VZJ Download to Citation Manager
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 8 March 2006
Published in Vadose Zone J 5:273-282 (2006)
DOI: 10.2136/vzj2005.0021
© 2006 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Karpouzas, D. G.
Right arrow Articles by Papadopoulou-Mourkidou, E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Karpouzas, D. G.
Right arrow Articles by Papadopoulou-Mourkidou, E.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Karpouzas, D. G.
Right arrow Articles by Papadopoulou-Mourkidou, E.
Related Collections
Right arrow Watershed and Landscape Processes
Right arrow Pesticides
Right arrow Vadose Zone Processes and Chemical Transport

SPECIAL SECTION: FROM FIELD- TO LANDSCAPE-SCALE VADOSE ZONE PROCESSES

Basin-Scale Risk Assessment in Rice Paddies

An Example Based on the Axios River Basin in Greece

Dimitrios G. Karpouzasa,*, Ettore Capria and Euphemia Papadopoulou-Mourkidoub

a Univ. Cattolica del Sacro Cuore, Istituto di Chimica Agraria ed Ambientale, Sezione Chimica Vegetale, Via E. Parmense 84, Piacenza 29100, Italy
b Aristotle Univ. of Thessaloniki, Lab. of Pesticide Science, Thessaloniki 54124, Greece
* Corresponding author (dimitrios.karpouzas{at}unicatt.it)

Received 8 February 2005.



   ABSTRACT
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
RICE (Oryza sativa L.) is cultivated in large river basins in Europe, where high loads of applied herbicides have resulted in the contamination of related surface water (SW) and groundwater (GW) systems. Therefore, risk assessment of pesticides used on rice should be performed at the basin scale. This study reports the development and validation of a basin-scale scenario, representative of rice cultivation in the main rice cultivation area in Greece, the Axios river basin. Soils of the rice-cultivated region of the basin are characterized as heavy clay, clay loams with poor infiltration. A 2000-ha rice-cultivated basin was simulated consisting of rice management blocks (200 ha) associated with drainage canals and a cross-linked river. This system was conceptualized using a combination of rice water quality (RICEWQ 1.6.2v) and river water quality (RIVWQ 2.02) models. Application of RICEWQ 1.6.2v for a 20-yr simulation period showed that the 90th percentile of the annual GW predicted environmental concentrations (PECs) for the herbicides propanil (3',4'-dichloropropionanilide) and molinate (S-ethyl perhydroazepine-1-carbothioate) were 3.2 fg L–1 and 1.9 pg L–1, respectively, suggesting low risk to GW. Pesticide mass and water releases from simulated paddies were utilized as inputs for canals and river, and the fate of propanil and molinate was simulated with the RIVWQ model. Comparison of the 90th percentile of maximum daily river PECs with the maximum measured concentrations of propanil and molinate in the Axios river in 1994 revealed acceptable agreement in both the magnitude of values and their temporal distribution. In general, the 90th percentiles of the maximum daily PECs of propanil in both canals (1.96 µg L–1) and rivers (1.15 µg L–1) were markedly lower than the corresponding PECs of molinate (6.07 and 0.394 µg L–1, respectively). Increases in the period of paddy closure after application of molinate resulted in reduced SW PECs, suggesting that the selection of water management practices should be done according to pesticide persistence and local agronomic conditions.

Abbreviations: DAS, days after seeding • DAT, days after treatment • DCT, days of closure time • FOCUS, FOrum for the Coordination in the USe of models • GAP, good agricultural practice • GW, groundwater • Med-Rice, Mediterranean Rice Group • PEC, predicted environmental concentration • RICEWQ, Rice Water Quality model • RIVWQ, River Water Quality model • SW, surface water • TER, toxicity exposure ratio • VADOFT, vadose zone flow and transport model


   INTRODUCTION
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
RICE is commonly cultivated in Europe in large river basins. Such basins can be located close to urban areas, such as the Axios river basin in northern Greece, or within zones of high ecological value, such as the National Park of Donana and the Lagoon of Valencia in Spain (Ramos et al., 2000). Monitoring studies conducted in the main rice cultivation basins in Greece (Papadopoulou-Mourkidou et al., 2004), Italy (Readman et al., 1993), and Portugal (Cerejeira et al., 2003) revealed the presence of relatively high concentrations of herbicides commonly applied to rice, including molinate and propanil, in related SW and GW systems. Artificial (e.g., drainage canals) and natural SW bodies (e.g., streams, rivers) are present in most of the rice-cultivated basins in Europe constituting a unique ecosystem hosting birds, arthropods, and other living organisms. Therefore, it is most relevant to consider pesticide risk assessment for rice crop at the basin-scale level.

The increasing cost of pesticide monitoring and field studies have made mathematical modeling an indispensable part of the pesticide registration process. Although a uniform approach for proper model use has now been established in Europe (FOrum for the Coordination in the USe of models [FOCUS], 2000, 2001), this could not be used for rice because of the unique flooding conditions applied to rice cropping in Europe. Currently, the PECs of pesticides applied in rice paddies are calculated with a Tier 1 spreadsheet, which was developed within the framework of the European Commission (Mediterranean Rice Group [Med-Rice], 2003). However, the use of more sophisticated mathematical models is essential in cases where refined modeling or application of mitigation strategies must be applied. The rice water quality (RICEWQ) 1.6.2v model is the only validated model at present that can adequately predict the environmental fate of pesticides in rice paddies and provide PECs for GW and immediately adjacent SW systems (Miao et al., 2003a; Karpouzas et al., 2005). However, recent studies by Miao et al. (2003b) revealed that RICEWQ 1.6.2v could be an overly conservative model for predicting the environmental concentrations of pesticides in related SW systems. In the same study, Miao et al. (2003b) revealed that the combination of RICEWQ with the river water quality (RIVWQ) 2.02v model provided a more realistic estimation of PECs in SW bodies associated with treated rice paddies.

The existence of significant differences in crop, water, and pesticide management practices among the various rice-cultivating countries in Europe necessitates the development of various well-defined national rice scenarios instead of one or a few pan-European scenarios (Med-Rice, 2003). Therefore, in a similar way to the FOCUS SW scenarios developed for other crops (FOCUS, 2001), the tiered approach for rice may consist of lower tier assessments (i.e., Step 1 and 2), involving the use of established conservative scenarios at the paddy-field scale, Step 3 assessment including site-specific consideration and calculation of PECs at paddy-field scale, and Step 4 assessments including landscape risk assessment using basin-scale scenarios, where mitigation processes could also be included (Karpouzas and Capri, 2004). Such mitigation processes in rice cultivation could be the controlled prolongation of the period of paddy closure after application of certain pesticides. This measure has been found to be a significant parameter controlling the magnitude of molinate concentrations in SW bodies (Crepeau and Kuivila, 2000). A closure period of 3 to 4 d is recommended in Japan (Vu et al., 2004). In contrast, in the Sacramento Valley in California, farmers are obliged to maintain a closure period of 28 d after application of molinate to minimize pesticide load to related SW bodies (Newhart and Ceesay, 2001).

The aims of this study were (i) to develop and validate a basin-scale scenario representative of rice cultivation in Greece, which could potentially be used for pesticide regulatory purposes, and (ii) to identify the effect of a possible mitigation practice on the PECs of molinate in SW systems.


   MATERIALS AND METHODS
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Description of RICEWQ 1.6.2v and RIVWQ 2.02v Models
RICEWQ 1.6.2v, using daily time steps, simultaneously tracks mass balance of the chemical in the rice foliage, the water column, and the sediment. Subsequently, the pesticide mass and water volume predicted to leach below paddy sediment are considered as inputs for the vadose zone flow and transport model (VADOFT), which subsequently calculates, at daily time steps, pesticide concentrations at the selected GW level. The top 5 cm of the soil profile is represented by the active sediment layer in RICEWQ, and the remainder of the soil profile is represented as multiple compartments in VADOFT. The bottom of the active sediment layer is the interface between the two subsystems, the paddy water–sediment system and the soil profile below the top 5 cm, represented by the two models. The model also provides daily summaries of the amount of pesticide and water lost from the paddy system due to runoff or overflow and controlled drainage. These losses are used as water and pesticide mass input for the chemical transport model for riverine environments in RIVWQ 2.02.

RIVWQ 2.02. simulates the transport and fate of organic chemicals in riverine systems on the basis of the theory of constituent mass balance. The system geometry is represented using a link-mode approach in which the water body is divided into a number of discrete junctions connected by flow channels. Dynamic constituent transport occurs between junctions via links and is a balance between river-driven flows and dispersion processes. Chemical transformation occurs within each node, including dilution, volatilization, partitioning between water and sediment, decay in water and sediment, and resuspension from bed sediments. A detailed description of the model is given in the user's manuals for RICEWQ 1.6.2v (Williams et al., 1999) and RIVWQ 2.02 (Williams et al., 2004) and also in previous works (Miao et al., 2003a).

Model Parameterization According to the Developed Scenario
A basin-scale scenario was developed based on the agricultural, water management, and pesticide practices used in the Axios river basin (Fig. 1 ). The southern part of the Axios river basin constitutes more than 75% of the whole rice cultivation region in Greece and comprises a well-developed system of irrigation and drainage canals. Irrigation of the studied field in the south part of the basin is achieved with water supplied solely by the Axios river (Ntanos, 1997, 2001). Soils of the rice-cultivated area in the Axios river basin are mainly heavy clay, clay loams with poor infiltration (Anonymous, 2000). The simulated system was an intensively rice- cultivated basin of 2000 ha consisting of 500 paddy fields (4 ha each). Paddy fields were grouped in 10 management blocks of 50 paddies. Each of the paddy blocks was associated with a drainage canal receiving pesticide mass and water releases as predicted by the RICEWQ 1.6.2v model. Drainage canals were subsequently discharged into a large river system, and the fate of pesticide loads in the canals and the river was simulated by RIVWQ 2.02v (Fig. 1).


Figure 1
View larger version (44K):
[in this window]
[in a new window]
  		Fig. 1. A site map of the study area and a schematic representation of the basin-scale scenario developed.

 
A simulation period of 26 yr for RICEWQ 1.6.2v was considered according to FOCUS GW scenarios. To appropriately set soil moisture in the soil profile before the simulation period and because residues may take more than 1 yr to leach for certain pesticides, a 6-yr warm-up period was added to the start of the simulation period. Thus, simulation results during the warm-up period were ignored (FOCUS, 2000). Daily meteorological data (26 yr) were derived from the Athens meteorological station. The model requires daily precipitation (cm) and daily evapotranspiration (cm) data during the simulation period. Evapotranspiration is one of the major hydrological components of water balance in the paddy field. Evapotranspiration from rice fields can be directly monitored or estimated from other meteorological data (Vu et al., 2005). In the model, evapotranspiration is assumed equal to pan evaporation, which is a valid assumption for an aquatic environment (Linsley and Franjini, 1979).

A series of 10 different seeding dates during 20 d, from 20 April up to 7 May of every year, were selected and were assigned to each of the 500 paddy fields of the simulated basin following a completely randomized block design. This period represents the main establishment period for the rice crop in Greece. Subsequent agronomic practices including emergence date, maturation date, and harvest date were applied 7, 110, and 130 d after seeding (DAS). RICEWQ 1.6.2v does not allow both irrigation and drainage to occur concurrently. The continuous flow into and out of rice paddies in Greece was represented by a simulated water management where initiation of irrigation was triggered when the water level in the paddy had fallen below 8 cm and irrigation was stopped every time the water level exceeded 11 cm. Previous validation studies with the RICEWQ model have shown that such modeling practice could adequately represent the water balance of a continuous flow-in-and-out system (Miao et al., 2003a; Karpouzas et al., 2005).

The pesticides included in the study were the rice herbicides propanil and molinate. Propanil and molinate are commonly applied in rice paddies in Greece as late and early postemergence applications, respectively, for the control of barnyardgrass (Echinochloa spp.) and other annual weeds. The maximum recommended dose for both of these herbicides was used for model parameterization (Table 1). We considered a single application per year for both herbicides at 35 and 10 DAS for propanil and molinate, respectively. The mode of pesticide application to rice paddies is different for propanil and molinate. Molinate is commonly applied in flooded paddies, while propanil is usually applied in drained paddies. Thus, for simulating the application of propanil, paddies were left to drain for 4 d before its application, and they were subsequently reflooded 2 d after its application. The water and pesticide management practices used for model parameterization were in agreement with their good agricultural practices (GAP) documents and with information derived by local agronomic experts (EU, 2003).


View this table:
[in this window]
[in a new window]
  		Table 1. Values of parameters used by RICEWQ model.

 
Degradation rates of pesticides molinate and propanil in paddy water and sediment, are required for RICEWQ parameterization. In accordance with Med-Rice recommendations, degradation rates of pesticides applied to rice should be obtained from studies employed in aerobic water–sediment systems (Med-Rice, 2003). Consequently, the degradation rates of pesticides used for model parameterization were calculated from literature half-life values obtained from such studies assuming first-order decay (USEPA, 2003a, 2003b) (Table 1). Adsorption coefficients, Kd, for each pesticide were estimated by their respective literature organic C adsorption coefficients, Koc (Wauchope et al., 1992; Augustijn-Beckers et al., 1994).

In accordance with lower-tier Med-Rice scenarios, a clay soil horizon with an infiltration rate of 0.1 cm d–1, representing a high risk for SW contamination was considered for the parameterization of the RICEWQ model. This was selected since previous studies have indicated that most of the soils in the rice-cultivated area of the Axios river basin are characterized as heavy clay, clay loam soils with a low infiltration rate (Anonymous, 2000; Papadopoulou-Mourkidou et al., 2004). A uniform soil horizon of 1-m depth, which was subdivided into three soil zones (0–30, 30–60, and 60–100 cm), was considered for the parameterization of VADOFT submodel in accordance with Med-Rice recommendations (Med-Rice, 2003). The first layer (0–30 cm) of the vadose zone, which is situated right beneath rice paddies, was considered as a saturated and anaerobic soil layer (Greppi, 2004). Therefore, degradation rates, which are essential for the parameterization of the VADOFT submodel, were obtained from anaerobic soil studies. Finally, half-life values from aerobic soil studies were utilized for calculating degradation rates in the other two soil layers (30–60 and 60–100 cm). Pesticide degradation in soil is predominantly microbial (Kruger et al., 1997; Jones and Norris, 1998). Several studies so far have documented the reduction of microbial population and soil organic matter content with increasing soil depth (Helweg, 1992; Kruger et al., 1993; Fomsgaard, 1995). Therefore, the different soil horizons differ in their ability to adsorb and degrade pesticides. The differences are based on default assumptions introduced by the Med-Rice group (Med-Rice, 2003) and FOCUS GW scenarios (FOCUS, 2000) and assume that in the second horizon the organic C is 50% of the value of the first horizon (biofactor = 0.5) and in the third horizon it is 30% of the value of the first horizon (biofactor = 0.3). Both sorption and the degradation rate are decreased by these biofactors in the second and third horizons. The reductions of pesticide sorption in the second and third soil layer were included in the calculations of the retardation coefficient (R).

Field capacity (cm3 cm–3) and wilting point (cm3 cm–3) were derived from Baumer pedotransfer functions with the soil parameters estimate software SOILPAR (Research Institute of Industrial Crops, Bologna, Italy) (Acutis and Donatelli, 2003). The initial soil moisture content (cm3 cm–3) was set to field capacity. The values of the variables used for the parameterization of RICEWQ 1.6.2v (RICEWQ and VADOFT) model are shown in Tables 1 and 2.


View this table:
[in this window]
[in a new window]
  		Table 2. Physicochemical and hydraulic parameters used in the VADOFT submodel contained within RICEWQ 1.6.2v.

 
For the parameterization of RIVWQ, a clay sediment was considered for both the drainage canals and the river system (Med-Rice, 2003). Hydrological and other characteristics of the simulated SW systems are listed in Table 3. Hydrological characteristics and dimensions, including flow velocity (m s–1) and water discharge (m3 s–1) were derived by Kampa et al. (2000) and local observations, respectively. The drainage canals had a rectangular shape, with a total length of 5 km, depth of 1.5 m, top width of 4 m, and a flow velocity of 0.1 m s–1. The river systems simulated in both scenarios had a rectangular shape with a total length of 10 km, depth of 3 m, top width of 60 m, and a flow velocity of 0.3 m s–1. Nodal spacings of 100 m for the drainage canals and 500 and 1000 m for the river system were selected for modeling purposes. Pesticide mass and water volume discharge from drainage canals into the river system occurs at 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, and 5.5 km along the length of the river.


View this table:
[in this window]
[in a new window]
  		Table 3. Values of parameters used by the RIVWQ model.

 
Validation of the Scenario Developed
The basin-scale scenario developed was validated by comparing river PECs derived by the application of RIVWQ 2.02v model with measured data obtained from a monitoring study conducted in the Axios river basin in 1994. Water samples were collected from the furthest downstream part of the Axios river during 1994 and were analyzed for pesticide residues. A more detailed description of the monitoring study was provided by Papadopoulou-Mourkidou et al. (2004).

Molinate and propanil were used as study compounds for the parameterization of RICEWQ and RIVWQ models. These compounds constitute the most common rice herbicides in Greece for the last 25 yr; monitoring studies in the specific studied area have documented their presence in the SW systems of the region. The 90th percentile of the yearly maximum daily PECs for propanil and molinate at a point 9 km along the length of the river were compared with the yearly maximum concentration of the two herbicides at the most downstream point of the Axios river during 1994. In addition, the temporal distribution of predicted and measured concentrations of molinate and propanil in the river were graphically assessed.

Application of Mitigation Measures
The effect of closure time applied to rice paddies after application of molinate on the magnitude of its PECs in the related drainage canals and river was examined. Three different closure periods at 5, 10, and 20 d after pesticide application were used for the parameterization of the RICEWQ model, and the predicted pesticide mass and water volume lost by rice paddies were used as loadings into the RIVWQ 2.02 model. This does not apply to propanil, for which no mitigation measures were tested due to its mode of application. Propanil is commonly applied onto bare soil after rice paddies are left to completely drain. Paddies are reflooded 24 h after application of propanil, and water circulation is reestablished 48 h later.

Risk Assessment for Propanil and Molinate
The risk of GW contamination from the use of molinate and propanil was assessed by calculating their annual average concentrations leaching below the 1-m soil zone for the 20-yr simulation. In accordance with FOCUS GW recommendations, the 80th percentile value was considered as the GW PECs for assessing the potential risk for GW resources.

The yearly maximum daily PECs at specific points (1, 2, 3, 4, and 5 km) along the length of each of the 10 drainage canals were obtained for all simulation years, and the 90th percentile values were calculated. The yearly maximum daily PECs at all junction points between drainage canals and river and also at specific points downstream of the last junction (7, 8, and 9 km) were summarized, and the 90th percentile values were determined. The 90th percentile PECs, considered as worst-case values, were used for the calculation of acute toxicity exposure ratios (TERacute) for aquatic organisms using the following equation:

Formula 1[1]
Acute (EC50, LC50 [µg L–1]) toxicological endpoints for the aquatic organisms used in standard risk assessment procedure for pesticide registration purposes were derived from literature databases (PAN Pesticide Ecotoxicity, 2005) and other sources (USEPA, 2003a, 2003b). The TERacute trigger values of 100 for fishes, invertebrates, and other aquatics, and 10 for algae and aquatic plants, in accordance with EC Directive 91/414/EC (http://europa.eu.int/comm/food/index_en.htm [verified 28 Nov. 2005]), are used as a first step for assessing the potential risk for acute effects on aquatic species. Our study focused on the assessment of the acute effects of pesticides on aquatic organisms since a calculation of acute/chronic effect ratio revealed that for the specific herbicides acute effects were more significant than chronic.

Statistical Analysis
We performed estimation of percentiles and other statistical analysis, including one-way analysis of variance between the three different closure periods applied as mitigation strategies for molinate, using the statistical package SPSS 11.0.1v (SPSS Inc., 2001).


   RESULTS AND DISCUSSION
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Groundwater Risk Assessment
A preliminary assessment of the effect of application date on the GW PECs of propanil and molinate indicated a significant effect for propanil (P < 0.05) and a less pronounced effect for molinate (data not shown). Therefore, the combination of agronomic practices (seeding dates) in simulated rice paddies, which produced the highest annual GW PECs for propanil and molinate, were obtained and used in GW risk assessment. The highest annual GW PECs of propanil and molinate were observed in rice fields that were seeded with rice 29 April and 23 April, respectively. The annual predicted concentrations (n = 20 yr) of propanil and molinate at the 1-m soil depth for the paddies treated on the above-mentioned dates were calculated, and the results are presented in Fig. 2 and 3. Propanil appeared to pose no risk for GW contamination in the specific scenario with the low infiltration rate (0.1 mm d–1). In accordance with FOCUS GW (2000), the 80th percentile annual average concentration of propanil for the 20-yr simulation period was 0.32 fg L–1 (Fig. 2). This is in agreement with previous GW monitoring studies in rice cultivated areas. In a USGS monitoring study employed in a representative rice-cultivated area in the USA, propanil concentrations never exceeded the detection limit of 0.05 µg L–1 (USEPA, 2003a). Propanil is a high-input herbicide, which is commonly applied in rice fields in Europe. However, its relatively high dissipation rate in water and sediment limits its movement to deeper layers of the soil horizon beneath rice paddy fields (Santos et al., 1998).


Figure 2
View larger version (13K):
[in this window]
[in a new window]
  		Fig. 2. Annual average groundwater predicted environmental concentrations of propanil for a 20-yr simulation period. The shaded bar represents the 80th percentile value.

 

Figure 3
View larger version (27K):
[in this window]
[in a new window]
  		Fig. 3. Annual average groundwater predicted environmental concentrations of molinate for a 20-yr simulation period when a paddy closure period of (a) 5, (b) 10, and (c) 20 d was applied. The shaded bar represents the 80th percentile value.

 
No risk for GW contamination was predicted for molinate in the specific scenario. The 80th percentile annual average concentrations of molinate for the 20-yr simulation period were 1.91, 1.90, and 1.88 pg L–1 for paddy closure times of 20, 10, and 5 d, respectively (Fig. 3). A slightly higher GW PEC was predicted by RICEWQ 1.6.2v model when the longer paddy closure period of 20 d was applied, compared with the corresponding PECs for the 10- and 5-d closure time. The longer period of paddy closure after application of molinate minimized pesticide losses by controlled drainage and allowed significant molinate mass to be available in paddy water for vertical movement to deeper soil layers. Molinate is considered a relatively mobile chemical, but its high volatility contributes to its rapid dissipation in the field and therefore diminishes the amount of pesticide in paddy water and sediment that is available for leaching (Soderquist et al., 1977; Ross and Sava, 1986). Charizopoulos and Papadopoulou-Mourkidou (1999) found that molinate was one of the most frequently detected pesticides in the rain water collected in the Axios river basin; this was attributed to the high volatility of molinate. Previous studies have indicated the difficulties associated with the parameterization of the volatilization routine of the RICEWQ model (Ferrari et al., 2005). Therefore, a lumped degradation rate value for molinate, including volatilization, was used for model parameterization.

Our results contrast the previous findings by Riparbelli et al. (1996), who reported the frequent detection of molinate at concentrations exceeding 0.1 µg L–1 in GW aquifers in Lombardy, northern Italy. This discrepancy could be attributed to the use of the clay Med-Rice scenario, which is characterized by low infiltration rate (0.1 cm d–1), thus representing low risk for GW contamination. According to Miao et al. (2003a), 40% of the rice cropping area in Italy consists of sand and gravel-based soils, which are characterized by higher infiltration rates, increasing the risk for GW contamination by molinate. In previous studies, application of the sandy Med-Rice scenario, which uses an infiltration rate of 1 cm d–1, resulted in GW PECs marginally exceeding 0.1 µg L–1, which is the maximum allowable limit for a single pesticide in drinking water in the EU (Karpouzas and Capri, 2006). Application of the RICEWQ 1.6.2v model enabled us to estimate GW PECs, which could be used for regulatory purposes. Furthermore, application of RICEWQ 1.6.2v in combination with a geographical information system could be a powerful modeling tool for assessing the vulnerability of specific rice-cultivated areas for GW contamination by certain rice pesticides.

Predicted Concentrations of Propanil and Molinate in Surface Water Systems
The yearly maximum daily PECs of propanil and molinate in the drainage canals and the river system of the local scenario are summarized in Table 4. Concentrations of propanil in the simulated drainage canals and river were markedly lower than the corresponding concentrations of molinate. This could be attributed to the rapid dissipation of propanil in rice paddy systems (Santos et al., 1998; USEPA, 2003a) compared with the slower dissipation of molinate (Ross and Sava, 1986). However, the maximum daily PECs of molinate in the simulated drainage canals were within the range reported in the literature. For example, in Australia, a monitoring study employed by the local Department of Water Resources (unpublished data, 1992) detected a concentration of 42 µg L–1 of molinate in drainage canals receiving waters from rice paddies.


View this table:
[in this window]
[in a new window]
  		Table 4. The 90th percentile of the maximum daily predicted environmental concentrations (µg L–1) and measured concentrations of propanil and molinate in drainage canals and river.

 
Application of longer closure periods after application of molinate resulted in significantly lower PECs (P < 0.05) in the simulated drainage canals and river. Therefore, the 90th percentile of the yearly maximum daily PECs of molinate in the drainage canals were 26.7, 11.6, and 6.1 µg L–1 for 5-, 10-, and 20-d closure time (DCT), respectively (Table 4). A similar trend was evident for the yearly maximum daily PECs of molinate in the river. This was not surprising since the half-life of molinate used for model parameterization was approximately 6 d (USEPA, 2003b). Therefore, at the end of a 5-d paddy closure after application of molinate, more than 50% of the initially applied herbicide (3.75 kg ha–1) will still be present in the paddy water, which is released to drainage canals when controlled drainage of the paddy commences. Consequently, the longer period of paddy closure after application of molinate allows longer time for pesticide dissipation (i.e., by biodegradation, photodegradation, volatilization) and diminishes the likelihood for contamination of adjacent SW systems with high levels of pesticides. Our simulation results are consistent with experimental observations reported by Crepeau and Kuivila (2000), who found that increasing the period of paddy closure after application of molinate from 19 to 28 d resulted in a lower concentration of the herbicide in the Sacramento River, which was receiving drainage water from rice paddy fields.

The duration of paddy closure after pesticide application should be considered in relation to pesticide persistence and efficacy. For example, in a field study employed in northern Italy with the rice herbicide pretilachlor [2-chloro-2',6'-diethyl-N-(2-propoxyethyl)acetanilide], Vidotto et al. (2004) found that a closure period of 18 to 24 d after pesticide application resulted in reduced pretilachlor residues in the drainage water of the treated paddy. In a similar field study with the rice herbicide cinosulfuron {1-(4,6-dimethoxy-1,3,5-triazin-2-yl)-3-[2-(2-methoxyethoxy)phenylsulfonyl]urea}, an increase in the duration of paddy closure from 14 to 22 d resulted in a significant reduction of pesticide residues in the water of the paddy outlet from 16.3 to 3.3 µg L–1, respectively (Ferrero et al., 2001). These results suggest that application of such a long closure period was efficient in preventing contamination of receiving SW systems with high amounts of pretilachlor and cinosulfuron. A longer paddy closure should be used as an effective mitigation strategy after the application of rather persistent pesticides. This would allow the dissipation of pesticides from paddy water during the closure time and thus diminish the mass of pesticide released to drainage canals at the end of the closure period.

Validation of Scenario with Measured Pesticide Concentrations
The yearly maximum measured concentrations of propanil and molinate in the Axios river are shown in Table 4. The maximum measured concentration of propanil in the Axios river in 1994 was in close agreement with the 90th percentile of the yearly maximum daily PECs of the herbicides in the river. The 90th percentile of the annual maximum PECs for propanil was 1.15 µg L–1 compared with the maximum measured concentrations of 1.0 µg L–1. The maximum measured concentration of molinate in the Axios river in 1994 (0.3 µg L–1) was in good agreement with the 90th percentile of the annual maximum daily PECs for molinate only when a closure period of 20 d was applied for model parameterization. Significantly higher PECs (P < 0.05) were obtained when a 5- or 10-d closure period was applied in simulated paddies after application of molinate. In Greece, a closure period of 5 to 20 d is recommended after application of molinate, depending on the visual efficacy of the herbicide against weeds.


View this table:
[in this window]
[in a new window]
  		Table 5. Acute toxicity exposure ratios (TERacute) for propanil and molinate in the drainage canals and the river of the simulated basin.

 
Generally, the maximum measured concentrations were lower than the 90th percentile of the maximum daily PECs for both herbicides. This could be attributed to an inherent conservatism of the developed scenario. In our study, all simulated paddies were annually treated with the maximum recommended dose of molinate and propanil, which is a rather conservative assumption. A recent survey employed in the Axios river basin showed that more than 70% of local rice farmers utilize molinate and/or propanil alone or as a tank mixture with other herbicides (e.g., pretilachlor, bentazone [3-isopropyl-1H-2,1,3-benzothiadiazin-4(3H)-one 2,2-dioxide]) as a standard herbicide treatment practice (DowHelanco, personal communication, 2002). Accordingly, when only 70% of the simulated basin was assumed treated with molinate and a 10-DCT of the paddy was used, the model, as expected, predicted lower yearly maximum daily PECs. Therefore, the 90th percentile value was 1.13 µg L–1, compared with the corresponding PECs of 1.76 µg L–1 when 100% of the rice-cultivated basin was assumed treated with molinate (Table 5). In a complementary study by Warren et al. (2004), only 25% of the simulated rice-cultivated area was considered treated with the rice-herbicide bensulfuron-methyl (methyl 2-[[[[[(4,6-dimethoxypyrimi-din-2-yl) amino]-carbonyl]-amino]-sulfonyl]methyl] benzoate). However, in Europe this herbicide is used less frequently than propanil and molinate. Consequently, in the developed scenario, the surface area of the simulated basin, which will be hypothetically treated with a studied pesticide, could be adjusted according to local information on the areas treated with the specific pesticide.

The range of predicted concentrations encompasses the measured pesticide concentrations in magnitude, and their temporal distribution in the riverine water co-occurs with the measured values of propanil and molinate as shown in Fig. 4 and 5, respectively. The maximum concentrations of propanil and molinate were measured in the water of the Axios river shortly after their application in local rice paddies (i.e., 19 June and 19 May for propanil and molinate, respectively), compared with 14 June and 16 May for the corresponding daily maximum PECs for the two herbicides. Several studies have reported a similar temporal trend of pesticide concentration in SW systems related to a rice-cultivated basin (Crepeau and Kuivila, 2000; Sudo et al., 2002; Papadopoulou-Mourkidou et al., 2004; Vu et al., 2004).


Figure 4
View larger version (22K):
[in this window]
[in a new window]
  		Fig. 4. Comparison of measured concentrations of propanil in the Axios River in 1994 with (a) the average daily predicted environmental concentrations of propanil in the simulated river for the 20-yr simulation period and (b) the daily predicted environmental concentrations of propanil in years when the minimum and maximum daily predicted environmental concentrations were observed. Error bars represent the standard deviation of the mean (n = 20 yr).

 

Figure 5
View larger version (23K):
[in this window]
[in a new window]
  		Fig. 5. Comparison of measured concentrations of propanil in the Axios River in 1994 with (a) the average daily predicted environmental concentrations of propanil in the simulated river for the 20-yr simulation period and (b) the daily predicted environmental concentrations of propanil in years when the minimum and maximum daily predicted environmental concentrations were observed. Error bars represent the standard deviation of the means (n = 20 yr).

 
Surface Water Risk Assessment
The 90th percentile of yearly maximum daily PECs of propanil and molinate in the drainage canals and the river (Table 4) were used for calculating TERacute for aquatic indicator species, which are routinely used in the pesticide regulatory process, and the results are shown in Table 5. The aquatic risk assessment suggested that propanil poses a low risk for aquatic organisms in both the drainage canals and river of the simulated basin.

The aquatic risk assessment suggested that molinate may pose a risk for certain aquatic organisms, including Daphnia magna, alga (Selenastrum capricornutum), fish (Oncorhyncus mykiss), and Gamarus spp. in drainage canals (Table 5). The drainage canals of the Axios river basin are not considered permanent SW bodies, although they do occasionally maintain a low water level during winter, when they are not operative. Therefore, the risk assessment analysis for fish might not be relevant and at a higher tier analysis the relevance of each organism in pesticide risk analysis should be considered on a case-by-case basis. Low risk for alga and Gamarus spp. in drainage canals was indicated when the closure time after application of molinate was more than 10 d. Molinate may also pose a risk for the indicator fish Oncorhynchus mykiss in the simulated river system only when the shortest paddy closure time of 5 d was applied. However, molinate appears to pose low risk for any aquatic organism in the river when the closure time after application of molinate was more than 10 d. This was expected since molinate is a high-input herbicide and the duration of paddy closure after application of molinate has been found to be a significant parameter controlling its appearance in adjacent SW bodies (Crepeau and Kuivila, 2000). Our results are in agreement with previous monitoring and toxicological studies that have documented the frequent detection of molinate in SW systems at concentrations posing a risk for aquatic organisms, including algae and fish (Tarazona et al., 2003). However, it should be stressed that the trigger TER values of 100 and 10 set for aquatic risk assessment are rather conservative trigger ratios, which are used to incorporate the inherent uncertainty of both exposure and effect calculations. Therefore, failing these trigger values does not necessarily mean there is a high risk.


   CONCLUSIONS
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
This study has led to the development of a well-defined basin-scale scenario representative of rice cultivation in Greece. In general, GW PECs obtained after application of the RICEWQ 1.6.2v model for molinate and propanil were within the range reported in the literature and constitute no risk for GW contamination. Surface water PECs obtained by the combination of RICEWQ and RIVWQ models were in acceptable agreement with measured values from the simulated area, suggesting that this scenario could be potentially used in the registration process for pesticides used in rice paddies. Application of mitigation measures, including prolongation of paddy closure after pesticide application could be a useful strategy for minimizing the risk for high pesticide loads onto related SW bodies. Further studies should focus on the development of similar representative scenarios for other rice-cultivated basins in Europe, which in accordance with already established GW and SW FOCUS scenarios for other crops could be utilized in a uniform risk assessment procedure.


   ACKNOWLEDGMENTS
 
This paper was produced within the framework of a Marie Curie individual fellowship "Environmental risk analysis leading to simulate a sustainable ecosystem management in rice area" held by Dr. D. Karpouzas (QLK5-CT-2002-51598).

We thank Universita Cattolica del Sacro Cuore for covering the publication cost, contribution 2005.


   REFERENCES
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
D. L. Corwin, J. Hopmans, and G. H. de Rooij
From Field- to Landscape-Scale Vadose Zone Processes: Scale Issues, Modeling, and Monitoring
Vadose Zone J., March 8, 2006; 5(1): 129 - 139.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Karpouzas, D. G.
Right arrow Articles by Papadopoulou-Mourkidou, E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Karpouzas, D. G.
Right arrow Articles by Papadopoulou-Mourkidou, E.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Karpouzas, D. G.
Right arrow Articles by Papadopoulou-Mourkidou, E.
Related Collections
Right arrow Watershed and Landscape Processes
Right arrow Pesticides
Right arrow Vadose Zone Processes and Chemical Transport

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome