A simulation model (PostPLANT-Soil) for predicting pesticide concentrations in succeeding leafy vegetables: II. Validation with experimental data on plant uptake in a growth chamber

Abstract

We validated a simulation model (PostPLANT-Soil) for predicting pesticide concentrations in succeeding leafy vegetables reported in our first paper in this series, which includes the pesticide sorption process into plant roots. As a result of the model validation with the measured data from a plant uptake study in a growth chamber, the model successfully simulated the concentration changes of pesticides in a plant shoot. However, the simulated shoot concentrations for several pesticides were overestimated compared to the measured values. The leafy vegetable (Brassica rapa) used in this study probably has a high metabolic ability for the fungicide flutolanil from the result of the uptake study under a hydroponic condition.

Introduction

Recently, some pesticides used for primary crops have been detected in succeeding crops at levels exceeding the uniform residue level of 0.01 mg/kg^1,2) via the residual pesticides in topsoil of upland fields. In general, the pesticide concentrations in succeeding crops depend largely on pesticide behavior in soil caused by application methods (e.g., usage, timing, and frequency) and physicochemical properties of the pesticides. In addition, cultivation systems (e.g., cultivated species, season, and plant-back intervals) influence pesticide residue levels. Therefore, it is important to adequately estimate the phytoavailable pesticides in soil and especially to comprehend uptake process of the pesticides by plants for evaluating the risks of pesticides to succeeding crops.

Cucurbitaceous plants are known to uptake and translocate organic chemicals having the value of logarithmic n-octanol–water partition coefficient (log P_ow)>5 such as dieldrin³⁾ and DDE.^4,5) We previously performed uptake experiments with 10 pesticides and dieldrin using 16 plant species.⁶⁾ As a result of this study, the concentrations of dieldrin in cucurbits were 20–50 times higher than those in cruciferous vegetables. However, the concentrations of pesticides with log P_ow values of −0.6 to 4.6 in leafy vegetables were about 2–10 times higher than those in other plant species. Therefore, there was no marked difference among plant species as observed for dieldrin. We also found that the uptake abilities of pesticides with log P_ow values of 2.6–4.6 from plant roots of Brassica rapa L. var. perviridis were lowest in very young seedlings and then constant until the harvesting stage, but the translocation abilities of the pesticides from root to shoot were constant during all growth stages.⁷⁾ These findings are important for understanding the pesticide residue mechanism in succeeding crops.

In our first report on a simulation model developed for predicting pesticide concentrations in succeeding leafy vegetables (PostPLANT-Soil),⁸⁾ from the results of the model validation using the measured data from a field experiment in an upland Andosol, the pesticide concentrations in the B. rapa shoot were correlated with the concentrations in the soil solution rather than those of the water-extracted pesticides. The model successfully simulated the concentration changes in the plant shoot when the simulated concentrations of the pesticides in the soil solution are fitted to the measured values by considering the key parameter (the corrective coefficient of the soil adsorption coefficient). However, the simulated shoot concentrations at 28 days after sowing (an appropriate harvest weight) for several pesticides were much larger than the measured values. This result indicates that B. rapa may have a metabolic ability for the pesticides.

To evaluate a degree of decomposition by plant metabolism of each pesticide, it is necessary to validate the PostPLANT-Soil model using the results of detailed uptake experiments for many pesticides with different log P_ow values under controlled growth conditions. Although many findings of the plant metabolism of various pesticides focused on identification of their metabolites have been reported,^9,10) there are few reports on the metabolic rate (e.g., half-life period) of individual pesticides in specific crops (especially, leafy vegetables). In addition, we suggested that the metabolic rate of four pesticides might be affected by growth conditions (temperature, day length, and soil water content) of B. rapa.¹¹⁾ Therefore, it is necessary to separately obtain the metabolic rate of target pesticides under the same growth condition as in the uptake experiment for the model validation.

The aim of this study is to validate the simulation model (PostPLANT-Soil) for predicting pesticide concentrations in succeeding leafy vegetables, which includes pesticide sorption process into plant roots. We compared the values calculated by the model with measured values from a plant uptake study of B. rapa in a growth chamber under controlled growth conditions. Furthermore, to obtain the metabolic rate of two fungicides (procymidone and flutolanil) with a similar log P_ow used for the plant uptake study, we conducted an uptake experiment for the fungicides under a hydroponic condition and discuss plant metabolism of the fungicides in B. rapa.

Materials and methods

1. Chemicals

Ten pesticides (dinotefuran, imidacloprid, clothianidin, thiacloprid, fosthiazate, metalaxyl, fenobucarb, procymidone, flutolanil, and tolclofos-methyl) were used for the plant uptake experiment in a growth chamber and for model validation. Detailed information about physicochemical properties of the pesticides is provided in our first report.⁸⁾

2. Model description

In our first report,⁸⁾ we considered the following situation to develop a simulation model (PostPLANT-Soil) for predicting residual pesticide concentrations in succeeding crops (Fig. 1): (1) topsoil of upland fields is composed of liquid and solid phases, and the gas phase is neglected; (2) a compartment system was employed for soil solution, two soil particles (solid phase 1 at reversible and solid phase 2 at irreversible sorption sites), and plant shoot, which are assumed to be in a completely mixed condition; (3) residual pesticides used for previous crops are uniformly distributed in the plowed layer at sowing or planting of succeeding crops; (4) pesticides are distributed between soil solution and particle compartments of the plowed layer depending on soil adsorption coefficient; (5) pesticides in the soil solution compartment are taken up via plant root of the succeeding crops and transported to the plant shoot compartment depending on the transpiration stream concentration factor¹²⁾; and (6) exchange of pesticides between the plant shoot and air and photolysis on the plant surface are not considered.

Fig. 1. Schematic diagram of pesticide behavior in topsoil of upland fields and succeeding crops according to a PostPLANT-Soil model including pesticide sorption process into plant roots. Single-headed arrows indicate mass flow of pesticides. Solid phases 1 and 2 are soil particle compartment at reversible and at irreversible sorption sites, respectively.

In this study, the PostPLANT-Soil model was included to predict residual pesticide concentrations in the plant root (Fig. 1) because root concentrations of B. rapa were measured in the plant uptake study for the model validation. We added the following situation: pesticides are distributed reversibly between soil solution and plant root compartments with a completely mixed condition depending on the root concentration factor.¹²⁾

2.1. Description of pesticide dissipation in soil

In general, dissipation behavior of total-extractable pesticides in soil shows a first-order dissipation curve or a bi-phasic pattern. In this study, a double first-order in parallel (DFOP) model¹³⁾ as a descriptor of the dissipation pattern was used in fitting to the measured data as follows:

  

(1)

where

• C_{S_total}(t)=concentration of the total-extractable pesticide in soil at time t [µg/g on a dry weight (DW) basis],
• C_{S_total}(0)=initial concentration of the pesticide in soil (µg/g DW),
• g=fraction of C_{S_total}(0) distributed in the first compartment (dimensionless),
• k₁=first-order dissipation rate constant in the first compartment (day⁻¹), and
• k₂=first-order dissipation rate constant in the second compartment (day⁻¹).

The DFOP model parameters, g, k₁, and k₂ were calculated by the nonlinear least-squares method with Microsoft Excel Solver Add-In software.¹⁴⁾

2.2. Sorption process of pesticides into plant roots

The root concentration factor, RCF [mL/g on a fresh weight (FW) basis], is defined as the chemical concentration ratio between the plant root and soil solution in rhizosphere,⁹⁾ and can be expressed at an equilibrium condition as follows:

  

(2)

where

• C_Root=pesticide concentration in the plant root (µg/g FW), and
• C_Water=pesticide concentration in the soil solution adjacent to the plant roots (µg/mL).

2.3. Growth rate of plant roots

In our first report, growth of leafy vegetable shoots was expressed by a logistic function. We assumed that growth of the roots is approximated by a logistic function the same as for shoots, and growth rate can be given as follows:

  

(3)

where

• m_Root=mass of the plant root (g FW/stock),
• m_Root,max=maximum mass of the plant root (g FW/stock), and
• λ_Root=first-order growth rate constant of the plant root (day⁻¹).

With an initial condition of m_Root (t=0)=m_Root,0, Eq. (3) can be integrated as follows:

  

(4)

The parameters m_Root,0, m_Root,max, and λ_Root were obtained by fitting the calculated results according to Eq. (4) with the measured data using nonlinear least-squares method with Microsoft Excel Solver Add-In software.

2.4. PostPLANT-Soil model including sorption process into plant roots

In this model, a two-site soil sorption model¹⁵⁾ was employed, which is composed of an irreversible and reversible sorption site (i.e., DFOP model). The model was constructed on the assumptions: (1) applied pesticides are instantaneously distributed between the soil solution and soil particle at irreversible sorption site and then degradation process occurs individually in each compartment without sorption process; (2) soil particle compartment at reversible sorption site is a temporally pool where no degradation occurs; (3) adsorption and desorption processes between soil solution and plant root compartments are expressed by a kinetic concept. If bulk density and volumetric water content of the plowed soil layer are constant, then the chemical mass balance equations in the plowed soil layer and the plant shoot and root are as follows:

for the soil solution compartment,

  

(5)

for the soil particle compartment at reversible sorption site,

  

(6)

for the soil particle compartment at irreversible sorption site,

  

(7)

for the plant shoot compartment,

  

(8)

and for the plant root compartment,

  

(9)

with

  

where

• M_app=amount of pesticide applied (g a.i./m²),
• C_Soil,1=pesticide concentration in the soil particle compartment at reversible sorption site (µg/g DW),
• C_Soil,2=pesticide concentration in the soil particle compartment at irreversible sorption site (µg/g DW),
• C_Shoot=pesticide concentration in the plant shoot (µg/g FW),
• m_Shoot=mass of the plant shoot (g FW/stock),
• A=area of experimental field (m²),
• ρ_b=bulk density of the plowed soil layer (g/mL),
• θ=volumetric water content of the plowed soil layer (v/v),
• d_S=depth of the plowed soil layer (m),
• Q_T=volumetric flow rate of the transpiration stream in the xylem (mL/stock/day),
• ρ_plant=planting density of the succeeding crop (stock/m²),
• TSCF=transpiration stream concentration factor (dimensionless),
• f_soil,1=mass fraction of soil particle at reversible sorption site (=1−f_soil,2, dimensionless),
• f_soil,2=mass fraction of soil particle at irreversible sorption site (dimensionless),
• K_d=soil adsorption coefficient measured by a batch method (mL/g),
• f_adj=corrective coefficient of K_d for a field condition (dimensionless)
• k_ads=adsorption rate constant for the soil and plant root (day⁻¹),
• k_des=desorption rate constant for the soil and plant root (day⁻¹),
• k_dis,S=first-order dissipation rate constant in the soil solution and soil particle at irreversible sorption site (day⁻¹),
• k_deg,P=first-order degradation rate constant in the plant shoot and root (day⁻¹),
• λ_Shoot=first-order growth rate constant of the plant shoot (day⁻¹), and
• t_uptake=time from sowing to the appearance of the uptake ability of pesticides by plant roots (day).

In this study, the adsorption and desorption rate constants for the plant root are assumed to be equal to those for the soil particle.

These ordinary differential equations can be solved by the Runge-Kutta-Gill method with the initial pesticide concentration in each compartment; the numerical solutions give the pesticide concentrations in the compartments of the soil water and particle and stem and leaf of the crop as a function of time t. A computer simulation program was developed with Visual Basic for Applications software (ver. 7.1) in Microsoft Excel 2016.

3. Plant uptake study in a growth chamber for model validation

The details of the plant uptake study in a growth chamber were reported previously.⁷⁾ In brief, 10 pesticides (analytical grade) were added to an uncontaminated Andosol [soil texture, loam; organic carbon, 52.1 g/kg; and water-holding capacity (WHC), 747.1 mL/kg]. Then, Wagner pots (1/5000 a) were filled with the prepared soil. We sowed seeds of the leafy vegetable (B. rapa L. var. perviridis ‘Yokattana’) in the soil on the next day and raised them in a growth chamber at 20°C under a 12 : 12 hr light : dark cycle. The drain hole of the pots was closed during the experimental period. We measured the weight of each pot and adjusted the soil moisture to maintain 50–70% WHC by supplying water accordingly, and we defined the total value of the decrease amount during the period from sowing to harvesting as the transpiration rate. At 10, 18, 25, 32, 39, and 60 days after sowing, shoots and roots were harvested. The roots were washed and sonicated in water to remove soil particles.¹⁶⁾ Fresh weights and moisture contents of the shoots and roots were measured in each sample, and then the samples were analyzed for pesticide concentrations. These uptake experiments were conducted in quadruplicate. Conditions of the plant uptake study are shown in Table 1.

Table 1. Conditions of the plant uptake study in a growth chamber

Specific conditions	Unit	Value
Surface area of Wagner pot	m2	0.02
Depth of the soil layer (ds)	m	0.17
Bulk density of the soil (ρb)	g/mL	0.6
Volumetric water content of the soil (θ)	v/v	0.27
Total carbon content of the soil	%	5.2
Grown period of leafy vegetable ‘Yokkatana’ (proper harvest time)	day	39
Planting density of the plant (ρplant)	stock/m2	50a)
Initial mass of the plant shoot (mShoot,0)	g FW/stock	0.003
Maximum mass of the plant shoot (mShoot,max)	g FW/stock	324.7
Growth rate constant of the plant shoot (λShoot)	day−1	0.26
Initial mass of the plant root (mRoot,0)	g FW/stock	0
Maximum mass of the plant root (mRoot,max)	g FW/stock	56.5
Growth rate constant of the plant root (λRoot)	day−1	0.27
Volumetric flow rate of transpiration stream of the plant (qT)	mL/g FW/day	2.0

^a) Initial value at sowing.

To analyze pesticides in a soil solution, the prepared soil including 10 pesticides adjusted to 60% WHC was used to fill quadruplicate stainless-steel vessels (100 mL) with a lid and placed in darkness at 20°C. After 1, 3, 7, 14, 22, and 36 days after treatment, the soil solution was collected by centrifuging, and a composite sample was used to quantify the pesticide concentrations. Experimental conditions, the analytical method, and the results are described in detail in our previous report.⁷⁾

4. Uptake and distribution of procymidone and flutolanil in B. rapa under hydroponic conditions

To evaluate a metabolic ability of B. rapa, we conducted an uptake experiment for procymidone and flutolanil with a similar log P_ow (3.3 and 3.77) under a hydroponic condition. The test solution for the experiment was prepared as follows.¹⁷⁾ Procymidone and flutolanil (Fujifilm Wako Pure Chemical Corporation, Osaka, Japan) were dissolved in acetone (Fujifilm Wako Pure Chemical) to 100 mg/L stock solution. A 600-µL aliquot of the stock solution was mixed into 600 mL of the solution containing 0.5 mM CaCl₂ and 2 mM 2-(N-morpholino) ethanesulfonic acid (MES) buffer (pH 5.8). Then, this solution was ultrasonicated for 30 min. The final concentrations in the test solution were 129.3±2.7 µg/L of procymidone and 140.7±14.0 µg/L of flutolanil.

The seedling of B. rapa was transferred to a stainless-steel vessel (140 mm height ×82 mm inner diameter) filled with 600 mL of the test solution. The experiment was conducted in a growth chamber (Koito Kogyo, Tokyo, Japan) at 20°C under a 12 : 12 hr light:dark cycle. The test solution was not aerated to avoid the volatilization of the pesticides during the uptake process. After 24 hr, the root of the seedling was rinsed in 600 mL Milli-Q water and then the seedling was transferred to the stainless-steel vessel with 600 mL of a hydroponic solution not including the pesticides with aeration. After 0, 1, 3, and 7 days, the shoots and roots of the seedling were sampled separately. Fresh weights of the shoots and roots were measured in each sample, and then the samples were analyzed for pesticide concentrations. Transpiration rate was calculated from the volume loss of the solution. The experiment was conducted in quadruplicate. The analytical method of the pesticide concentration in the solution, shoots, and roots are described our previous report.⁶⁾ We estimated the translocation from the roots to shoots and metabolic ability of B. rapa from the measured concentrations of the seedling.

Results and discussion

1. Shoot and root growth

The wet weight of the shoot and root of B. rapa grown in the Wagner pots increased exponentially to be 64 and 5.4 g FW/stock at 39 days after sowing, respectively, corresponding to an appropriate harvest weight, and then the growth rate decreased gradually (Fig. 2). Until the appropriate harvest time, FW of the plant root was about one-tenth of that for the plant shoot. In our first report,⁸⁾ we found that plant shoot growth in a field study can be expressed by a logistic equation. The growth of the shoot as well as the root was also approximated by a logistic function in the growth chamber study. The growth rate constants of the shoot (λ_Shoot) and root (λ_Root) were about the same: 0.26 and 0.27 (day⁻¹), respectively.

Fig. 2. Changes in the weight of the plant shoot and root: (●) and (▲) measured; and (—) and (—) fitted curve using a logistic function.

The transpiration rate of the leafy vegetable for the plant uptake study in a growth chamber increased in direct proportion to the plant shoot FW (Fig. 3), that is, the volumetric flow rate of transpiration stream on a FW basis (q_T) was constant (average value of 2.0 mL/g FW/day). We also found that the TSCF values of dinotefuran, imidacloprid, clothianidin, and thiacloprid [which had a high-water solubility (285–40,000 mg/L) and a low log P_ow (−0.549 to 1.26)] remained virtually constant for the growing period, whereas the values of pesticides (except for tolclofos-methyl) with a log P_ow of 1.68–3.77 at 10 days after sowing were lower than those from then on (Supplemental Fig. S1). From the results, the time from sowing to the appearance of the uptake ability of pesticides by plant roots (t_uptake) was set to be 8 days for dinotefuran, imidacloprid, clothianidin, and thiacloprid; the t_uptake value for the other pesticides was set at 10 days. In this period, we assumed that the transport process of pesticides into plant root as well as the uptake process had not occurred.

Fig. 3. Relationship between transpiration rate via the plant root and weight of the plant shoot.

2. Parameters for the PostPLANT-Soil model

2.1. Soil sorption and dissipation processes

The values of the soil adsorption coefficient (K_d) for the 10 pesticides were determined by the batch method, i.e., the soil:solution ratio is 1 : 5 under a slurry condition (Table 2).¹⁸⁾ Therefore, the K_d values must be adjusted to actual sorption ability under the soil moisture conditions (50–70% WHC) in the plant uptake study using the corrective coefficient (f_adj) in the same manner as in our first report.⁸⁾ The f_adj values of the 10 pesticides were estimated as described in Section 3.2.

Table 2. Parameters for soil sorption, dissipation, and root uptake process using the improved PostPLANT-Soil model

Pesticide	k1a)(day−1)	k2b)(day−1)	Soil adsorption coefficient, Kd(mL/g)c)	fadj for soil solutiond)	kdese)(day−1)	kdis,Sf)(day−1)	log Powg)	RCFh)
Dinotefuran	0.0099	0.011	3.7	0.45	0.010	0.011	−0.549 (25°C)	0.83
Imidacloprid	0.0084	0.0093	73.9	0.55	0.0088	0.0088	0.57 (21°C)	0.90
Clothianidin	0.0085	0.0094	48.6	0.55	0.0089	0.0089	0.7 (25°C)	0.92
Thiacloprid	0.015	0.017	170.9	0.45	0.016	0.016	1.26 (20°C)	1.10
Fosthiazate	0.024	0.027	4.9	0.35	0.025	0.026	1.68 (25°C)	1.41
Metalaxyl	0.022	0.025	4.6	0.30	0.023	0.023	1.75 (25°C)	1.49
Fenobucarb	0.023	0.026	5.4	0.30	0.024	0.024	2.67 (25°C)	4.25
Procymidone	0.026	0.029	30.3	0.50	0.018	0.018	3.3 (25°C)	11.3
Flutolanil	0.030	0.034	26.8	0.40	0.024	0.024	3.77 (25°C)	25.0
Tolclofos-methyl	0.0086	0.0096	289.1	0.50	0.0091	0.0091	4.56 (25°C)	98.8

^a) First-order dissipation rate constant in the first compartment for DFOP model. k₁=0.9·k₂. ^b) First-order dissipation rate constant in the second compartment for DFOP model. ^c) Measured values by a batch method for the andosol corresponding to an uptake study. ^d) Corrective coefficient for the K_d values for the pesticide concentrations in soil solution. ^e) Desorption rate constant in the soil solution and soil particle at reversible sorption site. Adsorption rate constant (k_ads) is assumed to be equal k_des. ^f) Dissipation rate constant in the soil solution and soil particle at ireversible sorption site. ^e) logarithmic n-octanol–water partition coefficient. ^h) Root concentration factor calculated from log P_ow by the equation of Briggs et al. (Ref. 12)

In our first report,⁸⁾ changes in concentrations of the total-extracted pesticides were fitted to a single first-order (SFO) model during the experimental period of 35 days. In the PostPLANT-Soil model, sorption process between soil solution and plant root was expressed by a kinetic concept. Therefore, we selected a DFOP model to describe the dissipation patterns of the total-extracted pesticides in soil because values of the adsorption and desorption rate constants (k_ads and k_des) must exceed 0 (i.e., when an SFO model is selected, the k_ads and k_des values are 0). If the k_ads value is assumed to be equal to k_des, the input parameters related to sorption and degradation processes can be expressed by the DFOP parameters (g, k₁, and k₂) as follows^10,19):

  

(10)

where

  

The values of g, k₁, and k₂ for the 10 pesticides were estimated as described in Section 3.2.

2.2. Uptake and transport process of pesticides for plant root

In our first report, the TSCF values of the 10 pesticides were calculated using the equation of Briggs et al.,⁹⁾ which was also used for the model simulation in this study. The volumetric flow rate of the transpiration stream (Q_T, mL/stock/day) in Eqs. (5) and (8) can be described by its flow rate on a FW basis (q_T) from the growth chamber study as described in Section 1.

  

(11)

Briggs et al.⁹⁾ also reported the relationship between RCF and log P_ow for the same condition:

  

(12)

The RCF values for the 10 pesticides were estimated using this relationship (Table 2).

3. Validation of the PostPLANT-Soil model

3.1. Sensitivity analysis for pesticide concentrations in plant root

In our first report,⁸⁾ a sensitivity analysis was performed to determine the relative importance of the input parameters used in the PostPLANT-Soil model. In the analysis, the model outputs were set to be the pesticide concentrations in the plant shoot at the appropriate harvest time. The analysis showed that the following input parameters were highly sensitive to shoot concentrations: TSCF, q_T, and K_d and its f_adj.

In this study, we performed a sensitivity analysis of the input parameters used in the PostPLANT-Soil model. The following parameters were selected in the analysis: K_d, the f_adj of K_d, RCF, k_deg,P, and λ_Shoot. The model outputs were set to be the pesticide concentrations in the plant root at the appropriate harvest time (39 days after sowing). Sensitivity of the model output to one input parameter can be measured by the sensitivity ratio (SR).²⁰⁾ Analysis was carried out independently for a +10% change of each input parameter value under the conditions. The higher the absolute SR value, the more sensitive is the model output to the input parameters. The RCF value calculated by Eq. (12) remained almost constant (RCF ≅ 1) at a log P_ow<1. The RCF value increased sharply with increasing log P_ow values at >3. Therefore, three pesticides, i.e., clothianidin, procymidone, and tolclofos-methyl with log P_ow of 0.7, 3.3, and 4.56, respectively, were selected for the analysis. Because information about degradability in the plant root is insufficient for these pesticides, a short and long dissipation time for 50% (DT₅₀) in the root was assumed to be 1 and 30 days, respectively, and the k_deg,P values were obtained.

The SR values for clothianidin, procymidone, and tolclofos-methyl applied to the PostPLANT-Soil model are shown in Fig. 4. The difference in the results among the pesticides was very small. The K_d and f_adj were most sensitive to decrease in the simulated pesticide concentrations of the plant root (−1.2< SR<−0.94). Meanwhile, increase in the RCF value led to an increase in these concentrations (SR ≅1). The results indicate that the parameters related to soil sorption processes are key factors for determining mass distribution of phytoavailable pesticides in the soil solution and that RCF represents the degree of sorption efficiency to the plant root.

Fig. 4. Sensitivity ratios of the PostPLANT-Soil model outputs to the input parameters. The model outputs for sensitivity analysis were the pesticide concentrations in the plant root at the appropriate harvest time (39 days after sowing).

The model output for easily degradable pesticides in the plant root (DT₅₀=1 day) was more sensitive than that for pesticides with low degradability (DT₅₀=30 days). Sensitivity to pesticide root concentrations may be subject to the degradation rate constant in the plant root (k_deg,P) as well as to the crop growth period.

The increase in the growth rate constant of the plant shoot (λ_Shoot) caused a decrease in the pesticide concentrations. This was easily derived from the dilution effect on the pesticide concentration with root growth.

3.2. Model validation with the results of plant uptake study

Changes in the measured pesticide concentrations in the soil solution (Supplemental Table S1), which were obtained by the centrifugation method from the soil samples in the stainless-steel vessels, could be approximated by a first-order reaction curve (Fig. 5). However, dissipation rate differed among the pesticides. In this study, pesticide concentrations of the total extracts in the soil samples were not analyzed. Therefore, changes in concentrations of the total-extracted pesticides were estimated from the results of the pesticide residue study in an upland Andosol field of our first report.⁸⁾ In the field study, measured concentrations of the total extracts for the 10 pesticides were fitted to an SFO model during the experimental period (35 days) and the first-order dissipation rate constant (k_SFO, day⁻¹) for each pesticide was obtained.

Fig. 5. Simulated (solid and dashed lines) and measured (symbols) pesticide concentrations in the plant shoot and root, and those in the soil solution. Open square symbols for the plant root (□) indicate under the limit of quantification (0.01 µg/g FW). The measured concentrations in the soil solution were analyzed for a composite sample from four vessels. The shoot and root concentrations indicate the average values (n=4).

In this study, we selected a DFOP model and assumed the model parameters as g=0.5 and k₁=0.9·k₂ in Eq. (1) for describing the dissipation curves of the 10 pesticides as mentioned in Section 2.1. Under this situation, changes in concentrations of the total-extracted pesticides in soil showed nearly first-order dissipation curves. The mass fraction of the soil particle at irreversible sorption site (f_soil,2) was 0.997, obtained using Eq. (10), i.e., the compartment at reversible sorption site was very small.

To estimate the dissipation behavior of the total-extracted pesticides in the soil, the k_SFO values obtained from the field data were adjusted to coincide with the dissipation curves of the pesticide concentrations in the soil solution. The adjusted k_SFO values of the 10 pesticides were regarded as the dissipation rate constants in the second compartment (k₂) for the DFOP model (Table 2). The k₂ values of dinotefuran, imidacloprid, clothianidin, metalaxyl, procymidone, and flutolanil with a longer DT₅₀ (28–75 days) in the field soil were equivalent to or larger than the original k_SFO values (by a factor of 1–2). In contrast, the k₂ values of thiacloprid, fosthiazate, fenobucarb, and tolclofos-methyl with a shorter DT₅₀ (7–16 days) were lower than the k_SFO values (by a factor of 0.2–0.5).

The f_adj of the K_d was calibrated individually to ensure the measured concentration of the pesticides in the soil solution as found in our first report.⁸⁾ However, the f_adj value of tolclofos-methyl with a high adsorption property (K_d=289 mL/g) was adjusted to ensure the measured concentration in the plant shoot, since the simulated shoot concentration of tolclofos-methyl with the calibrated f_adj value for the measured concentration in the soil solution was much higher than the measured shoot concentration. This is probably because the measured concentration of tolclofos-methyl in the solution using the centrifugation method was underestimated compared to the actual concentration due to its high adsorption property. The calibrated f_adj values of 0.30–0.55 in this study (Table 2) were almost the same as those for our field study (0.15–0.60). In our first report,⁸⁾ there was a trend toward an increase in f_adj values with increase in the log P_ow value of the pesticides, whereas no such trend was observed in the present study. This is likely because downward or lateral flow of soil water due to rainfall events on the field led to decreased concentration of the pesticides with a small log P_ow in the plowed soil layer, whereas there was no movement of the soil solution in the vessels in this study.

The leafy vegetables in the Wagner pots were cultivated under similar soil conditions as in the vessels, i.e., soil water did not flow out due to an adequate water supply because the drain hole of the pots was closed during the experimental period. The measured concentrations of the pesticides in the plant shoot sampled from the pots reached maxima at 19 days after sowing (except for procymidone at 33 days and flutolanil at 26 days) and then declined gradually until the appropriate harvest time of 39 days after sowing (Fig. 5 and Supplemental Table S1). The maximum concentrations varied greatly among the pesticides from 0.001 µg/g FW (tolclofos-methyl) to 0.37 µg/g FW (dinotefuran).

In the model simulation, the degradation rate constant in the plant shoot and root (k_deg,P) was set at 0 because there was little knowledge concerning plant metabolism for the pesticides. The simulated shoot concentrations of dinotefuran, imidacloprid, clothianidin, thiacloprid, and tolclofos-methyl were in good agreement with the measured values. However, the simulated concentrations of fosthiazate, metalaxyl, fenobucarb, and flutolanil were larger than the measured values, whereas those of procymidone were slightly small. To assess goodness of fit between the measured and simulated shoot concentrations, root-mean-square error (RMSE, %)²¹⁾ was calculated for the 10 pesticides as reported in our first report⁸⁾ (Table 3). In general, the lower the RMSE, the better the agreement between measured and simulated data. The RMSE values of dinotefuran, imidacloprid, clothianidin, thiacloprid, procymidone, and tolclofos-methyl were less than 130%, whereas those of fosthiazate, metalaxyl, fenobucarb, and flutolanil exceeded 200%. The ratio of the simulated concentration in the plant shoot at 39 days after sowing (appropriate harvest time) to the measured concentration (simulated/measured ratio) are shown in Table 3. The simulated/measured ratios of dinotefuran, imidacloprid, clothianidin, thiacloprid, procymidone, and tolclofos-methyl were 0.2–2.3, whereas those of fosthiazate, metalaxyl, fenobucarb, and flutolanil were 7.0–26.4.

Table 3. Statistical analysis of the measured and simulated concentrations of the 10 pesticides in the plant shoot and root based on the root-mean-square error (RMSE)

Pesticide	RMSE (%)		Simulated/measured ratioa)
	shoot	root	shoot	root
Dinotefuran	16.2	−b)	1.2	−b)
Imidacloprid	73.9	−	2.2	−
Clothianidin	128.6	−	2.3	−
Thiacloprid	76.1	−	1.9	−
Fosthiazate	766.2	−	14.0	−
Metalaxyl	212.5	−	26.4	−
Fenobucarb	2410.1	681.6	18.1	4.6
Procymidone	76.1	25.1	0.2	0.7
Flutolanil	789.7	267.3	7.0	2.6
Tolclofos-methyl	48.0	98.7	1.4	0.4

^a) Ratio of the simulated concentration in the plant shoot and root to the measured one at 39 days after sowing. ^b) Not calculated because the measured concentrations were less than the limit of quantification (0.01 µg/g FW).

The measured root concentrations of the 10 pesticides are shown in Supplemental Table S1. Dinotefuran, imidacloprid, clothianidin, thiacloprid, fosthiazate, and metalaxyl were below the limit of quantification (LOQ, 0.01 µg/g FW), likely due to the high-water solubility (285–40,000 mg/L) and low log P_ow (−0.549–1.75) of these pesticides. In contrast, procymidone, flutolanil, and tolclofos-methyl had low water solubility (1.1–6.63 mg/L) and large log P_ow (3.3–4.56) and were detected at concentrations above the LOQ. The calculated RMSE values for the root concentrations were also less than 100% for procymidone and tolclofos-methyl and larger than 200% for fenobucarb and flutolanil (Table 3). The simulated/measured ratios of procymidone and tolclofos-methyl were 0.7 and 0.4, respectively, and those for fenobucarb and flutolanil were 4.6 and 2.6, respectively. The model validation for the root concentrations was insufficient in this study because available data were limited to only four pesticides. Therefore, the model needs to be improved and validated with another experimental data on plant uptake.

3.3. Metabolic ability of B. rapa from uptake study under hydroponic condition

In this study, the simulated concentrations of the plant shoot for fosthiazate, metalaxyl, fenobucarb, and flutolanil were overestimated compared to measured values. Therefore, these pesticides may be involved in plant metabolism. In an uptake experiment using tomato seedlings grown in hydroponic solutions,²²⁾ fosthiazate was not detected in the plant shoot after 10 days of treatment and four polar metabolites were detected. Similar uptake experiment for metalaxyl was reported in which the parent compound quickly decreased to be 33% of total residual radioactivity in the plant shoot at 3 days after treatment.²³⁾

Changes in amount of procymidone and flutolanil in root and shoot were shown in Fig. 6 from the result of our uptake experiment under a hydroponic condition. Amounts of the two pesticides in the plant root decreased rapidly at 1 day after transplantation and were then constant until 7 days. In the plant shoot, the amount of flutolanil also decreased rapidly at 1 day, whereas that of procymidone did not decrease for 7 days. These results show that procymidone did not degrade in the shoot, whereas flutolanil was readily taken up by the plant root and easily translocated and/or metabolized in the plant shoot. The DT₅₀ value in the shoot was estimated to be 0.28 days (k_deg,P=2.5 day⁻¹) by fitting the measured data according to an SFO model. We simulated the shoot concentrations of flutolanil with consideration of the dissipation in the shoot by the model. The simulated flutolanil concentrations agreed closely with the measured concentrations (dashed line in Fig. 5), because the RMSE value of the shoot markedly decreased from 789.7 to 70.2.

Fig. 6. Changes in amount of procymidone and flutolanil in root and shoot of B. rapa. After uptake process of the pesticides for 24 hr, the seedlings of B. rapa were transferred to a hydroponic solution without procymidone and flutolanil. Error bars indicate SEM (n=4). Data were analyzed by one-way ANOVA followed by Tukey’s multiple comparison test (p<0.01). Within a day, means followed by the same letter are not significantly different.

In contrast, the simulated concentrations of procymidone, which did not degrade in the plant shoot, were underestimated compared to measured values. The TSCF value of the pesticides from the uptake experiment described above was derived from the pesticide concentration in xylem sap of the plant shoot divided by that in the test solution at the end of the uptake process for 24 hr. However, it is difficult to directly measure the concentration in xylem sap, therefore, it was estimated as the amount of pesticide in the shoot divided by the transpiration volume (Table 4).^24,25) The TSCF value of 0.52 for procymidone (Table 4) was about twice the estimated value of 0.30 from log P_ow value using the equation of Briggs et al.⁹⁾ The simulated shoot concentrations of procymidone using the measured TSCF were larger than those using the estimated TSCF and agreed more closely with the measured concentrations (dashed line in Fig. 5), because the RMSE of 56.8 for the measured TSCF was smaller than the 76.1 for the estimated TSCF.

Table 4. Concentration distribution and transpiration stream concentration factor (TSCF) of procymidone and flutolanil at the end of the uptake process of the pesticides for 24 hr

Pesticide	Concentration (ng/g FW)			TSCFb)
	Test solutiona)	Root	Shoot
Procymidone	128.5±3.8	252.3±7.7	391.8±13.1	0.52±0.01
Flutolanil	126.0±5.6	172.3±4.4	62.0±4.6	0.08±0.01

^a) Hydroponic solution containing procymidone and flutolanil.

^b) TSCF=(amount in shoots/transpiration volume)/(concentration in test solution). The shoot weight was 5.32±0.58 g FW, and the transpiration volume was 31±3 mL at the end of the uptake process.

Conclusion

We validated a simulation model (PostPLANT-Soil) explained in our first report for predicting pesticide concentrations in succeeding leafy vegetables. The model validation using the measured data from a detailed plant uptake study in a growth chamber showed that the pesticide shoot concentrations of B. rapa were correlated with the concentrations in the soil solution as in our first report.⁸⁾ The model successfully simulated the concentration changes of pesticides in plant shoot by considering the key parameters: f_adj, and TSCF. However, the simulated concentrations of the shoot for several pesticides were overestimated compared to the measured values. The leafy vegetable (B. rapa) used in this study probably have a high metabolic ability for flutolanil from the result of the uptake study under a hydroponic condition. Therefore, appropriately assessing the risk of pesticide residues in succeeding crops requires knowledge of the plant metabolism of each pesticide. Furthermore, the model must be improved and validated for predicting pesticide concentrations more precisely in plant shoot as well as root.

Acknowledgements

This work was supported by the Environmental Research and Technology Development Fund (5-1703) of the Ministry of the Environment, Japan.

Electronic supplementary materials

The online version of this article contains supplementary material, which is available at https://www.jstage.jst.go.jp/browse/jpestics/

References

• 1) S. Hori: The Japanese positive list system for agricultural chemical residues in foods and food additives. Foods Food Ingredients J. Jpn. 212, 496–508 (2007), in Japanese.
• 2) M. Tanigawa: Status of pesticides exceeding residue standards based on information published on the Internet. J. Pestic. Sci. 35, 537–541 (2010), in Japanese.
• 3) T. Otani, N. Seike and Y. Sakata: Differential uptake of dieldrin and endrin from soil by several plant families and Cucurbita genera. Soil Sci. Plant Nutr. 53, 86–94 (2007).
• 4) J. C. White: Plant-facilitated mobilization and translocation of weathered 2,2-bis(p-chlorophenyl)-1,1-dichloroethylene(p,p′-DDE) from an agricultural soil. Environ. Toxicol. Chem. 20, 2047–2052 (2001).
• 5) M. P. N. Gent, J. C. White, Z. D. Parrish, M. Isleyen, B. D. Eitzer and M. I. Mattina: Uptake and translocation of p,p′-dichlorodiphenyldichloroethylene supplied in hydroponics solution to Cucurbita. Environ. Toxicol. Chem. 26, 2467–2475 (2007).
• 6) S. Namiki, T. Otani, Y. Motoki, N. Seike and T. Iwafune: Differential uptake and translocation of organic chemicals by several plant species from soil. J. Pestic. Sci. 43, 96–107 (2018).
• 7) S. Namiki, N. Seike and Y. Motoki: Relationship between growth stage of Brassica rapa var. perviridis and the abilities for uptake and translocation of pesticides in soil. J. Pestic. Sci. 44, 1–8 (2019).
• 8) S. Namiki, K. Inao, Y. Motoki, N. Seike and E. Watanabe: A simulation model (PostPLANT-Soil) for predicting pesticide concentrations in succeeding leafy vegetables: I. Validation with experimental data in a Japanese Andosol field. J. Pestic. Sci., (2023), in press.
• 9) L. L. van Eerd, R. E. Hoagland, R. M. Zablotowicz and J. C. Hall: Pesticide metabolism in plants and microorganisms. Weed Sci. 51, 472–495 (2003).
• 10) K. A. Ford and J. E. Casida: Comparative metabolism and pharmacokinetics of seven neonicotinoid insecticides in spinach. J. Agric. Food Chem. 56, 10168–10175 (2008).
• 11) S. Namiki, T. Otani, Y. Motoki and N. Seike: The influence of Brassica rapa var. perviridis growth conditions on the uptake and translocation of pesticides. J. Pestic. Sci. 43, 248–254 (2018).
• 12) G. G. Briggs, R. H. Bromilow and A. A. Evans: Relationships between lipophilicity and root uptake and translocation of non-ionised chemicals by barley. Pestic. Sci. 13, 495–504 (1982).
• 13) https://esdac.jrc.ec.europa.eu/public_path/projects_data/focus/dk/docs/FOCUSkineticsvc1.1Dec2014.pdf/ (Accessed 20 Feb., 2023)
• 14) https://esdac.jrc.ec.europa.eu/projects/degradation-kinetics-software/ (Accessed 20 Feb., 2023)
• 15) M. M. S. ter Horst, J. J. T. I. Boesten, W. van Beinum and S. Beulke: Acceptability of inversely-modelled parameters for non-equilibrium sorption of pesticides in soil. Environ. Model. Softw. 46, 260–270 (2013).
• 16) J. L. Gomez-Eyles, M. T. O. Jonker, M. E. Hodson and C. D. Collins: Passive samplers provide a better prediction of PAH bioaccumulation in earthworms and plant roots than exhaustive, mild solvent, and cyclodextrin extractions. Environ. Sci. Technol. 46, 962–969 (2012).
• 17) https://www.oecd-ilibrary.org/environment/guidance-document-on-aquatic-toxicity-testing-of-difficult-substances-and-mixtures_0ed2f88e-en/ (Accessed 11 Jul., 2023)
• 18) Y. Motoki, T. Iwafune, N. Seike, K. Inao and T. Otani: Effect of Time-dependent sorption on the dissipation of water-extractable pesticides in soils. J. Agric. Food Chem. 64, 4478–4486 (2016).
• 19) https://esdac.jrc.ec.europa.eu/public_path/projects_data/focus/gw/NewDocs/focusGWReportOct2014.pdf/ (Accessed 20 Feb., 2023)
• 20) U.S.EPA: Risk Assessment Guidance for Superfund: Volume III-Part A, Process for conducting probabilistic risk assessment, EPA 540-R-02-002, 2001.
• 21) D. G. Karpouzas, S. Cervelli, H. Watanabe, E. Capri and A. Ferrero: Pesticide exposure assessment in rice paddies in Europe: a comparative study of existing mathematical models. Pest Manag. Sci. 62, 624–636 (2006).
• 22) http://www.fsc.go.jp/fsciis/attachedFile/download?retrievalId=kya20220323041&fileId=210/ (Accessed 20 Feb., 2023)
• 23) M. Kubicki, M. Lamshöft, A. Lagojda and M. Spiteller: Metabolism and spatial distribution of metalaxyl in tomato plants grown under hydroponic conditions. Chemosphere 218, 36–41 (2019).
• 24) M. G. T. Shone, O. B. Bartlett and A. V. Wood: A comparison of the uptake and translocation of some organic herbicides and a systemic fungicide by barley. II. Relationship between uptake by roots and translocation to shoots. J. Exp. Bot. 25, 401–409 (1974).
• 25) S. Trapp: Modelling uptake into roots and subsequent translocation of neutral and ionisable organic compounds. Pest Manag. Sci. 56, 767–778 (2000).