigated cotton (Gossypium hirsutum L.) indus-

Airborne transport pathways all involve three se-try in northern New South Wales, Australia, the in-quential processes, associated with (i) the source of thesecticide endosulfan is widely applied by aerial sprayingtransported entity (spray droplets, contaminant vapor,from November to January. A major water quality mon-or contaminant-bearing dust particles); (ii) the disper-itoring program (Cooper, 1996; Muschal, 1997, 1998)sion of the entity by wind and turbulence in the atmo-has shown that during and just after this spraying season,sphere; and (iii) the deposition of the entity to the waterendosulfan concentrations in rivers near and down-body. This paper assembles a suite of simple, compatiblestream of cotton-growing areas are broadly in the rangealgebraic descriptions for the component processes, and0.02 to 0.2 mg L21, with occasional higher peaks. Thesetests these descriptions both separately and togethervalues significantly exceed the present Australian envi-with available experimental evidence. The major focusronmental guideline for protection of ecosystems, cur-is on the spray and vapor pathways, since the dust path-rently 0.01 mg L21 (Australian and New Zealand Envi-way is not significant in the present context (Paper I).ronment and Conservation Council, 1992). To manage

The notation follows Paper I: subscripts “a” and “w”and ameliorate such environmental contamination, it isdistinguish concentrations and other properties in airvital to understand the pathways by which agriculturaland water, and the a, b, and sulfate species are distin-chemicals such as endosulfan move through the environ-guished by superscripts a, b, or g (where g denotes thement. Knowledge of the relative magnitudes and behav-sulfate). Thus, the total endosulfan concentrations inior patterns of these pathways is needed both for effec-air and water (in kg endosulfan m23 ) are respectivelytive management of major pathways and also to avoidCa 5 Ca

a 1 Cba 1 Cg

a and Cw 5 Caw 1 Cb

w 1 Cgw. Aexpensive efforts aimed at closing minor pathways.

superscript s denotes an arbitrary species.This paper is the second of a pair, which have theoverall aims of quantifying the magnitude, behavior,and relative importance of each of four major pathways THEORY(spray drift, vapor transport, dust transport, and runoff)

Dispersion of Particles and Vaporby which endosulfan can move from farms to rivers, andTo determine contaminant transport via any airborne path-

M.R. Raupach and P.R. Briggs, CSIRO Land and Water, Canberra, way (spray drift, vapor transport, or dust transport), it is neces-GPO Box 1666, Canberra, ACT 2601, Australia. N. Ahmad, Austra- sary to calculate deposition from a cloud of particles or gaslian Water Technologies, Sydney, NSW, Australia. V.E. Edge, NSW dispersing in the air. Here we use a simple model based onAgriculture, Locked Bag 21, Orange, NSW 2800, Australia. Received

mass conservation and a Gaussian-plume assumption. The8 Oct. 1999. *Corresponding author ([email protected]).starting point is the conservation equation for a scalar entityin the air:Published in J. Environ. Qual. 30:729–740 (2001).

729

730 J. ENVIRON. QUAL., VOL. 30, MAY–JUNE 2001

Physically, this means that Q equals the total mass of scalar]C]t

1 u]C]x

5 2]fy

]y2

]fz

]z1 S [1] (the time-integrated mass flux) crossing the yz plane just dow-

nwind of x 5 0. However, as the scalar travels downwind itis removed by deposition to the surface. To describe this, wewhere x, y, and z are along-wind, cross-wind, and verticalconsider the mass fraction s(x):position coordinates (with z 5 0 at the ground), t is time, C

is the scalar concentration, fy and fz are the cross-wind andvertical flux densities, S is the source density, and u(z) is the s(x) 5

1Q

#∞

0

#∞

2∞

u(z) C(x, y, z) dy dz [6]wind velocity. In this section the subscript “a”, denoting airconcentrations, can be omitted without confusion. Fluxes in

which is the total mass of scalar crossing the yz plane at x,the air (fy and fz in Eq. [1]) are distinguished from fluxesnormalized by the original mass Q of scalar at x 5 0. Clearly,entering water bodies (Fw(i) in Eq. [1] of Paper I) because fzs(0) 5 1. An equation for the evolution of s(x) can be foundis positive upward but Fw(i) is positive downward across theby integrating Eq. [3] over the yz plane. Using Eq. [3] to [6],water surface for the spray, vapor, and dust pathways. All oftaking Wd as uniform over all x and y, and noting that fluxesthe quantities C, S, fy, and fz are functions of x, y, z, and tvanish as y → 6∞ and z → 1∞, we obtain:in general, but the wind speed u is assumed to vary only

with height z (that is, the wind field is steady in time andQ

dsdx

5 2 #∞

2∞

D(x, y) dy 5 2Wd #∞

2∞

C(x, y, zr ) dy [7]horizontally homogeneous).We restrict attention to “puff” releases of scalar, defined

as releases that are localized in time (but not necessarily instan- It is necessary to close this equation by relating the finaltaneous), so that C, S, fy, and fz have finite integrals of the concentration integral to s(x). This can be done generally byform: means of a similarity hypothesis about the concentration field

(as will be shown in a later paper), but here it is sufficient toC(x, y, z) 5 #

1∞

2∞

C(x, y, z, t) dt [2] use the much more restricted assumption that the concentra-tion field is Gaussian. In a Gaussian puff from a point releaseof mass Q at location x 5 0, y 5 0, and release height z 5where a caret denotes integration over all time. Then, timehs, in a mean wind of speed u along the x axis, the time-integration of Eq. [1] yields:integrated concentration is (Csanady, 1973; Hanna et al.,1982):

u]C]x

5 2]fy

]y2

]fz

]z1 S [3]

C(x, y, z) 5Q

2p u sY sZ

exp12y2

2s2Y2 3exp12(z 2 zs)2

2s2Z

2The time-integrated downward flux at the ground is thedeposition D(x,y) 5 2fz (x,y,0), with units (kg m22 ). Thisis specified by a deposition velocity Wd, which relates the 1 exp12(z 1 zs)2

2s2Z

24 [8]downward flux of scalar at the surface to the concentrationat a reference height zr just above the surface:

where sY(x) is the plume width, sZ(x) is the plume depth,and zs(x) is the height of the plume centroid. We can refer2fz(x, y, 0, t) 5 WdC(x, y, zr, t) orto sY(x), sZ(x), and zs(x) as "plume" characteristics because,

D(x, y) 5 WdC(x, y, zr ) [4] as noted above, the time-integrated concentration field froma puff is congruent with the concentration in a steady plume.Equation [4] is the lower boundary condition on Eq. [1]If the scalar is settling under gravitation with terminal velocityand [3]. The initial condition is that C is zero far upstreamWt, then the plume centroid can be assumed to obey zs 5and as t → 2∞. We note the congruence between a time-max(hs 2 xWt/u, 0), while for a non-settling scalar materialintegrated concentration field for a puff release and the con-such as a gas, Wt 5 0 and zs 5 hs. Empirical forms for sY(x)centration field for a steady plume release with the sameand sZ(x) are often specified in terms of Pasquill stabilitysource geometry; the former is described by Eq. [3] and theclasses, for example by Hanna et al. (1982) as summarized inlatter by Eq. [1] without the time derivative term, with similarTable 1. The relationship of these classes to more physicalboundary conditions in each case.measures of stability, such as the dimensionless RichardsonThe deposition velocity Wd is a crucial parameter. In gen-number, the Monin–Obhukov stability parameter, and theeral, it is a conductance (a transfer coefficient with the dimen-terrain roughness length, are given in Golder (1972) andsion of velocity) specifying the air-to-surface flux of particlesHanna et al. (1982). When the integral in Eq. [7] is evaluated(2fz ) through an equation of the form 2fz 5 Wd(Cr 2 Cs )using Eq. [8], we obtain:where Cs and Cr are the air concentrations at the surface and

at a reference level zr. For particles being deposited onto a dsdx

5 212p2

1/2 Wd s(x)sZ(x) u

exp12z2s(x)

2s2Z (x)2 [9]surface it is generally assumed that Cs 5 0, yielding Eq. [4].

The deposition velocity is controlled by different processesfor particles and gases, leading to different models for the two which is an ordinary differential equation fully specifying s(x),cases as described below. given the starting condition s(0) 5 1. This is similar to the

Suppose now that mass Q of scalar is released from a source result of the source-depletion method (Hanna et al., 1982),in the yz plane at x 5 0, not necessarily instantaneously or at which calculates s(x) by using standard Gaussian formulae asa single point (y,z), but from a source sufficiently localized if the source strength were Qs(x). The present derivationthat S(x,y,z,t) has a finite integral over t, y, and z, and S is makes the assumptions easier to identify and generalize.zero except at x 5 0. Then an integral mass balance over the Equation [9] is easily solved numerically once hs, u, Wd, Wt,yz plane x 5 0 shows that: and sZ(x) are specified. Of these, hs and u are measured

parameters, and sZ(x) is given in Table 1. The terminal veloc-ity Wt is zero for a gas, while for particles it depends on particleQ 5 #

∞

0

#∞

2∞

u(z) C(0, y, z) dy dz [5]diameter, particle-to-air density ratio, and air viscosity, and

RAUPACH ET AL.: ENDOSULFAN TRANSPORT: II 731

Table 1. Formulae recommended by Hanna et al. (1982) for thewidth sY(x ) and depth sZ(x ) of a Gaussian plume from a pointsource near the ground, as a function of downwind distance x(km), valid for 0.1 , x , 10 km. A "super-unstable" class Zhas been added.

Plume depthPasquill stability class Plume width sY(x ) sZ(x )

Z (super-unstable) 0.40x(1 1 0.1x )21/2 0.40xA (very unstable) 0.22x(1 1 0.1x )21/2 0.20xB (moderately unstable) 0.16x(1 1 0.1x )21/2 0.12xC (slightly unstable) 0.11x(1 1 0.1x )21/2 0.08x(1 1 0.2x )21/2

D (neutral) 0.08x(1 1 0.1x )21/2 0.06x(1 1 1.5x )21/2

E (slightly stable) 0.06x(1 1 0.1x )21/2 0.03x(1 1 0.3x )21

F (moderately stable) 0.04x(1 2 0.1x )21/2 0.016x(1 1 0.3x )21

can be calculated by standard methods summarized in Mal-colm and Raupach (1991). The model for the deposition veloc-ity Wd is described in the next subsection. In the restrictedcase where sZ(x) 5 sZ0 1 ax (with constant a) and the releaseheight hs 5 0 (so that zs 5 0), Eq. [9] has the analytic solution:

s(x)s(0)

5 1sz(x)sZ0

22b

, b 5 12p2

1/2 Wd

au[10]

which provides a useful check on numerical solutions. We usethe numerical solution in practice, because it is not subject tothe restrictions on Eq. [10].

Once s(x) is determined, the deposition D(x,y) from a pointrelease of mass Q of scalar is given from Eq. [4] and [8] by:

Fig. 1. Results from particle dispersion model for four different parti-D(x, y) 5 WdC(x, y, zr ) 5

Qs(x)p u sY sZ

exp12y2

2s2Y2 [11] cle size distributions (log-normal with median particle diameter 5

80, 160, 240, and 320 mm and standard deviation ln(1.8) through-out). Conditions assumed throughout: canopy height hc 5 0.50 m;where the exponentials involving z in Eq. [8] have been simpli-leaf dimension dc 5 0.01 m; field length xp 5 1000 m; spray releasefied to unity by assuming that both the reference height zr height hs 5 2 m; initial plume height sZ0 5 1 m; atmospheric

and the source height zs are small compared with sZ(x), which stability is neutral (D ); wind speed ur 5 2 m s21. Panels: (a )happens at sufficiently large x (in practice, quite quickly). depleting mass fraction s(x ); (b ) deposition 2ds/dx 5 2s9(x ) for

For line and plane sources, simpler expressions can be ob- a laterally uniform unit line source; (c ) drift deposition fractionfdrift(x ) for a plane source from Eq. [12].tained by using mass conservation as expressed by the first

equality in Eq. [7], which shows that for a unit source (a sourceof unit strength), the deposition is 2ds/dx. This applies both basic behavior of the model is shown in Fig. 1, by plotting theto the laterally integrated deposition from a unit point source, depleting mass fraction s(x) and the deposition 2ds/dx 5and to the deposition D(x) from a laterally uniform unit line 2s9(x) for a laterally uniform unit line source (both foundsource (an extensive line source across the wind direction). from numerical solution of Eq. [9]), and the drift depositionExtending to area sources, the deposition from a laterally fraction fdrift(x) for a plane source (from Eq. [12]). Resultsuniform unit plane source (extending over 2xp , x , 0 and are shown for several particle size distributions, in typical2∞ , y , ∞) is the integral of the unit line-source deposition, conditions specified in the figure legend and defined below.2ds/dx. Hence, if the actual plane source strength is D0 and The effect of particle size in limiting downwind drift is evident.the deposition is D(x), then the deposition from a unit plane In this work we do not explicitly consider the effects ofsource is: spray droplet evaporation. However, present indications are

that droplet evaporation is not a major factor for the applica-tion described in Paper I, as discussed there.

D(x)D0

5 2 #0

2xp

s9(x 2 x1) dx1 5 s(x) 2 s(x 1 xp) [12]

Particle Depositionwhere the prime denotes differentiation, so that s9(x) 5 ds/dx. In a spraying operation, D0 is the mass of spray released The deposition of particles to a vegetated surface occursper unit area or the intended dose, and the ratio D(x)/D0 is by three processes acting in parallel: direct gravitational set-the drift deposition fraction fdrift(x) used in Paper I to quantify tling to the surface, inertial impaction of particles on individualspray drift. elements (leaves and stems), and Brownian diffusion of parti-

In practice, we are usually concerned with the dispersion cles through the boundary layers of individual elements. Theof a cloud containing a distribution P(d) of particle sizes where subject has been reviewed comprehensively by Davidson andd is the particle diameter. Particle size influences s(x) and Wu (1990). (Some authors, including Davidson and Wu, distin-thence the deposition D through the deposition velocity Wd, guish a fourth process, interception, which we treat as a formwhich is a strong function of d (discussed in the next subsec- of impaction). All of these processes are strong functions oftion). Hence, the deposition is calculated by solving separately particle diameter (Fuchs, 1964; Chamberlain, 1967): gravita-for each particle size and summing the results, weighted by tional settling dominates for large particles (.100 mm), im-P(d). paction dominates in a middle size range centred around 10

mm, and Brownian diffusion dominates for very small particlesAn experimental test of this model is presented later. The


(,,1 mm). Consequently, a plot of Wd against particle diame- temperature, ra the air density, and lpath the mean free pathof air molecules (about 2 3 1027 m at sea level).ter exhibits a complex structure with a minimum at around

For the particle impaction conductance Gimp, the hypothe-1 mm where none of the three processes is effective. Impactionsis is:and Brownian diffusion are also strong functions of the length

scale of the canopy elements (such as a leaf width or stemGimp 5 afEimp GaM (form) [15]diameter) and the local wind speed about the canopy elements,

so Wd also depends on wind speed and canopy architecture. where Eimp is the particle impaction efficiency (0 , Eimp , 1)The first models used to describe this complex set of processes and af is another factor of order 1 accounting for differentialwere single-layer models in which the vegetated surface was sheltering effects. The motivations for Eq. [15] are that a directtreated as a composite, bulk entity (Owen and Thomson, 1963; analogy between the conductances for particle impaction andChamberlain, 1967; Sehmel, 1980). Later work endeavored to form drag is only likely to be valid for particles with impactionresolve some of the complexities associated with the structure efficiencies close to 1, and that the role of particle diameterof the vegetation canopy through multilayer models that re- in impaction can be quantified directly by the impaction effi-solve the vertical variation of the wind field, particle concen- ciency. This can be specified as a function of the Stokes numbertration, and deposition processes inside the canopy (Bache, St:1979a,b; Slinn, 1982; Davidson et al., 1982; Ferrandino andAylor, 1985; Raupach, 1993). While many of these multilayer Eimp 5 1 St

St 1 p2q

; St 52tstokeUc

dc

; tstoke 5rpd2

18ranamodels provide good representations of data from specificexperiments, they are too demanding on data and parameter-

[16]izations to be useful in the present context. Therefore, we usea single-layer model here. where rp is the particle density, dc is the dimension of the

The model is based on antecedents for single-layer models canopy elements upon which impaction occurs, and Uc is theof gas transfer to vegetation (Chamberlain, 1966; 1967; Shref- flow velocity about the canopy elements (which we character-fler, 1978; Hicks et al., 1985), but makes use of more recent ize by the friction velocity u*). The stokes number St is theunderstanding of leaf-scale processes as reviewed by Davidson ratio of the Stokes relaxation time tstoke to the radial traversaland Wu (1990). The deposition velocity is treated as a bulk time dc/(2Uc ). Equation [16] uses a commonly assumed empiri-(single-layer) conductance made up of three component bulk cal form for the function Eimp(St), for which Bache (1981) andconductances acting in parallel: Peters and Eiden (1992) proposed p 5 0.8 and q 5 2 for

several element shapes on the basis of fits to data.Wd 5 Wt 1 Gimp 1 Gbrow [13] Combining Eq. [13] to [15], the final form of the single-

layer model for Wd is:where Wt (the terminal velocity) accounts for gravitationalsettling, Gimp for impaction, and Gbrow for Brownian diffusion.

Wd 5 Wt 1 GaM[ fformafEimp 1 (1 2 fform) av Sc22/3] [17]Of these three, Wt is a well-known function of particle diame-ter, particle-to-air density ratio, and air viscosity (Malcolm in which af and av are parameters to be determined empirically.and Raupach, 1991). The bulk impaction conductance Gimp Although this model is a great simplification of the real,and Brownian conductance Gbrow are both calculated by ap- multilayer physics, it has advantages for the present purpose:pealing to the analogy between particle transfer to the surface it includes enough physics to capture the dependence of theand momentum transfer. three major processes (settling, impaction, and Brownian dif-

The bulk aerodynamic conductance for momentum (GaM 5 fusion) on particle diameter and wind speed; its two empiricalcoefficients are sufficient to permit matching to experimentalu2

*/ur, where u* is the friction velocity and ur the mean windreality but not enough to introduce parameterization prob-speed at the reference level zr ) is a known property of the

surface that can be specified in terms of the aerodynamicroughness length and thence the architecture of the canopy(Raupach, 1992, 1994). Two processes contribute to GaM: formor pressure drag, and viscous or skin-friction drag (Thom,1971), so we may write GaM 5 GaM(form) 1 GaM(visc). For atypical canopy, Thom (1971) estimated the ratio of the contri-butions of form and viscous drag to the total drag to be about3:1. If fform is the fraction of the total canopy drag exerted asform drag, this implies that GaM(form) 5 fformGaM andGaM(visc) 5 (1 2 fform )GaM, with fform 5 0.75.

Our hypothesis is that the Brownian conductance Gbrow isproportional to GaM(visc), and the impaction conductance Gimp

is proportional to GaM(form). For Gbrow, the relationship is:

Gbrow 5 avSc22/3GaM (visc) [14]

where Sc is the particle Schmidt number (Sc 5 na/kp, wherena is the kinematic viscosity of air and kp the Brownian diffusiv- Fig. 2. Test of the single-layer model for the deposition velocity Wd,

Eq. [15], against wind tunnel measurements of particle depositionity for particles in air), and av is a factor of order 1 accountingto a sticky grass surface by Chamberlain (1967). Predictions arefor different effects of interelement sheltering on the molecu-of Wd as a function of particle diameter d for vegetation of heightlar transfer of particles and momentum. The factor Sc22/3 ac-0.06 m and leaf area index 1, using the model of Raupach (1992,counts for the different molecular diffusivities of particles and1994) to calculate roughness length and other canopy aerodynamic

momentum (Monteith, 1973). The Brownian diffusivity is properties. Data and predictions are for three wind speeds givinggiven as a function of particle diameter d by the Stokes– friction velocities (u*) of 0.35, 0.70, and 1.40 m s21. The solid lineEinstein formula (Fuchs, 1964): kp 5 (1 1 2.5lpath/d)[kT/ is the predicted terminal velocity Wt as a function of d (Malcolm

and Raupach, 1991).(3pranad)], where k is the Boltzmann constant, T the absolute


lems; and it makes full use of information about bulk momen- air concentration Csa when a vessel containing water, air, and

tum transfer to characterize the aerodynamic properties of the species s is allowed to come to equilibrium. Data reviewedthe canopy. in Paper I show that As is of order 103 for a-endosulfan, 104

Figure 2 shows a test of Eq. [17] against the classical wind for b-endosulfan, and 105 for endosulfan sulfate, decreasingtunnel data set of Chamberlain (1967), for deposition of parti- roughly by a factor of two for each 68C increase in temperature.cles of various sizes to sticky (artificial) short grass of height Equation [18] shows that vapor deposition to water is driven0.06 m. The coefficients af and av are treated as adjustable by the difference between Cw and Ca (weighted by As ) and isparameters and set at af 5 2 and av 5 8. The main observed a bidirectional process: endosulfan moves from air to waterfeatures of the variation of Wd with d are reproduced quantita- when Ca is high and revolatilizes back to air when Ca is low.tively by the model, including the minimum in Wd(d) around More precisely, if the water concentration Cs

w in a water bodyd 5 1 mm (where none of Wt, Gimp, or Gbrow is effective), and of depth H responds to the air concentration Cs

a only by air–the convergence of Wd to the settling velocity Wt for large water exchange, then Cs

w obeys the mass balance equationparticles. The model also satisfactorily reproduces the trend ]Cs

w/]t 5 F sw(vapor)/H (see Paper I). This can be written as

of the data with wind speed. Given the simplicity of the model, ]Csw/]t 5 (AsCs

a 2 Csw )/T s, where T s is the time scale AsH/Vs

d.the agreement with measurements is satisfactory: it is better Hence, we have a first-order linear system in which Cs

w tendsthan some multilayer models (for instance Raupach [1993]) to track AsCs

a, with time averaging over a time of order T s.and is sufficient for the present purpose. For exposure to a steady air concentration Cs

a, the water con-Figure 3 shows the contributions of the three terms Wt, centration approaches the equilibrium value Cs

w 5 AsCsa, which

Gimp, and Gbrow to Wd, over the particle size range 0.01 to 1000 is at least 1000 times larger than the air concentration and ismm. The significant range for spray drift is broadly 10 to independent of water depth. This equilibrium is approached1000 mm. Around 10 mm, impaction is the dominant process in a time scale T s 5 AsH/Vs

d, which depends on water depth,contributing to Wd, and around 1000 mm, Wd is dominated by being about 1 d when H is 0.1 m and 10 d when H is 1 m (forparticle settling. Since Brownian diffusion is not a significant As 5 103 and Vs

d 5 1 mm s21 ).contributor to Wd in the range of interest for spray drift, the The vapor deposition velocity Vs

d determines the air–waterparameter av has no bearing and Eq. [17] is effectively a single- flux. According to the Deacon model (Deacon, 1977; Den-parameter model in the present context. mead and Freney, 1992), the inverse of Vs

d is a transfer resis-tance (Rs

d 5 1/Vsd ) given by the series sum of contributions

from three sequential parts of the water–air pathway: transferVapor Deposition to Water Surfacesthrough the water (Rw ), the quasilaminar air sublayer occu-

For vapor transport of pesticides from fields to water bodies, pying the lowest millimetre or so of the air (Rb ), and thethe main deposition process of concern is deposition of vapor turbulent atmosphere (Rt ). This contrasts with Eq. [13] forto the water body itself. For a species s, the deposition flux particle deposition velocity, which is a parallel sum of threeof vapor into a water body is given by (Denmead and Freney, conductances describing different, parallel processes op-1992): erating over the entire pathway from air to surface. The model

gives the vapor deposition velocity as:F s

w(vapour) 5 Vsd1Cs

a 2Cs

w

As2 [18]

Vsd 5

1Rt 1 Rb 1 Rw/Aswhere Vs

d is a deposition velocity describing the (two-way)exchange of vapor between water and air, and As is the water–

R21t 5 kVKu*/ln(zr/zb)air partition coefficient, defined as the partition of the species

s between water and air in a system at thermodynamic equilib- R21b 5 0.066(na/ka)20.61 u*

rium:R21

w 5 0.082(ra/rw)1/2 (nw/kw)22/3 u* [20]As 5 (Cs

w/Csa)eq [19]

where ra and rw are the densities of air and water, ka and kwThus, As is the ratio of the water concentration Cs

w to the are molecular diffusivities of endosulfan in air and water, na

and nw are the kinematic viscosities of air and water, u* is thefriction velocity, kVK (5 0.4) is the von Karman constant, zr

is the reference height for the air concentration, and zb is thethickness of the quasilaminar air sublayer (given by zb 5 50na/u*). The molecular diffusivities for endosulfan in air and waterwere estimated by multiplying the corresponding diffusivitiesfor CO2 by (Mc/Me )1/2 5 (44/407)1/2, where Mc and Me are themolecular weights for CO2 and endosulfan, respectively.

Figure 4 shows the behavior of the model by plotting Vsd,

Rt, Rb, and Rw/As against u*· Typical values for Vsd are about

2 mm s21 at u* 5 0.2 m s21, varying almost linearly with u*.All of the resistance terms in Eq. [20] are significant.

Source Terms for Spray Drift and Vapor Transport

Spray Drift

The source for spray drift is the intended deposition ordose D0 over the target area. Typically, the total dose ofFig. 3. Contributions of the three terms Wt (settling), Gimp (impaction),endosulfan in a single spray is D0 5 0.72 kg (endosulfan)and Gbrow (Brownian diffusion) to the deposition velocity Wd. Condi-

tions as in Figure 1 with u* 5 1.40 m s21. ha21 5 0.72 3 1024 kg m22, partitioned in the ratio 2:1 between


the crop into the air (kg m22 s21 ) at time t since spraying,and Ts

vol is a time constant for the volatilization of species s,which can be determined from measurements (Kennedy etal., 1998). The air concentration produced by this flux canbe evaluated using a simple Gaussian dispersion model (seeabove), yielding:

Csa(x, y, t) 5

AF svol(t)

psYsZ uexp12y2

2s2Y2 [22]

where x and y are the downwind and crosswind coordinatesof the receptor point relative to the source, and A is thearea of the source (assumed to be a quasi–point source withdimensions small compared with x and y).

Fig. 4. The deposition velocity of endosulfan vapor to water, from RESULTSthe Deacon model, Eq. [18]. Assumed parameters: ra 5 1.2 kg

Comparison of Vapor Transport Modelm23, rw 5 1000 kg m23, na 5 1.5 3 1025 m2 s21, nw 5 1 3 1026 m2

s21, ka 5 4.8 3 1026 m2 s21, kw 5 5.3 3 10210 m2 s21. with Field Data

Experimentsthe a and b isomers, so Da0 and Db

0 are respectively 0.48 31024 and 0.24 3 1024 kg m22. Ahmad et al. (1995) and Edge et al. (1998) describe

the field experiments used here for comparisons withVapor Transport the model. Two experimental sprays were carried out

at Auscott Warren, NSW, at 0400 on 21 Dec. 1994Of the total endosulfan dose D0 applied to the crop, afraction fvol is ultimately volatilized into the atmosphere (the (Spray 1) and 1600 on 7 Jan. 1995 (Spray 2). In eachrest ultimately degrading in situ or being transported away spray, 3 L ha21 of ULV Thiodan (240 g L21 endosulfan)from the crop by waterborne pathways). Kennedy et al. (1998) plus 0.5 L ha21 of a Bt formulation was applied aeriallyfound that fvol is about 0.7. To specify the vapor source, we to 182 ha of cotton in Fields 4 and 5 (see Fig. 5). Samplingassume that the flux of endosulfan vapor into the air decays of endosulfan in the environment was maintained forexponentially with time after spraying such that the total 5 d after each spray, as follows:amount of endosulfan volatilized is fvolD0 kg m22. This impliesthat: • Leaf samples were collected from Field 4 to deter-

mine the initial spray deposit and the subsequentF s

vol(t) 5fvolDs

0

T svol

exp(2t/T svol) [21] loss rate of endosulfan species (Ahmad et al., 1995).

• Air samplers were placed within Field 4 and 200 mto its south at station S200m (Fig. 5). These yieldedwhere Fs

vol is the volatilized endosulfan (species s) flux from

Fig. 5. Layout of the field experiment at Auscott Warren, 1994–1995.


air concentrations of endosulfan species averaged of endosulfan per spraying, D0 5 0.72 3 1024 kg m22,is partitioned 2:1 between the a and b isomers. Thereover 4-h intervals for the first 24 h and then over

12-h intervals to 5 d. Unfortunately, no air concen- is no sulfate in the applied dose. Although some oxida-tion to sulfate occurs on leaves in the time betweentration data were recovered from the S200m station

for Spray 2. spraying and volatilization, this was assumed to degradein situ because the vapor pressure of the sulfate is low• Water samplers consisted of water-filled galvanized

iron trays, placed along four transects extending compared with the a and b isomers (Paper I, AppendixA). The fraction volatilized was taken as f s

vol 5 0.5 forapproximately south, north, east, and west of Fields4 and 5, to distances up to 1000 m (Fig. 5). Each all species, to account for the sum of off-target spray

drift and other removal mechanisms. The main alterna-tray was 0.5 m2 in surface area and 0.1 m deep, andwas filled to depth 0.05 m with water. Tightly fitting tive removal mechanism is probably assimilation into

the soil, which can operate only on the fraction of thelids were placed over the trays during spraying toprevent contamination of the water by direct spray spray cloud that lands on the soil or is subsequently

washed into the soil by rain. Endosulfan on the soil stilldrift. The lids were removed 1 h after spraying.Water samples were removed from the trays every volatilizes, though at a slower rate than from leaves

(Hoechst Aktiengesellschaft, 1993).24 h, to 5 d, for analysis.The volatilization time scale in Eq. [21] was deter-In addition to the above experiments, two additional

mined from the leaf samples, which showed that theexperiments (Sprays 3 and 4) were carried out in theendosulfan present on leaves decays approximately ex-following season at Auscott Warren, on 18 and 27 Dec.ponentially as implied by Eq. [21], with time constants1995. In Spray 3, ULV Thiodan was applied aerially, inof 0.5 d (a) and 1.5 d (b) for Spray 1, and 1.0 d (a) andthe same way as for Sprays 1 and 2, to 120 ha of cotton5.0 d (b) for Spray 2. The longer decay time for the bin Field 7 (adjacent to Fields 4 and 5, which were fallow).isomer is caused by its lower vapor pressure (Paper I,In Spray 4, Thiodan EC (350 g L21 endosulfan) wasAppendix A). The difference in decay times was causedapplied aerially to Field 7 at 2.1 L ha21 (together withby warmer conditions for Spray 1 than for Spray 2,Pix at 0.45 L ha21, with a total application rate of 20 Land the strong dependence of the vapor pressure onha21 ). A major difference between Spray 4 and othertemperature. The average maximum temperature forsprays was in the droplet size distribution, which had athe 2 d after spraying was 408C for Spray 1 and 298Cmedian of around 240 mm for the EC formulation (Sprayfor Spray 2. These values for decay times are consistent4) and around 80 mm for the ULV formulation (otherwith wind tunnel measurements by Hoechst Aktienge-sprays). The movement of volatilized endosulfan wassellschaft (1993), who found values of less than 1 anddetermined after Sprays 3 and 4 using water-filled traysabout 3 d for the volatilization time scale of technicalas described above. No air concentrations were mea-endosulfan from leaves and soil, respectively. Singh etsured. Site configuration prevented samples being takenal. (1991) found a much larger value, about 10 d, bymore than 400 m from the edge of Field 7. The samplingexposing the liquid spray mixture in a petri dish on theperiods were reduced from 5 d to 3 d (Spray 3) and 2 droof of a laboratory, but this value is not applicable(Spray 4), with sampling only every 24 h, because thebecause of the unrealistic exposure technique.results from Sprays 1 and 2 had shown that peak concen-

• Meteorological parameters: Meteorological datatrations were achieved after 2 d. Because of these moreincluding wind speed, wind direction, and air tempera-restricted data sets we have not modeled Sprays 3 andture were recorded at height 2 m during the months4 in detail, but we use them for qualitative comparisons.surrounding the trials by Dr. Nicholas Woods and col-leagues, CPAS, Gatton College, University of Queens-

Model Specifications land. We used these data in half-hourly averaged form.Two gaps in the data (0830 to 1400 on 5 Jan. 1995, andThe vapor transport model was run for Sprays 1 and0700 on 8 January to 2230 on 9 January) were filled2, using the parameters summarized in Table 2. Theseby copying data from periods of similar meteorologicalwere assigned as follows:

• Volatilization parameters: The total applied dose conditions, as judged from weather maps and data from

Table 2. Parameters used for comparisons of the vapor transport model with field data.

Quantity Symbol Units Value

Endosulfan dose D0 kg m22 0.72 3 1024, partitioned (2:1:0) between (a, b, g)Fraction volatilized vol 0.5Volatization time scales Ta

vol d 0.5 (Spray 1); 1.0 (Spray 2)Tb

vol 1.5 (Spray 1); 5.0 (Spray 2)Elementary source area A ha 22.75 5 182/8Wind speed u m s21 local measurementsWind direction local measurementsStability classes for 8 three-hour periods E,E,B,A,A,A,E,E (u . 2 m s21 )

F,F,A,Z,Z,Z,F,F (u , 2 m s21 )Friction velocity u* m s21 0.01uReference height zr m 2.0Water depth in trays H m 0.05Endosulfan chemistry see Paper I


nearby meteorological stations. Fortunately, the gapsoccurred during periods of reasonably predictable windand temperature. The missing 5.5 h on 5 January werefilled using the same period on 11 January, and themissing 38.5 h on 8 and 9 January were filled with thesame period on 3 and 4 January.

In the absence of direct information, stability classeswere assigned according to time of day. Each 24-h pe-riod (0000–2400) was divided into eight 3-h blocks(0000–0300,..., 2100–2400). The stability classes for theseeight blocks were prescribed by the sequence (E,E,B,A,A,A,E,E) on occasions of moderate wind (u .umin 5 2 m s21 ), and by the sequence (F,F,A,Z,Z,Z,F,F)for occasions of light wind (u , umin ). Sensitivity testshave confirmed that the model results are insensitiveto the exact choice of stability classifications.

• Deposition and chemical parameters: The waterdepth in the trays was H 5 0.05 m. The depositionvelocity for vapor of species s to water, Vs

d, was deter-mined by Eq. [20]. The chemical properties of endosul-fan (Paper I, Appendix A) were evaluated by assumingthat the temperature of the water in the trays was equalto the measured air temperature at the meteorologicalreference height. The friction velocity u* was estimatedfrom the mean wind speed (u) at the reference height(zr 5 2 m), since the ratio u*/u is reasonably independentof wind speed. Over a rough surface such as an agricul-tural area, this ratio is about 0.1. However, the waterin each experimental tray was nearly aerodynamicallysmooth and was significantly sheltered by additionalmechanisms, mainly a tray side wall extending 50 mmabove the water surface. To account for this we assumea shelter factor (the ratio of u* within the tray to theambient u*) of 0.1. Thus, for the water in the trays, u* 50.01u. No stability corrections were applied. Sensitivitytests have shown that the model results are insensitiveto assumptions about u*. The reason is that u* affects

Fig. 6. Predicted and measured endosulfan air concentrations (a, b,only the deposition velocity Vsd, and therefore the rate

g, total) at stations within Field 4 and at S200m outside Field 4.at which the air–water equilibration proceeds (that is,the rate at which the flux in Eq. [18] approaches zero

endosulfan concentrations within Field 4 for Spray 1when Csa is steady). It does not affect the final equilib-

are overpredicted by the model for the first 24 h afterrium value of Csw, which is AsCs

a. The rate of equilibrationspraying, but tend to be underpredicted at later times.is given by the time constant AsH/Vs

d, which is fairlyThe comparisons between model and measurements areshort (around 1.5 h for H 5 0.05 m, As 5 1000, Vs

d 5quite good at S200m for Spray 1, and also for Spray 20.001 m s21 ). Therefore, over time scales of a day or so,(for which only within-field data are available). It isthe air and water concentrations are close to equilibriumconcluded that, despite major simplifications, the modeland the rate of the exchange (influenced by u*) isis capable of reproducing the order of magnitude ofnearly immaterial.measured air concentrations. This is sufficient for thepurpose at hand.Results of Comparisons Turning to water concentrations, the model was used

We first consider air concentrations. Measurements to predict the water concentrations (Caw, Cb

w, Cgw and the

total Cw 5 Caw 1 Cb

w 1 Cgw) at each receptor point (tray),of air concentrations were made at stations within Field

4 and at the S200m station, 200 m to its south (Fig. 5). for the 5 d following each spray. A representative sam-ple of these predictions (the westward trays) are shownThese are compared with predictions from the model

in Fig. 6. There is a strong diurnal cycle in both the in Fig. 7 for Spray 1 (upper frames) and Spray 2 (lowerframes), together with the measured concentrations atpredicted and measured air concentrations, caused by

the entrapment of air with high concentrations near the these trays. The predicted water concentrations rise rap-idly from zero at the time of spraying to typical maximaground at night when atmospheric dispersion is much

weaker than by day. The measured concentrations are around 0.5 g L21 for the trays close to the crop and 0.1g L21 for more distant trays. There is a diurnal oscillationsimilar within Field 4 and at S200m. Total measured


Fig. 7. Predicted and measured endosulfan water concentrations (a, b, g, total) at four westward receptor locations for Sprays 1 and 2.

in the predictions (though less strong than in the pre- centrations (over all 5 d of observation) for each endo-sulfan species. Again, each point represents one traydicted air concentrations), with highest concentrations

occurring at night, because the water concentrations for one spray. For Spray 1 the b-endosulfan concentra-tions are overpredicted by the model while the otherfollow the air concentrations with a temporal smoothing.

To simplify quantitative comparisons, Fig. 8 shows two species are underpredicted, leading to a reasonableprediction for the total endosulfan concentration (Fig.the measured and modeled total concentrations Cw at

each tray, averaged in time over all five observationdays. Each point represents one tray for one spray, andis an average over five values (the measurements at24-h intervals and the predicted values at these times).For both Sprays 1 and 2 the model and measurementsagree to within about 30% on average, a good resultgiven the uncertainties involved in the modeling andthe inevitable scatter in the measured data. Figure 9compares measured and modeled time-averaged con-

Fig. 9. Time-averaged comparison (over 5 d) of measured and mod-Fig. 8. Time-averaged comparison (over 5 d) of measured and mod-eled total endosulfan concentrations at 13 receptor points, for eled endosulfan species concentrations (a, b, g, total) at 13 receptor

points, for (a ) Spray 1; (b ) Spray 2.Sprays 1 and 2.


out for 21 particle sizes determined by the particle sizedistribution P(d). The final fdrift for a spray with a broadparticle size spectrum was calculated by integratingfdrift(x,d)P(d) over d, where fdrift(x,d) is the value appro-priate at distance x downwind of the field for a singleparticle size d.

For the ‘‘standard case’’ of Bird et al. (1996), 36 exper-imental replicates are available for a single-swathe aerialboom spray with a median droplet diameter of close to250 mm and an approximately log-normal particle sizedistribution P(d). We inferred mean measured valuesof fdrift(x) from plots at three different wind speeds corre-sponding to the 20th, 50th, and 80th percentiles of thewind speed range covered by their 36 replicates. Figure10 compares these data with predictions for each windFig. 10. Comparison of particle dispersion (spray drift) model withspeed, using model parameters (given in the figure cap-data from Bird et al. (1996; standard case) for three wind speeds

u. Parameters: canopy height hc 5 0.15 m; leaf dimension dc 5 tion) selected to match the experimental conditions. The0.005 m; field length xp 5 274 m; spray release height hs 5 2.5 m; predictions capture the magnitude and the streamwiseinitial plume height sZ0 5 1 m; median particle diameter 5 250 trend of fdrift(x) well, slightly overestimating the magni-mm; particle size distribution is log-normal with standard deviation

tude of fdrift(x) at higher wind speeds. Given the majorln(1.7); atmospheric stability is neutral (D ) or slightly unstable (C ).simplifications of the model, this is a satisfactory resultthat provides confidence that the values of fdrift being8). For Spray 2, the prediction of both isomer concentra-estimated by the model are reasonable.tions is reasonable but the sulfate concentration is un-

derpredicted.CONCLUSIONSEndosulfan sulfate was not present in the applied

spray or in the measured air concentrations. Also, a and We have developed and tested models for two signifi-b endosulfan are oxidized to sulfate only very slowly, if cant airborne pathways transporting endosulfan andat all, in clean water or in water with added clean dust other agricultural chemicals from farms to environmen-(Edge et al., 1998). For this reason, we used a very long tal receptors, especially water bodies. For both path-time constant (100 d) to characterize oxidation (Paper ways, the models consist of three components respec-I). The predictions accordingly show very little sulfate tively describing source strengths, dispersion, and(Fig. 9). The presence of sulfate in some of the observa- deposition. Economy of description has been gained bytions was probably due to the deposition of contami- considering these pathways together, as much of thenated dust by vehicular movement or by wind. underlying process physics (especially for dispersion) is

The water concentration data from Sprays 3 and 4 common to both. Throughout, we have used the simplest(1995–1996 season) were generally consistent with those possible process descriptions consistent with often com-from Sprays 1 and 2, though somewhat lower on aver- plex physical realities; for example, a Gaussian-plumeage. Temperatures following spraying were also lower. model is used to describe both spray and vapor dis-The most significant aspect of these data was that there persion.was no obvious difference between the EC and ULV Spray drift depends not only on dispersion but alsoformulations. on deposition to the underlying (usually vegetated) sur-

Given the difficulties involved in both measurements face. A single-layer model for this deposition processand modeling, the comparisons in Fig. 6 to 9 are gener- has been formulated that is well supported by the labo-ally encouraging. One reason for the disagreements is ratory data of Chamberlain (1967) and that shows thatlikely to be uncertainties in the chemical coefficients. for spray droplets of around 100 mm or smaller, im-The water–air partition coefficients As are the most impaction and related processes dominate over gravita-portant parameters determining the water concentra- tional settling in particle deposition. By combining thistions of the a and b isomers, while the rate constants model with a simple dispersion model based on a settlingfor oxidation to sulfate determine the sulfate concentra- Gaussian plume, a model for spray drift is obtained thattions at intermediate times of a few days. Both the values agrees very well with field data from Bird et al. (1996).and the assumed temperature dependencies of these For vapor transport, the model consists of a simplequantities are still uncertain. exponential-decay assumption for the post-spray volatil-

ization of endosulfan from a sprayed crop, the sameComparison of Spray Drift Model with Field Data dispersion model used for spray drift, and a model fordeposition of vapor to a water surface due to DeaconThe spray drift model for the drift deposition fraction

fdrift has been compared with field data from Bird et al. (1977) and Denmead and Freney (1992), based on akinetic formulation for the air–water exchange and a(1996). The model consists of Eq. [12] for fdrift, with

the airborne particle fraction s(x) determined by the deposition velocity. The dispersion model and field mea-surements (Edge et al., 1998) both show that volatiliza-numerical (Runge–Kutta) solution of Eq. [9], using Eq.

[17] to specify the deposition velocity Wd as a function tion and dispersion of endosulfan from sprayed cropsproduces air concentrations (Ca ) in the range 1 to 5of particle size d. The calculation of fdrift was carried


waters. Australian and New Zealand Environment and Conserva-mg m23 at sites within and adjacent to sprayed fields,tion Council, Environment Priorities and Coordination Group, En-immediately after spraying. Values are much higher atvironment Australia, Canberra, Australia.

night than by day, because of entrapment of air with Bache, D.H. 1979a. Particle transport within plant canopies—I. Ahigh concentrations near the ground at night. As volatil- framework for analysis. Atmos. Environ. 13:1257–1262.

Bache, D.H. 1979b. Particulate transport within plant canopies—II.ization proceeds, air concentrations decay with a timePrediction of deposition velocities. Atmos. Environ. 13:1681–1687.scale of 1 to 5 d, decreasing with increasing temperature.

Bache, D.H. 1981. Analysing particulate deposition to plant canopies.The complete model of the vapor transport pathway Atmos. Environ. 15:1759–1761.(including volatilization, dispersion, deposition, and wa- Bird, S.L., D.M. Esterly, and S.G. Perry. 1996. Off-target deposition

of pesticides from agricultural aerial spray applications. J. Environ.ter chemistry) produces predicted endosulfan waterQual. 25:1095–1104.concentrations (Cw ) of order 0.1 to 0.5 mg L21 in a water

Briggs, P.R., M.R. Raupach, B. Cooper, and M. Muschal. 1998. Integ-body of depth 0.05 m within 1 km of a sprayed cotton rative modelling of transport and fate of endosulfan in the riverinecrop, over the first few days after spraying. These predic- environment. Part I: Use of observed riverine concentrations totions are in broad (though far from perfect) agreement identify contributions of airborne and waterborne transport. Con-

sultancy Rep. 98-14. CSIRO Land & Water, Canberra, Australia.with field observations by Edge et al. (1998). For experi-Chamberlain, A.C. 1966. Transport of gases to and from grass andmental trays containing water of depth 0.05 m, they

grass-like surfaces. Proc. Royal Soc. London A290:236–265.found water concentrations, averaged over the 5 d fol- Chamberlain, A.C. 1967. Transport of Lycopodium spores and otherlowing spraying, ranging from 0.3 mg L21 (at a distance small particles to rough surfaces. Proc. Roy. Soc. London A296:

45–70.of 200 m from the crop) to 0.18 mg L21 (at 1 km) forCooper, B. 1996. Central and north west regions water quality pro-Spray 1; the corresponding range for Spray 2 was 0.18

gram: 1995/96 report on pesticides monitoring. TS96.048. NSWmg L21 (at 200 m) to 0.1 mg L21 (at 1 km). The highest Dep. of Land and Water Conserv., Parramatta, Australia.observed concentrations in individual trays were around Csanady, G.T. 1973. Turbulent diffusion in the environment. D. Reidel0.5 mg L21, recorded around 48 h after spraying. Moni- Publ. Co., Dordrecht, the Netherlands.

Davidson, C.I., J.M. Miller, and M.A. Pleskow. 1982. Influence oftoring of two additional sprays (Sprays 3 and 4) in thesurface structure on predicted particle dry deposition to naturalfollowing season gave similar (though somewhat lower)grass canopies. Water Air Soil Pollut. 18:25–43.concentrations, but temperatures were also lower. No Davidson, C.I., and Y.L. Wu. 1990. Dry deposition of particles and

differences were observed between ULV and EC for- vapors. p. 103–216. In S.E. Lindberg et al. (ed.) Acidic precipitation,Vol. 3: Sources, deposition, and canopy interactions. Springer–mulations, for which median spray droplet diametersVerlag, New York.are around 80 mm and 240 mm, respectively.

Deacon, E.L. 1977. Gas transfer to and across an air–water interface.This level of agreement is sufficient for the purpose ofTellus 29:363–374.

establishing the overall significance of vapor transport Denmead, O.T., and J.R. Freney. 1992. Transfer coefficients for air–relative to other possible transport pathways. However, water exchange of ammonia, carbon dioxide and methane. Ecol.

Bull. 42:31–41.modeling uncertainty remains in the physical and chemi-Edge, V.E., N. Ahmad, and P. Rohas. 1998. Aerial transport of endo-cal properties of endosulfan species. It should also be

sulfan: Vapour and dust movement. p. 23–27. In Minimising thenoted that the water concentrations observed in the impact of pesticides on the riverine environment: Key findingsexperimental trays cannot be directly scaled to infer the from research with the cotton industry: 1998 Conference, Canberra.concentrations arising from vapor transport in a river, 21–22 July 1998. LWRRDC Occasional Paper 23/98. Land and

Water Resour. Res. and Development Corp., Canberra, Australia.which has a quite different (generally much greater)Ferrandino, F.J., and D.E. Aylor. 1985. An explicit equation for depo-depth and for which the exposure pattern is quite differ-

sition velocity. Boundary-Layer Meteorol. 31:197–201.ent. To make this conversion, one must consider the Fuchs, N.A. 1964. The mechanics of aerosols. Pergamon Press, Ox-dynamical factors built into the model described in ford, UK.Part I. Golder, D. 1972. Relations among stability parameters in the surface

layer. Boundary-Layer Meteorol. 3:47–58.Hanna, S.R., G.A. Briggs, and R.P. Hosker. 1982. Handbook on atmo-ACKNOWLEDGMENTS

spheric diffusion. Tech. Info. Center, U.S. Dep. of Energy, OakRidge, TN.We are grateful for many discussions with Nick Schofield,

Hicks, B.B., D.D. Baldocchi, R.P. Hosker Jr., B.A. Hutchison, D.R.Dave Anthony, Phillip Ford, John Leys, Frank Larney, IvanMatt, R.T. McMillen, and L.C. Satterfield. 1985. On the use ofKennedy, Brian Hearn, Monika Muschal, and Bruce Cooper.monitored air concentrations to infer dry deposition. NOAA Tech.We likewise thank Nicholas Woods for both discussions andMemo. ERL ARL-141. NOAA, Silver Spring, MD.for access to the meteorological data recorded during the field Hoechst Aktiengesellschaft. 1993. HOE 002671. Endosulfan—Vol-

experiments. We acknowledge with appreciation the support atility and photochemical oxidative degradation in the atmosphere.of the Land and Water Resources Research and Development Document A50940. Hoechst Aktiengesellschaft, Frankfurt,Corporation (LWRRDC), the Cotton Research and Develop- Germany.ment Corporation (CRDC), and the Murray–Darling Basin Kennedy, I.R., F. Sanchez-Bayo, S.W.L. Kimber, H. Beasley, and N.

Ahmad. 1998. Movement and fate of endosulfan on-farm (NewCommission (MDBC), through the joint research programSouth Wales). p. 33–37. In Minimising the impact of pesticides on‘‘Minimising the Impact of Pesticides on the Riverine Envi-the riverine environment: Key findings from research with theronment’’.cotton industry: 1998 Conference, Canberra. 21–22 July 1998.LWRRDC Occasional Paper 23/98. Land and Water Resour. Res.REFERENCES and Development Corp., Canberra, Australia.

Malcolm, L.P., and M.R. Raupach. 1991. Measurements in an airAhmad, N., V.E. Edge, and P. Rohas. 1995. Aerial transport of endo-settling tube of the terminal velocity distribution of soil material.sulfan. In Proc. Workshop on Pesticides and the Riverine Environ.,J. Geophys. Res. 96:15275–15286.Sydney. Aug. 1995. Land and Water Resources Research and De-

Monteith, J.L. 1973. Principles of environmental physics. Arnold,velopment Corporation, Canberra, Australia.London.Australian and New Zealand Environment and Conservation Council.

1992. Australian water quality guidelines for fresh and marine Muschal, M. 1997. Central and north west regions water quality pro-


gram: 1996/97 report on pesticides monitoring. CNR97.063. NSW port and fate of endosulfan in the riverine environment, Part II:Summary report. Consultancy Rep. 98-50. CSIRO Land & Water,Dep. of Land and Water Conserv., Parramatta, Australia.

Muschal, M. 1998. Central and north west regions water quality pro- Canberra, Australia.Raupach, M.R., P.R. Briggs, P.W. Ford, M. Muschal, B. Cooper, andgram: 1997/98 report on pesticides monitoring. CNR98.038. NSW

Dep. of Land and Water Conserv., Parramatta, Australia. V.E. Edge. 2001. Endosulfan transport: I. Integrative assessmentof airborne and waterborne pathways. J. Environ. Qual. 30:714–728Owen, P.R., and W.R. Thomson. 1963. Heat transfer across rough

surfaces. J. Fluid Mech. 15:321–334. (this issue).Raupach, M.R., P.W. Ford, and P.R. Briggs. 1996. Modelling thePeters, K., and R. Eiden. 1992. Modelling the dry deposition velocity

of aerosol particles to a spruce forest. Atmos. Environ. 26A: aerial transport of endosulfan to rivers. Part I: The vapour transportpathway. Tech. Rep. 113. CSIRO Centre for Environ. Mechanics,2555–2564.

Raupach, M.R. 1992. Drag and drag partition on rough surfaces. Canberra, Australia.Sehmel, G.A. 1980. Particle and gas dry deposition: A review. Atmos.Boundary-Layer Meteorol. 60:375–395 [corrigendum: 76:303–304].

Raupach, M.R. 1993. Dry deposition of gases and particles to vegeta- Environ. 14:983–1011.Shreffler, J.H. 1978. Factors affecting dry deposition of SO2 on foreststion. Clean Air 27:200–203.

Raupach, M.R. 1994. Simplified expressions for vegetation roughness and grasslands. Atmos. Environ. 12:1497–1503.Singh, N.C., T.P. Dasgupta, E.V. Roberts, and A. Mansingh. 1991.length and zero-plane displacement as functions of canopy height

and area index. Boundary-Layer Meteorol. 71:211–216 [corrigen- Dynamics of pesticides in tropical conditions. 1. Kinetic studies ofvolatilisation, hydrolysis and photolysis of dieldrin and endosulfan.dum: 76:303–304].

Raupach, M.R., and P.R. Briggs. 1996. Modelling the aerial transport J. Agric. Food Chem. 39:575–579.Slinn, W.G.N. 1982. Predictions for particle deposition to vegetativeof endosulfan to rivers. Part II: Transport by multiple pathways.

Tech. Rep. 121. CSIRO Centre for Environ. Mechanics, Can- canopies. Atmos. Environ. 16:1785–1794.Thom, A.S. 1971. Momentum absorption by vegetation. Q. J. R. Mete-berra, Australia.

Raupach, M.R., and P.R. Briggs. 1998. Integrative modelling of trans- orol. Soc. 97:414–428.

Profenofos Residues in Wild Fish from Cotton-Growing Areasof New South Wales, Australia

A. Kumar and John C. Chapman*

ABSTRACT not enough toxicity and fate data available to permitadequate risk assessment of sites adjacent to the spray-The organophosphorus (OP) pesticide profenofos (O-4-bromo-2-ing. Profenofos is reported to be highly toxic to somechlorophenyl O-ethyl S-propyl phosphorothioate) is used heavily inaquatic organisms (Shaw, 1995), but only limited ecotox-cotton-growing areas of eastern Australia toward the end of the grow-

ing season. European carp (Cyprinus carpio ), bony bream (Nemata- icology data are available. Batley and Peterson (1992)losa erebi ), and mosquitofish (Gambusia holbrooki ) were collected ranked its risk as about mid-range on the list of 12from the cotton-growing areas around Wee Waa, New South Wales, priority cotton pesticides.to determine the relationship between profenofos residues and acetyl- It is often assumed that the low bioconcentration po-cholinesterase (AChE) activity in wild fish. Profenofos concentrations tential of OPs by aquatic organisms rapidly reduces thein water, sediment, and fish tissue reflected its general level of use; risk of profenofos exposure in ecosystems (Eto, 1974;levels in March 1994 were significantly higher than in 1993 and gener-

Yu and Sandborn, 1975) after application. Hence, evalu-ally decreased in May, 6 wk after cessation of spraying. Residues ination of the rate of recovery of the ecosystem wouldcarp and bony bream generally correlated with concentrations in waterpermit adequate assessment of any long-term conse-and sediment, although residues in fish tend to persist longer at somequences of profenofos spraying.sites. Acetylcholinesterase inhibition was a useful indicator of profen-

ofos exposure within a season, particularly if linked with residue Measurement of the concentrations of pesticide inmeasurements. Bony bream and gravid female mosquitofish recov- spot samples of water may underestimate the exposureered AChE levels more slowly than carp or nongravid mosquitofish. levels encountered by inhabiting aquatic organisms.Recovery in creeks was generally more rapid than in lagoons. However, an analysis of accumulated profenofos resi-

dues in the sediment or fish tissue may give a betterindication of exposure (Nowak and Julli, 1991) and toxi-

Profenofos is a broad-spectrum organophosphorus cological risk. In general, the most significant route of(OP) insecticide registered in Australia for control exposure for fish appears to be the direct uptake of

of agricultural pests in cotton-growing areas, usually late insecticide from water (Tsuda et al., 1994). This studyin the growing season. It is usually applied by air at attempts to establish whether or not wild fish accumu-around 2 kg a.i./ha (Shaw, 1995), but there are currently late sufficient residues of profenofos for it to represent

a risk to the aquatic system.Acetylcholinesterase (AChE) activity in fish has beenA. Kumar, Dep. of Biological Sciences, Macquarie Univ., North Ryde

used as an indicator to monitor poisoning by OP pesti-NSW 2113 Australia. John C. Chapman, Environment ProtectionAuthority, NSW, located at EPA and Univ. of Technology, Sydney, cides (Coppage and Braidech, 1976; Zinkl et al., 1991).Centre for Ecotoxicology, Westbourne Street, Gore Hill NSW 2065 The use of AChE activity measurements in fish couldAustralia. A. Kumar, present address: School of Pharmaceutical Sci- be developed as a useful indicator, provided some priorences, Univ. of South Australia, Adelaide, SA, 5000, Australia. Re-ceived 8 Oct. 1999. *Corresponding author ([email protected].

Abbreviations: AChE, acetylcholinesterase; ANOVA, analysis ofgov.au).variance; OP, organophosphorus; PCCM, Pearson Correlation Coeffi-cient Matrix.Published in J. Environ. Qual. 30:740–750 (2001).

endosulfan transport: ii. modeling airborne dispersal …dust.ess.uci.edu/ppr/ppr_rba01.pdf ·...

Documents